diff --git a/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/bash_tool.yaml b/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/bash_tool.yaml
deleted file mode 100644
index 659eaaae938f3bd0b07d6f5743ec4603c6959ad6..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/bash_tool.yaml
+++ /dev/null
@@ -1,223 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 886
- host: ""
- body: '{"max_tokens":40,"messages":[{"content":[{"text":"Generate a concise title for the following content:\n\nuse bash to create a file named test.txt with content ''hello bash''. do not print its timestamp\n \u003cthink\u003e\n\n\u003c/think\u003e","type":"text"}],"role":"user"}],"model":"claude-haiku-4-5-20251001","system":[{"text":"You will generate a short title based on the first message a user begins a conversation with.\n\n\u003crules\u003e\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\u003c/rules\u003e\n\n /no_think","type":"text"}],"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-haiku-4-5-20251001","id":"msg_01VzT95Yg52yRncWMcaLrmfC","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":177,"cache_creation_input_tokens":0,"cache_read_input_tokens":0,"cache_creation":{"ephemeral_5m_input_tokens":0,"ephemeral_1h_input_tokens":0},"output_tokens":3,"service_tier":"standard","inference_geo":"not_available"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"Bash create"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" file without timestamp"} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":177,"cache_creation_input_tokens":0,"cache_read_input_tokens":0,"output_tokens":9} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 1.104341792s
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31043
- host: ""
- body: "{\"max_tokens\":10000,\"messages\":[{\"content\":[{\"text\":\"\\u003csystem_reminder\\u003eThis is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\\u003c/system_reminder\\u003e\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"},{\"text\":\"use bash to create a file named test.txt with content 'hello bash'. do not print its timestamp\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"role\":\"user\"}],\"model\":\"claude-sonnet-4-6\",\"system\":[{\"text\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\u003ccritical_rules\\u003e\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default \\u003c4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\u003c/critical_rules\\u003e\\n\\n\\u003ccommunication_style\\u003e\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\u003c/communication_style\\u003e\\n\\n\\u003ccode_references\\u003e\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\u003c/code_references\\u003e\\n\\n\\u003cworkflow\\u003e\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\u003c/workflow\\u003e\\n\\n\\u003cdecision_making\\u003e\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\u003c/decision_making\\u003e\\n\\n\\u003cediting_files\\u003e\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\u003c/editing_files\\u003e\\n\\n\\u003cwhitespace_and_exact_matching\\u003e\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\u003c/whitespace_and_exact_matching\\u003e\\n\\n\\u003ctask_completion\\u003e\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\u003c/task_completion\\u003e\\n\\n\\u003cerror_handling\\u003e\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\u003c/error_handling\\u003e\\n\\n\\u003cmemory_instructions\\u003e\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\u003c/memory_instructions\\u003e\\n\\n\\u003ccode_conventions\\u003e\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\u003c/code_conventions\\u003e\\n\\n\\u003ctesting\\u003e\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\u003c/testing\\u003e\\n\\n\\u003ctool_usage\\u003e\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\u003cbash_commands\\u003e\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `\\u0026` for background processes that won't stop on their own (e.g., `node server.js \\u0026`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status \\u0026\\u0026 git diff HEAD \\u0026\\u0026 git log -n 3`)\\n\\u003c/bash_commands\\u003e\\n\\u003c/tool_usage\\u003e\\n\\n\\u003cproactiveness\\u003e\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\u003c/proactiveness\\u003e\\n\\n\\u003cfinal_answers\\u003e\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\u003c/final_answers\\u003e\\n\\n\\u003cenv\\u003e\\nWorking directory: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/bash_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\u003c/env\\u003e\\n\\n\\n\\n\\n\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"tool_choice\":{\"disable_parallel_tool_use\":false,\"type\":\"auto\"},\"tools\":[{\"input_schema\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"},\"name\":\"bash\",\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\u003ccross_platform\\u003e\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\u003c/cross_platform\\u003e\\n\\n\\u003cexecution_steps\\u003e\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with \\u003ccwd\\u003e\\u003c/cwd\\u003e tags\\n\\u003c/execution_steps\\u003e\\n\\n\\u003cusage_notes\\u003e\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '\\u0026\\u0026', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\u003c/usage_notes\\u003e\\n\\n\\u003cbackground_execution\\u003e\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `\\u0026` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\u003c/background_execution\\u003e\\n\\n\\u003cgit_commits\\u003e\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in \\u003ccommit_analysis\\u003e tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\u003ccrush@charm.land\\u003e\\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\u003c/git_commits\\u003e\\n\\n\\u003cpull_requests\\u003e\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in \\u003cpr_analysis\\u003e tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n\\n ## Summary\\n\\n \\u0026lt;1-3 bullet points\\u003e\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\u003c/pull_requests\\u003e\\n\\n\\u003cexamples\\u003e\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar \\u0026\\u0026 pytest tests\\n\\u003c/examples\\u003e\\n\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"},\"name\":\"download\",\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"},\"name\":\"edit\",\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\"},{\"input_schema\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"},\"name\":\"multiedit\",\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\"},{\"input_schema\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"},\"name\":\"fetch\",\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\"},{\"input_schema\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"glob\",\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\"},{\"input_schema\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"grep\",\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\"},{\"input_schema\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"},\"name\":\"ls\",\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\"},{\"input_schema\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"},\"name\":\"sourcegraph\",\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"},\"name\":\"view\",\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\"},{\"input_schema\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"},\"name\":\"write\",\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-sonnet-4-6","id":"msg_01G92Yy7Hi9yLMnYSFt6LSAC","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":3,"cache_creation_input_tokens":8288,"cache_read_input_tokens":0,"cache_creation":{"ephemeral_5m_input_tokens":8288,"ephemeral_1h_input_tokens":0},"output_tokens":59,"service_tier":"standard","inference_geo":"global"}}}
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"tool_use","id":"toolu_019uY95x1kBWzGRvM8Vv9Sfk","name":"bash","input":{},"caller":{"type":"direct"}} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"{\"desc"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"ription\""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":": \"Create "} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"test.txt wit"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"h conte"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"nt\""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":", \"com"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"mand"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"\": \"echo"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":" 'hello ba"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"sh' > "} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"/tm"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"p/crush-t"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"est/T"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"estCod"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"erAgen"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"t/anthro"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"pic-sonnet"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"/bash_t"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"ool/test.txt"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"\"}"} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"tool_use","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":3,"cache_creation_input_tokens":8288,"cache_read_input_tokens":0,"output_tokens":105} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 2.545587042s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31452
- host: ""
- body: "{\"max_tokens\":10000,\"messages\":[{\"content\":[{\"text\":\"\\u003csystem_reminder\\u003eThis is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\\u003c/system_reminder\\u003e\",\"type\":\"text\"},{\"text\":\"use bash to create a file named test.txt with content 'hello bash'. do not print its timestamp\",\"type\":\"text\"}],\"role\":\"user\"},{\"content\":[{\"id\":\"toolu_019uY95x1kBWzGRvM8Vv9Sfk\",\"input\":{\"command\":\"echo 'hello bash' \\u003e /tmp/crush-test/TestCoderAgent/anthropic-sonnet/bash_tool/test.txt\",\"description\":\"Create test.txt with content\"},\"name\":\"bash\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"tool_use\"}],\"role\":\"assistant\"},{\"content\":[{\"tool_use_id\":\"toolu_019uY95x1kBWzGRvM8Vv9Sfk\",\"cache_control\":{\"type\":\"ephemeral\"},\"content\":[{\"text\":\"no output\",\"type\":\"text\"}],\"type\":\"tool_result\"}],\"role\":\"user\"}],\"model\":\"claude-sonnet-4-6\",\"system\":[{\"text\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\u003ccritical_rules\\u003e\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default \\u003c4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\u003c/critical_rules\\u003e\\n\\n\\u003ccommunication_style\\u003e\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\u003c/communication_style\\u003e\\n\\n\\u003ccode_references\\u003e\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\u003c/code_references\\u003e\\n\\n\\u003cworkflow\\u003e\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\u003c/workflow\\u003e\\n\\n\\u003cdecision_making\\u003e\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\u003c/decision_making\\u003e\\n\\n\\u003cediting_files\\u003e\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\u003c/editing_files\\u003e\\n\\n\\u003cwhitespace_and_exact_matching\\u003e\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\u003c/whitespace_and_exact_matching\\u003e\\n\\n\\u003ctask_completion\\u003e\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\u003c/task_completion\\u003e\\n\\n\\u003cerror_handling\\u003e\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\u003c/error_handling\\u003e\\n\\n\\u003cmemory_instructions\\u003e\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\u003c/memory_instructions\\u003e\\n\\n\\u003ccode_conventions\\u003e\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\u003c/code_conventions\\u003e\\n\\n\\u003ctesting\\u003e\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\u003c/testing\\u003e\\n\\n\\u003ctool_usage\\u003e\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\u003cbash_commands\\u003e\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `\\u0026` for background processes that won't stop on their own (e.g., `node server.js \\u0026`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status \\u0026\\u0026 git diff HEAD \\u0026\\u0026 git log -n 3`)\\n\\u003c/bash_commands\\u003e\\n\\u003c/tool_usage\\u003e\\n\\n\\u003cproactiveness\\u003e\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\u003c/proactiveness\\u003e\\n\\n\\u003cfinal_answers\\u003e\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\u003c/final_answers\\u003e\\n\\n\\u003cenv\\u003e\\nWorking directory: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/bash_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\u003c/env\\u003e\\n\\n\\n\\n\\n\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"tool_choice\":{\"disable_parallel_tool_use\":false,\"type\":\"auto\"},\"tools\":[{\"input_schema\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"},\"name\":\"bash\",\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\u003ccross_platform\\u003e\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\u003c/cross_platform\\u003e\\n\\n\\u003cexecution_steps\\u003e\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with \\u003ccwd\\u003e\\u003c/cwd\\u003e tags\\n\\u003c/execution_steps\\u003e\\n\\n\\u003cusage_notes\\u003e\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '\\u0026\\u0026', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\u003c/usage_notes\\u003e\\n\\n\\u003cbackground_execution\\u003e\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `\\u0026` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\u003c/background_execution\\u003e\\n\\n\\u003cgit_commits\\u003e\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in \\u003ccommit_analysis\\u003e tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\u003ccrush@charm.land\\u003e\\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\u003c/git_commits\\u003e\\n\\n\\u003cpull_requests\\u003e\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in \\u003cpr_analysis\\u003e tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n\\n ## Summary\\n\\n \\u0026lt;1-3 bullet points\\u003e\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\u003c/pull_requests\\u003e\\n\\n\\u003cexamples\\u003e\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar \\u0026\\u0026 pytest tests\\n\\u003c/examples\\u003e\\n\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"},\"name\":\"download\",\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"},\"name\":\"edit\",\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\"},{\"input_schema\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"},\"name\":\"multiedit\",\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\"},{\"input_schema\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"},\"name\":\"fetch\",\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\"},{\"input_schema\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"glob\",\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\"},{\"input_schema\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"grep\",\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\"},{\"input_schema\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"},\"name\":\"ls\",\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\"},{\"input_schema\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"},\"name\":\"sourcegraph\",\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"},\"name\":\"view\",\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\"},{\"input_schema\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"},\"name\":\"write\",\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-sonnet-4-6","id":"msg_01CoFddpZjzAtL66Fv6R6wbF","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":1,"cache_creation_input_tokens":121,"cache_read_input_tokens":8288,"cache_creation":{"ephemeral_5m_input_tokens":121,"ephemeral_1h_input_tokens":0},"output_tokens":1,"service_tier":"standard","inference_geo":"global"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"Done"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"."} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":1,"cache_creation_input_tokens":121,"cache_read_input_tokens":8288,"output_tokens":5} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 1.475161917s
diff --git a/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/download_tool.yaml b/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/download_tool.yaml
deleted file mode 100644
index e6f1d54a04af36cb2487a80f1e7ccaff7b83b853..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/download_tool.yaml
+++ /dev/null
@@ -1,280 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 907
- host: ""
- body: '{"max_tokens":40,"messages":[{"content":[{"text":"Generate a concise title for the following content:\n\ndownload the file from https://example-files.online-convert.com/document/txt/example.txt and save it as example.txt\n \u003cthink\u003e\n\n\u003c/think\u003e","type":"text"}],"role":"user"}],"model":"claude-haiku-4-5-20251001","system":[{"text":"You will generate a short title based on the first message a user begins a conversation with.\n\n\u003crules\u003e\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\u003c/rules\u003e\n\n /no_think","type":"text"}],"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-haiku-4-5-20251001","id":"msg_01FrLY1GjC7uhHuN4Wm3K2Er","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":185,"cache_creation_input_tokens":0,"cache_read_input_tokens":0,"cache_creation":{"ephemeral_5m_input_tokens":0,"ephemeral_1h_input_tokens":0},"output_tokens":1,"service_tier":"standard","inference_geo":"not_available"}}}
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"Download"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" and save example.txt file"} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":185,"cache_creation_input_tokens":0,"cache_read_input_tokens":0,"output_tokens":10} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 849.119083ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31068
- host: ""
- body: "{\"max_tokens\":10000,\"messages\":[{\"content\":[{\"text\":\"\\u003csystem_reminder\\u003eThis is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\\u003c/system_reminder\\u003e\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"},{\"text\":\"download the file from https://example-files.online-convert.com/document/txt/example.txt and save it as example.txt\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"role\":\"user\"}],\"model\":\"claude-sonnet-4-6\",\"system\":[{\"text\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\u003ccritical_rules\\u003e\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default \\u003c4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\u003c/critical_rules\\u003e\\n\\n\\u003ccommunication_style\\u003e\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\u003c/communication_style\\u003e\\n\\n\\u003ccode_references\\u003e\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\u003c/code_references\\u003e\\n\\n\\u003cworkflow\\u003e\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\u003c/workflow\\u003e\\n\\n\\u003cdecision_making\\u003e\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\u003c/decision_making\\u003e\\n\\n\\u003cediting_files\\u003e\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\u003c/editing_files\\u003e\\n\\n\\u003cwhitespace_and_exact_matching\\u003e\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\u003c/whitespace_and_exact_matching\\u003e\\n\\n\\u003ctask_completion\\u003e\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\u003c/task_completion\\u003e\\n\\n\\u003cerror_handling\\u003e\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\u003c/error_handling\\u003e\\n\\n\\u003cmemory_instructions\\u003e\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\u003c/memory_instructions\\u003e\\n\\n\\u003ccode_conventions\\u003e\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\u003c/code_conventions\\u003e\\n\\n\\u003ctesting\\u003e\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\u003c/testing\\u003e\\n\\n\\u003ctool_usage\\u003e\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\u003cbash_commands\\u003e\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `\\u0026` for background processes that won't stop on their own (e.g., `node server.js \\u0026`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status \\u0026\\u0026 git diff HEAD \\u0026\\u0026 git log -n 3`)\\n\\u003c/bash_commands\\u003e\\n\\u003c/tool_usage\\u003e\\n\\n\\u003cproactiveness\\u003e\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\u003c/proactiveness\\u003e\\n\\n\\u003cfinal_answers\\u003e\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\u003c/final_answers\\u003e\\n\\n\\u003cenv\\u003e\\nWorking directory: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/download_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\u003c/env\\u003e\\n\\n\\n\\n\\n\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"tool_choice\":{\"disable_parallel_tool_use\":false,\"type\":\"auto\"},\"tools\":[{\"input_schema\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"},\"name\":\"bash\",\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\u003ccross_platform\\u003e\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\u003c/cross_platform\\u003e\\n\\n\\u003cexecution_steps\\u003e\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with \\u003ccwd\\u003e\\u003c/cwd\\u003e tags\\n\\u003c/execution_steps\\u003e\\n\\n\\u003cusage_notes\\u003e\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '\\u0026\\u0026', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\u003c/usage_notes\\u003e\\n\\n\\u003cbackground_execution\\u003e\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `\\u0026` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\u003c/background_execution\\u003e\\n\\n\\u003cgit_commits\\u003e\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in \\u003ccommit_analysis\\u003e tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\u003ccrush@charm.land\\u003e\\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\u003c/git_commits\\u003e\\n\\n\\u003cpull_requests\\u003e\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in \\u003cpr_analysis\\u003e tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n\\n ## Summary\\n\\n \\u0026lt;1-3 bullet points\\u003e\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\u003c/pull_requests\\u003e\\n\\n\\u003cexamples\\u003e\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar \\u0026\\u0026 pytest tests\\n\\u003c/examples\\u003e\\n\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"},\"name\":\"download\",\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"},\"name\":\"edit\",\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\"},{\"input_schema\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"},\"name\":\"multiedit\",\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\"},{\"input_schema\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"},\"name\":\"fetch\",\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\"},{\"input_schema\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"glob\",\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\"},{\"input_schema\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"grep\",\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\"},{\"input_schema\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"},\"name\":\"ls\",\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\"},{\"input_schema\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"},\"name\":\"sourcegraph\",\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"},\"name\":\"view\",\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\"},{\"input_schema\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"},\"name\":\"write\",\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-sonnet-4-6","id":"msg_013L7WUa6efMvkni2WdapPMT","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":3,"cache_creation_input_tokens":4563,"cache_read_input_tokens":3733,"cache_creation":{"ephemeral_5m_input_tokens":4563,"ephemeral_1h_input_tokens":0},"output_tokens":54,"service_tier":"standard","inference_geo":"global"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"tool_use","id":"toolu_01Jwoa3BFj9VHiSy3J9stPir","name":"download","input":{},"caller":{"type":"direct"}} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"{\"url\": \""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"https://ex"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"ample-fi"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"les.onl"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"ine-conver"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"t.com/docum"}}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"ent/txt"}}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"/ex"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"ample.t"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"xt\""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":", \"file_pa"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"th\": "} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"\"/tm"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"p/cr"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"ush-te"}}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"st/TestCode"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"rAgent"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"/anthrop"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"ic"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"-sonn"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"et/d"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"ownload_t"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"ool/examp"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"le.txt\"}"} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"tool_use","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":3,"cache_creation_input_tokens":4563,"cache_read_input_tokens":3733,"output_tokens":113} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 3.697473292s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 0
- host: example-files.online-convert.com
- headers:
- User-Agent:
- - crush/1.0
- url: https://example-files.online-convert.com/document/txt/example.txt
- method: GET
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |-
- TXT test file
- Purpose: Provide example of this file type
- Document file type: TXT
- Version: 1.0
- Remark:
-
- Example content:
- The names "John Doe" for males, "Jane Doe" or "Jane Roe" for females, or "Jonnie Doe" and "Janie Doe" for children, or just "Doe" non-gender-specifically are used as placeholder names for a party whose true identity is unknown or must be withheld in a legal action, case, or discussion. The names are also used to refer to acorpse or hospital patient whose identity is unknown. This practice is widely used in the United States and Canada, but is rarely used in other English-speaking countries including the United Kingdom itself, from where the use of "John Doe" in a legal context originates. The names Joe Bloggs or John Smith are used in the UK instead, as well as in Australia and New Zealand.
-
- John Doe is sometimes used to refer to a typical male in other contexts as well, in a similar manner to John Q. Public, known in Great Britain as Joe Public, John Smith or Joe Bloggs. For example, the first name listed on a form is often John Doe, along with a fictional address or other fictional information to provide an example of how to fill in the form. The name is also used frequently in popular culture, for example in the Frank Capra film Meet John Doe. John Doe was also the name of a 2002 American television series.
-
- Similarly, a child or baby whose identity is unknown may be referred to as Baby Doe. A notorious murder case in Kansas City, Missouri, referred to the baby victim as Precious Doe. Other unidentified female murder victims are Cali Doe and Princess Doe. Additional persons may be called James Doe, Judy Doe, etc. However, to avoid possible confusion, if two anonymous or unknown parties are cited in a specific case or action, the surnames Doe and Roe may be used simultaneously; for example, "John Doe v. Jane Roe". If several anonymous parties are referenced, they may simply be labelled John Doe #1, John Doe #2, etc. (the U.S. Operation Delego cited 21 (numbered) "John Doe"s) or labelled with other variants of Doe / Roe / Poe / etc. Other early alternatives such as John Stiles and Richard Miles are now rarely used, and Mary Major has been used in some American federal cases.
-
-
-
- File created by https://www.online-convert.com
- More example files: https://www.online-convert.com/file-type
- Text of Example content: Wikipedia (https://en.wikipedia.org/wiki/John_Doe)
- License: Attribution-ShareAlike 4.0 (https://creativecommons.org/licenses/by-sa/4.0/)
-
- Feel free to use and share the file according to the license above.
- headers:
- Content-Type:
- - text/plain; charset=UTF-8
- status: 200 OK
- code: 200
- duration: 1.587507958s
-- id: 3
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31576
- host: ""
- body: "{\"max_tokens\":10000,\"messages\":[{\"content\":[{\"text\":\"\\u003csystem_reminder\\u003eThis is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\\u003c/system_reminder\\u003e\",\"type\":\"text\"},{\"text\":\"download the file from https://example-files.online-convert.com/document/txt/example.txt and save it as example.txt\",\"type\":\"text\"}],\"role\":\"user\"},{\"content\":[{\"id\":\"toolu_01Jwoa3BFj9VHiSy3J9stPir\",\"input\":{\"file_path\":\"/tmp/crush-test/TestCoderAgent/anthropic-sonnet/download_tool/example.txt\",\"url\":\"https://example-files.online-convert.com/document/txt/example.txt\"},\"name\":\"download\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"tool_use\"}],\"role\":\"assistant\"},{\"content\":[{\"tool_use_id\":\"toolu_01Jwoa3BFj9VHiSy3J9stPir\",\"cache_control\":{\"type\":\"ephemeral\"},\"content\":[{\"text\":\"Successfully downloaded 2574 bytes to example.txt (Content-Type: text/plain; charset=UTF-8)\",\"type\":\"text\"}],\"type\":\"tool_result\"}],\"role\":\"user\"}],\"model\":\"claude-sonnet-4-6\",\"system\":[{\"text\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\u003ccritical_rules\\u003e\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default \\u003c4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\u003c/critical_rules\\u003e\\n\\n\\u003ccommunication_style\\u003e\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\u003c/communication_style\\u003e\\n\\n\\u003ccode_references\\u003e\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\u003c/code_references\\u003e\\n\\n\\u003cworkflow\\u003e\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\u003c/workflow\\u003e\\n\\n\\u003cdecision_making\\u003e\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\u003c/decision_making\\u003e\\n\\n\\u003cediting_files\\u003e\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\u003c/editing_files\\u003e\\n\\n\\u003cwhitespace_and_exact_matching\\u003e\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\u003c/whitespace_and_exact_matching\\u003e\\n\\n\\u003ctask_completion\\u003e\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\u003c/task_completion\\u003e\\n\\n\\u003cerror_handling\\u003e\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\u003c/error_handling\\u003e\\n\\n\\u003cmemory_instructions\\u003e\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\u003c/memory_instructions\\u003e\\n\\n\\u003ccode_conventions\\u003e\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\u003c/code_conventions\\u003e\\n\\n\\u003ctesting\\u003e\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\u003c/testing\\u003e\\n\\n\\u003ctool_usage\\u003e\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\u003cbash_commands\\u003e\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `\\u0026` for background processes that won't stop on their own (e.g., `node server.js \\u0026`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status \\u0026\\u0026 git diff HEAD \\u0026\\u0026 git log -n 3`)\\n\\u003c/bash_commands\\u003e\\n\\u003c/tool_usage\\u003e\\n\\n\\u003cproactiveness\\u003e\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\u003c/proactiveness\\u003e\\n\\n\\u003cfinal_answers\\u003e\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\u003c/final_answers\\u003e\\n\\n\\u003cenv\\u003e\\nWorking directory: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/download_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\u003c/env\\u003e\\n\\n\\n\\n\\n\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"tool_choice\":{\"disable_parallel_tool_use\":false,\"type\":\"auto\"},\"tools\":[{\"input_schema\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"},\"name\":\"bash\",\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\u003ccross_platform\\u003e\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\u003c/cross_platform\\u003e\\n\\n\\u003cexecution_steps\\u003e\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with \\u003ccwd\\u003e\\u003c/cwd\\u003e tags\\n\\u003c/execution_steps\\u003e\\n\\n\\u003cusage_notes\\u003e\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '\\u0026\\u0026', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\u003c/usage_notes\\u003e\\n\\n\\u003cbackground_execution\\u003e\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `\\u0026` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\u003c/background_execution\\u003e\\n\\n\\u003cgit_commits\\u003e\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in \\u003ccommit_analysis\\u003e tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\u003ccrush@charm.land\\u003e\\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\u003c/git_commits\\u003e\\n\\n\\u003cpull_requests\\u003e\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in \\u003cpr_analysis\\u003e tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n\\n ## Summary\\n\\n \\u0026lt;1-3 bullet points\\u003e\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\u003c/pull_requests\\u003e\\n\\n\\u003cexamples\\u003e\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar \\u0026\\u0026 pytest tests\\n\\u003c/examples\\u003e\\n\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"},\"name\":\"download\",\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"},\"name\":\"edit\",\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\"},{\"input_schema\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"},\"name\":\"multiedit\",\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\"},{\"input_schema\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"},\"name\":\"fetch\",\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\"},{\"input_schema\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"glob\",\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\"},{\"input_schema\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"grep\",\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\"},{\"input_schema\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"},\"name\":\"ls\",\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\"},{\"input_schema\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"},\"name\":\"sourcegraph\",\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"},\"name\":\"view\",\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\"},{\"input_schema\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"},\"name\":\"write\",\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-sonnet-4-6","id":"msg_016nawrQnyqX44my8ax6XWQA","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":1,"cache_creation_input_tokens":152,"cache_read_input_tokens":8296,"cache_creation":{"ephemeral_5m_input_tokens":152,"ephemeral_1h_input_tokens":0},"output_tokens":3,"service_tier":"standard","inference_geo":"global"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"Saved as"}}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" `example.txt` (2574 bytes)."} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0}
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":1,"cache_creation_input_tokens":152,"cache_read_input_tokens":8296,"output_tokens":17} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 1.427229833s
diff --git a/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/fetch_tool.yaml b/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/fetch_tool.yaml
deleted file mode 100644
index ffc6bcf1b97bf1b8350e2caaf72826339d62e34f..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/fetch_tool.yaml
+++ /dev/null
@@ -1,281 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 928
- host: ""
- body: '{"max_tokens":40,"messages":[{"content":[{"text":"Generate a concise title for the following content:\n\nfetch the content from https://example-files.online-convert.com/website/html/example.html and tell me if it contains the word ''John Doe''\n \u003cthink\u003e\n\n\u003c/think\u003e","type":"text"}],"role":"user"}],"model":"claude-haiku-4-5-20251001","system":[{"text":"You will generate a short title based on the first message a user begins a conversation with.\n\n\u003crules\u003e\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\u003c/rules\u003e\n\n /no_think","type":"text"}],"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-haiku-4-5-20251001","id":"msg_014TzoSX5ygfeVjxiBpMuo5b","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":192,"cache_creation_input_tokens":0,"cache_read_input_tokens":0,"cache_creation":{"ephemeral_5m_input_tokens":0,"ephemeral_1h_input_tokens":0},"output_tokens":1,"service_tier":"standard","inference_geo":"not_available"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"Check"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" if example.html contains 'John Doe'"} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":192,"cache_creation_input_tokens":0,"cache_read_input_tokens":0,"output_tokens":15} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 912.685792ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31086
- host: ""
- body: "{\"max_tokens\":10000,\"messages\":[{\"content\":[{\"text\":\"\\u003csystem_reminder\\u003eThis is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\\u003c/system_reminder\\u003e\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"},{\"text\":\"fetch the content from https://example-files.online-convert.com/website/html/example.html and tell me if it contains the word 'John Doe'\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"role\":\"user\"}],\"model\":\"claude-sonnet-4-6\",\"system\":[{\"text\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\u003ccritical_rules\\u003e\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default \\u003c4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\u003c/critical_rules\\u003e\\n\\n\\u003ccommunication_style\\u003e\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\u003c/communication_style\\u003e\\n\\n\\u003ccode_references\\u003e\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\u003c/code_references\\u003e\\n\\n\\u003cworkflow\\u003e\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\u003c/workflow\\u003e\\n\\n\\u003cdecision_making\\u003e\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\u003c/decision_making\\u003e\\n\\n\\u003cediting_files\\u003e\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\u003c/editing_files\\u003e\\n\\n\\u003cwhitespace_and_exact_matching\\u003e\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\u003c/whitespace_and_exact_matching\\u003e\\n\\n\\u003ctask_completion\\u003e\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\u003c/task_completion\\u003e\\n\\n\\u003cerror_handling\\u003e\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\u003c/error_handling\\u003e\\n\\n\\u003cmemory_instructions\\u003e\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\u003c/memory_instructions\\u003e\\n\\n\\u003ccode_conventions\\u003e\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\u003c/code_conventions\\u003e\\n\\n\\u003ctesting\\u003e\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\u003c/testing\\u003e\\n\\n\\u003ctool_usage\\u003e\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\u003cbash_commands\\u003e\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `\\u0026` for background processes that won't stop on their own (e.g., `node server.js \\u0026`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status \\u0026\\u0026 git diff HEAD \\u0026\\u0026 git log -n 3`)\\n\\u003c/bash_commands\\u003e\\n\\u003c/tool_usage\\u003e\\n\\n\\u003cproactiveness\\u003e\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\u003c/proactiveness\\u003e\\n\\n\\u003cfinal_answers\\u003e\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\u003c/final_answers\\u003e\\n\\n\\u003cenv\\u003e\\nWorking directory: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/fetch_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\u003c/env\\u003e\\n\\n\\n\\n\\n\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"tool_choice\":{\"disable_parallel_tool_use\":false,\"type\":\"auto\"},\"tools\":[{\"input_schema\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"},\"name\":\"bash\",\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\u003ccross_platform\\u003e\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\u003c/cross_platform\\u003e\\n\\n\\u003cexecution_steps\\u003e\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with \\u003ccwd\\u003e\\u003c/cwd\\u003e tags\\n\\u003c/execution_steps\\u003e\\n\\n\\u003cusage_notes\\u003e\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '\\u0026\\u0026', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\u003c/usage_notes\\u003e\\n\\n\\u003cbackground_execution\\u003e\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `\\u0026` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\u003c/background_execution\\u003e\\n\\n\\u003cgit_commits\\u003e\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in \\u003ccommit_analysis\\u003e tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\u003ccrush@charm.land\\u003e\\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\u003c/git_commits\\u003e\\n\\n\\u003cpull_requests\\u003e\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in \\u003cpr_analysis\\u003e tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n\\n ## Summary\\n\\n \\u0026lt;1-3 bullet points\\u003e\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\u003c/pull_requests\\u003e\\n\\n\\u003cexamples\\u003e\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar \\u0026\\u0026 pytest tests\\n\\u003c/examples\\u003e\\n\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"},\"name\":\"download\",\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"},\"name\":\"edit\",\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\"},{\"input_schema\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"},\"name\":\"multiedit\",\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\"},{\"input_schema\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"},\"name\":\"fetch\",\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\"},{\"input_schema\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"glob\",\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\"},{\"input_schema\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"grep\",\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\"},{\"input_schema\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"},\"name\":\"ls\",\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\"},{\"input_schema\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"},\"name\":\"sourcegraph\",\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"},\"name\":\"view\",\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\"},{\"input_schema\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"},\"name\":\"write\",\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-sonnet-4-6","id":"msg_01HSBRtvbotaLjiVj7Qj1qer","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":3,"cache_creation_input_tokens":4570,"cache_read_input_tokens":3733,"cache_creation":{"ephemeral_5m_input_tokens":4570,"ephemeral_1h_input_tokens":0},"output_tokens":63,"service_tier":"standard","inference_geo":"global"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"tool_use","id":"toolu_01CHPnVTFReUM8NiE1Cmeg1q","name":"fetch","input":{},"caller":{"type":"direct"}} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"{\"u"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"rl\": "} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"\"h"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"ttps:/"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"/example-fi"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"les"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":".o"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"nline-co"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"nvert.com/w"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"ebsite/htm"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"l/example.h"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"tml\""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":", \"form"}}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"at\": \"t"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"ext\"}"} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"tool_use","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":3,"cache_creation_input_tokens":4570,"cache_read_input_tokens":3733,"output_tokens":87} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 1.936287791s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 0
- host: example-files.online-convert.com
- headers:
- User-Agent:
- - crush/1.0
- url: https://example-files.online-convert.com/website/html/example.html
- method: GET
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |-
-
-
-
-
-
- HTML test file
-
- Purpose: Provide example of this file type
- Document file type: HTML
- Version: 1.0
- Remark:
-
- 
-
- Example content:
- The names "John Doe" for males, "Jane Doe" or "Jane Roe" for females, or "Jonnie Doe" and "Janie Doe" for children, or just "Doe" non-gender-specifically are used as placeholder names for a party whose true identity is unknown or must be withheld in a legal action, case, or discussion. The names are also used to refer to acorpse or hospital patient whose identity is unknown. This practice is widely used in the United States and Canada, but is rarely used in other English-speaking countries including the United Kingdom itself, from where the use of "John Doe" in a legal context originates. The names Joe Bloggs or John Smith are used in the UK instead, as well as in Australia and New Zealand.
-
- John Doe is sometimes used to refer to a typical male in other contexts as well, in a similar manner to John Q. Public, known in Great Britain as Joe Public, John Smith or Joe Bloggs. For example, the first name listed on a form is often John Doe, along with a fictional address or other fictional information to provide an example of how to fill in the form. The name is also used frequently in popular culture, for example in the Frank Capra film Meet John Doe. John Doe was also the name of a 2002 American television series.
-
- Similarly, a child or baby whose identity is unknown may be referred to as Baby Doe. A notorious murder case in Kansas City, Missouri, referred to the baby victim as Precious Doe. Other unidentified female murder victims are Cali Doe and Princess Doe. Additional persons may be called James Doe, Judy Doe, etc. However, to avoid possible confusion, if two anonymous or unknown parties are cited in a specific case or action, the surnames Doe and Roe may be used simultaneously; for example, "John Doe v. Jane Roe". If several anonymous parties are referenced, they may simply be labelled John Doe #1, John Doe #2, etc. (the U.S. Operation Delego cited 21 (numbered) "John Doe"s) or labelled with other variants of Doe / Roe / Poe / etc. Other early alternatives such as John Stiles and Richard Miles are now rarely used, and Mary Major has been used in some American federal cases.
-
- 
-
- File created by https://www.online-convert.com
- More example files: https://www.online-convert.com/file-type
- Text of "Example content": Wikipedia
- License: Attribution-ShareAlike 3.0 Unported
-
- 
-
- Feel free to use and share the file according to license above.
-
-
-
- headers:
- Content-Type:
- - text/html; charset=UTF-8
- status: 200 OK
- code: 200
- duration: 480.971292ms
-- id: 3
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 33934
- host: ""
- body: "{\"max_tokens\":10000,\"messages\":[{\"content\":[{\"text\":\"\\u003csystem_reminder\\u003eThis is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\\u003c/system_reminder\\u003e\",\"type\":\"text\"},{\"text\":\"fetch the content from https://example-files.online-convert.com/website/html/example.html and tell me if it contains the word 'John Doe'\",\"type\":\"text\"}],\"role\":\"user\"},{\"content\":[{\"id\":\"toolu_01CHPnVTFReUM8NiE1Cmeg1q\",\"input\":{\"format\":\"text\",\"url\":\"https://example-files.online-convert.com/website/html/example.html\"},\"name\":\"fetch\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"tool_use\"}],\"role\":\"assistant\"},{\"content\":[{\"tool_use_id\":\"toolu_01CHPnVTFReUM8NiE1Cmeg1q\",\"cache_control\":{\"type\":\"ephemeral\"},\"content\":[{\"text\":\"HTML test file Purpose: Provide example of this file type Document file type: HTML Version: 1.0 Remark: Example content: The names \\\"John Doe\\\" for males, \\\"Jane Doe\\\" or \\\"Jane Roe\\\" for females, or \\\"Jonnie Doe\\\" and \\\"Janie Doe\\\" for children, or just \\\"Doe\\\" non-gender-specifically are used as placeholder names for a party whose true identity is unknown or must be withheld in a legal action, case, or discussion. The names are also used to refer to acorpse or hospital patient whose identity is unknown. This practice is widely used in the United States and Canada, but is rarely used in other English-speaking countries including the United Kingdom itself, from where the use of \\\"John Doe\\\" in a legal context originates. The names Joe Bloggs or John Smith are used in the UK instead, as well as in Australia and New Zealand. John Doe is sometimes used to refer to a typical male in other contexts as well, in a similar manner to John Q. Public, known in Great Britain as Joe Public, John Smith or Joe Bloggs. For example, the first name listed on a form is often John Doe, along with a fictional address or other fictional information to provide an example of how to fill in the form. The name is also used frequently in popular culture, for example in the Frank Capra film Meet John Doe. John Doe was also the name of a 2002 American television series. Similarly, a child or baby whose identity is unknown may be referred to as Baby Doe. A notorious murder case in Kansas City, Missouri, referred to the baby victim as Precious Doe. Other unidentified female murder victims are Cali Doe and Princess Doe. Additional persons may be called James Doe, Judy Doe, etc. However, to avoid possible confusion, if two anonymous or unknown parties are cited in a specific case or action, the surnames Doe and Roe may be used simultaneously; for example, \\\"John Doe v. Jane Roe\\\". If several anonymous parties are referenced, they may simply be labelled John Doe #1, John Doe #2, etc. (the U.S. Operation Delego cited 21 (numbered) \\\"John Doe\\\"s) or labelled with other variants of Doe / Roe / Poe / etc. Other early alternatives such as John Stiles and Richard Miles are now rarely used, and Mary Major has been used in some American federal cases. File created by https://www.online-convert.com More example files: https://www.online-convert.com/file-type Text of \\\"Example content\\\": Wikipedia License: Attribution-ShareAlike 3.0 Unported Feel free to use and share the file according to license above.\",\"type\":\"text\"}],\"type\":\"tool_result\"}],\"role\":\"user\"}],\"model\":\"claude-sonnet-4-6\",\"system\":[{\"text\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\u003ccritical_rules\\u003e\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default \\u003c4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\u003c/critical_rules\\u003e\\n\\n\\u003ccommunication_style\\u003e\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\u003c/communication_style\\u003e\\n\\n\\u003ccode_references\\u003e\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\u003c/code_references\\u003e\\n\\n\\u003cworkflow\\u003e\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\u003c/workflow\\u003e\\n\\n\\u003cdecision_making\\u003e\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\u003c/decision_making\\u003e\\n\\n\\u003cediting_files\\u003e\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\u003c/editing_files\\u003e\\n\\n\\u003cwhitespace_and_exact_matching\\u003e\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\u003c/whitespace_and_exact_matching\\u003e\\n\\n\\u003ctask_completion\\u003e\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\u003c/task_completion\\u003e\\n\\n\\u003cerror_handling\\u003e\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\u003c/error_handling\\u003e\\n\\n\\u003cmemory_instructions\\u003e\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\u003c/memory_instructions\\u003e\\n\\n\\u003ccode_conventions\\u003e\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\u003c/code_conventions\\u003e\\n\\n\\u003ctesting\\u003e\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\u003c/testing\\u003e\\n\\n\\u003ctool_usage\\u003e\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\u003cbash_commands\\u003e\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `\\u0026` for background processes that won't stop on their own (e.g., `node server.js \\u0026`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status \\u0026\\u0026 git diff HEAD \\u0026\\u0026 git log -n 3`)\\n\\u003c/bash_commands\\u003e\\n\\u003c/tool_usage\\u003e\\n\\n\\u003cproactiveness\\u003e\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\u003c/proactiveness\\u003e\\n\\n\\u003cfinal_answers\\u003e\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\u003c/final_answers\\u003e\\n\\n\\u003cenv\\u003e\\nWorking directory: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/fetch_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\u003c/env\\u003e\\n\\n\\n\\n\\n\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"tool_choice\":{\"disable_parallel_tool_use\":false,\"type\":\"auto\"},\"tools\":[{\"input_schema\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"},\"name\":\"bash\",\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\u003ccross_platform\\u003e\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\u003c/cross_platform\\u003e\\n\\n\\u003cexecution_steps\\u003e\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with \\u003ccwd\\u003e\\u003c/cwd\\u003e tags\\n\\u003c/execution_steps\\u003e\\n\\n\\u003cusage_notes\\u003e\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '\\u0026\\u0026', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\u003c/usage_notes\\u003e\\n\\n\\u003cbackground_execution\\u003e\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `\\u0026` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\u003c/background_execution\\u003e\\n\\n\\u003cgit_commits\\u003e\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in \\u003ccommit_analysis\\u003e tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\u003ccrush@charm.land\\u003e\\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\u003c/git_commits\\u003e\\n\\n\\u003cpull_requests\\u003e\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in \\u003cpr_analysis\\u003e tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n\\n ## Summary\\n\\n \\u0026lt;1-3 bullet points\\u003e\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\u003c/pull_requests\\u003e\\n\\n\\u003cexamples\\u003e\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar \\u0026\\u0026 pytest tests\\n\\u003c/examples\\u003e\\n\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"},\"name\":\"download\",\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"},\"name\":\"edit\",\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\"},{\"input_schema\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"},\"name\":\"multiedit\",\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\"},{\"input_schema\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"},\"name\":\"fetch\",\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\"},{\"input_schema\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"glob\",\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\"},{\"input_schema\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"grep\",\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\"},{\"input_schema\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"},\"name\":\"ls\",\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\"},{\"input_schema\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"},\"name\":\"sourcegraph\",\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"},\"name\":\"view\",\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\"},{\"input_schema\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"},\"name\":\"write\",\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-sonnet-4-6","id":"msg_01Vumdh8JemqVgZyMrPMmvZg","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":1,"cache_creation_input_tokens":714,"cache_read_input_tokens":8303,"cache_creation":{"ephemeral_5m_input_tokens":714,"ephemeral_1h_input_tokens":0},"output_tokens":1,"service_tier":"standard","inference_geo":"global"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"Yes"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":", the page contains the word **\"John Doe\"** — multiple times in fact. It appears in the context of an"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" explanation about placeholder names used in legal contexts, popular culture, and TV shows."} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":1,"cache_creation_input_tokens":714,"cache_read_input_tokens":8303,"output_tokens":46} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 1.620128333s
diff --git a/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/glob_tool.yaml b/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/glob_tool.yaml
deleted file mode 100644
index 97432040978d373b31cb59056d2b7beb6856193f..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/glob_tool.yaml
+++ /dev/null
@@ -1,170 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 847
- host: ""
- body: '{"max_tokens":40,"messages":[{"content":[{"text":"Generate a concise title for the following content:\n\nuse glob to find all .go files in the current directory\n \u003cthink\u003e\n\n\u003c/think\u003e","type":"text"}],"role":"user"}],"model":"claude-haiku-4-5-20251001","system":[{"text":"You will generate a short title based on the first message a user begins a conversation with.\n\n\u003crules\u003e\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\u003c/rules\u003e\n\n /no_think","type":"text"}],"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-haiku-4-5-20251001","id":"msg_016gFGvY1mJARHPK5geBtZca","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":167,"cache_creation_input_tokens":0,"cache_read_input_tokens":0,"cache_creation":{"ephemeral_5m_input_tokens":0,"ephemeral_1h_input_tokens":0},"output_tokens":1,"service_tier":"standard","inference_geo":"not_available"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"Finding"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" Go files with glob in current directory"} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":167,"cache_creation_input_tokens":0,"cache_read_input_tokens":0,"output_tokens":11} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 522.915792ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31004
- host: ""
- body: "{\"max_tokens\":10000,\"messages\":[{\"content\":[{\"text\":\"\\u003csystem_reminder\\u003eThis is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\\u003c/system_reminder\\u003e\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"},{\"text\":\"use glob to find all .go files in the current directory\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"role\":\"user\"}],\"model\":\"claude-sonnet-4-6\",\"system\":[{\"text\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\u003ccritical_rules\\u003e\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default \\u003c4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\u003c/critical_rules\\u003e\\n\\n\\u003ccommunication_style\\u003e\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\u003c/communication_style\\u003e\\n\\n\\u003ccode_references\\u003e\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\u003c/code_references\\u003e\\n\\n\\u003cworkflow\\u003e\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\u003c/workflow\\u003e\\n\\n\\u003cdecision_making\\u003e\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\u003c/decision_making\\u003e\\n\\n\\u003cediting_files\\u003e\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\u003c/editing_files\\u003e\\n\\n\\u003cwhitespace_and_exact_matching\\u003e\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\u003c/whitespace_and_exact_matching\\u003e\\n\\n\\u003ctask_completion\\u003e\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\u003c/task_completion\\u003e\\n\\n\\u003cerror_handling\\u003e\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\u003c/error_handling\\u003e\\n\\n\\u003cmemory_instructions\\u003e\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\u003c/memory_instructions\\u003e\\n\\n\\u003ccode_conventions\\u003e\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\u003c/code_conventions\\u003e\\n\\n\\u003ctesting\\u003e\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\u003c/testing\\u003e\\n\\n\\u003ctool_usage\\u003e\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\u003cbash_commands\\u003e\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `\\u0026` for background processes that won't stop on their own (e.g., `node server.js \\u0026`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status \\u0026\\u0026 git diff HEAD \\u0026\\u0026 git log -n 3`)\\n\\u003c/bash_commands\\u003e\\n\\u003c/tool_usage\\u003e\\n\\n\\u003cproactiveness\\u003e\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\u003c/proactiveness\\u003e\\n\\n\\u003cfinal_answers\\u003e\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\u003c/final_answers\\u003e\\n\\n\\u003cenv\\u003e\\nWorking directory: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/glob_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\u003c/env\\u003e\\n\\n\\n\\n\\n\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"tool_choice\":{\"disable_parallel_tool_use\":false,\"type\":\"auto\"},\"tools\":[{\"input_schema\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"},\"name\":\"bash\",\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\u003ccross_platform\\u003e\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\u003c/cross_platform\\u003e\\n\\n\\u003cexecution_steps\\u003e\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with \\u003ccwd\\u003e\\u003c/cwd\\u003e tags\\n\\u003c/execution_steps\\u003e\\n\\n\\u003cusage_notes\\u003e\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '\\u0026\\u0026', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\u003c/usage_notes\\u003e\\n\\n\\u003cbackground_execution\\u003e\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `\\u0026` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\u003c/background_execution\\u003e\\n\\n\\u003cgit_commits\\u003e\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in \\u003ccommit_analysis\\u003e tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\u003ccrush@charm.land\\u003e\\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\u003c/git_commits\\u003e\\n\\n\\u003cpull_requests\\u003e\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in \\u003cpr_analysis\\u003e tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n\\n ## Summary\\n\\n \\u0026lt;1-3 bullet points\\u003e\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\u003c/pull_requests\\u003e\\n\\n\\u003cexamples\\u003e\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar \\u0026\\u0026 pytest tests\\n\\u003c/examples\\u003e\\n\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"},\"name\":\"download\",\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"},\"name\":\"edit\",\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\"},{\"input_schema\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"},\"name\":\"multiedit\",\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\"},{\"input_schema\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"},\"name\":\"fetch\",\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\"},{\"input_schema\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"glob\",\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\"},{\"input_schema\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"grep\",\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\"},{\"input_schema\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"},\"name\":\"ls\",\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\"},{\"input_schema\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"},\"name\":\"sourcegraph\",\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"},\"name\":\"view\",\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\"},{\"input_schema\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"},\"name\":\"write\",\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-sonnet-4-6","id":"msg_013VWNT1aWS42zuFYc5UY28H","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":3,"cache_creation_input_tokens":4545,"cache_read_input_tokens":3733,"cache_creation":{"ephemeral_5m_input_tokens":4545,"ephemeral_1h_input_tokens":0},"output_tokens":53,"service_tier":"standard","inference_geo":"global"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"tool_use","id":"toolu_015J9ZKKfSNuZNQwWrrHsVpX","name":"glob","input":{},"caller":{"type":"direct"}} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"{\"pat"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"tern\": \"*"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":".go\"}"} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"tool_use","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":3,"cache_creation_input_tokens":4545,"cache_read_input_tokens":3733,"output_tokens":53} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 1.651803583s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31337
- host: ""
- body: "{\"max_tokens\":10000,\"messages\":[{\"content\":[{\"text\":\"\\u003csystem_reminder\\u003eThis is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\\u003c/system_reminder\\u003e\",\"type\":\"text\"},{\"text\":\"use glob to find all .go files in the current directory\",\"type\":\"text\"}],\"role\":\"user\"},{\"content\":[{\"id\":\"toolu_015J9ZKKfSNuZNQwWrrHsVpX\",\"input\":{\"pattern\":\"*.go\"},\"name\":\"glob\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"tool_use\"}],\"role\":\"assistant\"},{\"content\":[{\"tool_use_id\":\"toolu_015J9ZKKfSNuZNQwWrrHsVpX\",\"cache_control\":{\"type\":\"ephemeral\"},\"content\":[{\"text\":\"/tmp/crush-test/TestCoderAgent/anthropic-sonnet/glob_tool/main.go\",\"type\":\"text\"}],\"type\":\"tool_result\"}],\"role\":\"user\"}],\"model\":\"claude-sonnet-4-6\",\"system\":[{\"text\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\u003ccritical_rules\\u003e\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default \\u003c4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\u003c/critical_rules\\u003e\\n\\n\\u003ccommunication_style\\u003e\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\u003c/communication_style\\u003e\\n\\n\\u003ccode_references\\u003e\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\u003c/code_references\\u003e\\n\\n\\u003cworkflow\\u003e\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\u003c/workflow\\u003e\\n\\n\\u003cdecision_making\\u003e\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\u003c/decision_making\\u003e\\n\\n\\u003cediting_files\\u003e\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\u003c/editing_files\\u003e\\n\\n\\u003cwhitespace_and_exact_matching\\u003e\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\u003c/whitespace_and_exact_matching\\u003e\\n\\n\\u003ctask_completion\\u003e\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\u003c/task_completion\\u003e\\n\\n\\u003cerror_handling\\u003e\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\u003c/error_handling\\u003e\\n\\n\\u003cmemory_instructions\\u003e\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\u003c/memory_instructions\\u003e\\n\\n\\u003ccode_conventions\\u003e\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\u003c/code_conventions\\u003e\\n\\n\\u003ctesting\\u003e\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\u003c/testing\\u003e\\n\\n\\u003ctool_usage\\u003e\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\u003cbash_commands\\u003e\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `\\u0026` for background processes that won't stop on their own (e.g., `node server.js \\u0026`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status \\u0026\\u0026 git diff HEAD \\u0026\\u0026 git log -n 3`)\\n\\u003c/bash_commands\\u003e\\n\\u003c/tool_usage\\u003e\\n\\n\\u003cproactiveness\\u003e\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\u003c/proactiveness\\u003e\\n\\n\\u003cfinal_answers\\u003e\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\u003c/final_answers\\u003e\\n\\n\\u003cenv\\u003e\\nWorking directory: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/glob_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\u003c/env\\u003e\\n\\n\\n\\n\\n\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"tool_choice\":{\"disable_parallel_tool_use\":false,\"type\":\"auto\"},\"tools\":[{\"input_schema\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"},\"name\":\"bash\",\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\u003ccross_platform\\u003e\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\u003c/cross_platform\\u003e\\n\\n\\u003cexecution_steps\\u003e\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with \\u003ccwd\\u003e\\u003c/cwd\\u003e tags\\n\\u003c/execution_steps\\u003e\\n\\n\\u003cusage_notes\\u003e\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '\\u0026\\u0026', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\u003c/usage_notes\\u003e\\n\\n\\u003cbackground_execution\\u003e\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `\\u0026` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\u003c/background_execution\\u003e\\n\\n\\u003cgit_commits\\u003e\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in \\u003ccommit_analysis\\u003e tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\u003ccrush@charm.land\\u003e\\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\u003c/git_commits\\u003e\\n\\n\\u003cpull_requests\\u003e\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in \\u003cpr_analysis\\u003e tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n\\n ## Summary\\n\\n \\u0026lt;1-3 bullet points\\u003e\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\u003c/pull_requests\\u003e\\n\\n\\u003cexamples\\u003e\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar \\u0026\\u0026 pytest tests\\n\\u003c/examples\\u003e\\n\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"},\"name\":\"download\",\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"},\"name\":\"edit\",\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\"},{\"input_schema\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"},\"name\":\"multiedit\",\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\"},{\"input_schema\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"},\"name\":\"fetch\",\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\"},{\"input_schema\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"glob\",\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\"},{\"input_schema\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"grep\",\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\"},{\"input_schema\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"},\"name\":\"ls\",\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\"},{\"input_schema\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"},\"name\":\"sourcegraph\",\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"},\"name\":\"view\",\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\"},{\"input_schema\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"},\"name\":\"write\",\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-sonnet-4-6","id":"msg_01R3dNA8AxFM52yqMkochA1R","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":1,"cache_creation_input_tokens":4636,"cache_read_input_tokens":3733,"cache_creation":{"ephemeral_5m_input_tokens":4636,"ephemeral_1h_input_tokens":0},"output_tokens":1,"service_tier":"standard","inference_geo":"global"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"One"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" `.go` file found: `main.go`"} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":1,"cache_creation_input_tokens":4636,"cache_read_input_tokens":3733,"output_tokens":15}}
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 2.491762208s
diff --git a/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/grep_tool.yaml b/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/grep_tool.yaml
deleted file mode 100644
index a938649fc0c9e9b85060d7e1dfaee7e8dd23de30..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/grep_tool.yaml
+++ /dev/null
@@ -1,183 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 845
- host: ""
- body: '{"max_tokens":40,"messages":[{"content":[{"text":"Generate a concise title for the following content:\n\nuse grep to search for the word ''package'' in go files\n \u003cthink\u003e\n\n\u003c/think\u003e","type":"text"}],"role":"user"}],"model":"claude-haiku-4-5-20251001","system":[{"text":"You will generate a short title based on the first message a user begins a conversation with.\n\n\u003crules\u003e\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\u003c/rules\u003e\n\n /no_think","type":"text"}],"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-haiku-4-5-20251001","id":"msg_01DDuSaAjzBVC2GMGFH71zGK","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":169,"cache_creation_input_tokens":0,"cache_read_input_tokens":0,"cache_creation":{"ephemeral_5m_input_tokens":0,"ephemeral_1h_input_tokens":0},"output_tokens":3,"service_tier":"standard","inference_geo":"not_available"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"Grep search"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" for 'package' in Go files"}}
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":169,"cache_creation_input_tokens":0,"cache_read_input_tokens":0,"output_tokens":14} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 748.122792ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31002
- host: ""
- body: "{\"max_tokens\":10000,\"messages\":[{\"content\":[{\"text\":\"\\u003csystem_reminder\\u003eThis is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\\u003c/system_reminder\\u003e\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"},{\"text\":\"use grep to search for the word 'package' in go files\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"role\":\"user\"}],\"model\":\"claude-sonnet-4-6\",\"system\":[{\"text\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\u003ccritical_rules\\u003e\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default \\u003c4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\u003c/critical_rules\\u003e\\n\\n\\u003ccommunication_style\\u003e\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\u003c/communication_style\\u003e\\n\\n\\u003ccode_references\\u003e\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\u003c/code_references\\u003e\\n\\n\\u003cworkflow\\u003e\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\u003c/workflow\\u003e\\n\\n\\u003cdecision_making\\u003e\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\u003c/decision_making\\u003e\\n\\n\\u003cediting_files\\u003e\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\u003c/editing_files\\u003e\\n\\n\\u003cwhitespace_and_exact_matching\\u003e\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\u003c/whitespace_and_exact_matching\\u003e\\n\\n\\u003ctask_completion\\u003e\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\u003c/task_completion\\u003e\\n\\n\\u003cerror_handling\\u003e\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\u003c/error_handling\\u003e\\n\\n\\u003cmemory_instructions\\u003e\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\u003c/memory_instructions\\u003e\\n\\n\\u003ccode_conventions\\u003e\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\u003c/code_conventions\\u003e\\n\\n\\u003ctesting\\u003e\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\u003c/testing\\u003e\\n\\n\\u003ctool_usage\\u003e\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\u003cbash_commands\\u003e\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `\\u0026` for background processes that won't stop on their own (e.g., `node server.js \\u0026`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status \\u0026\\u0026 git diff HEAD \\u0026\\u0026 git log -n 3`)\\n\\u003c/bash_commands\\u003e\\n\\u003c/tool_usage\\u003e\\n\\n\\u003cproactiveness\\u003e\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\u003c/proactiveness\\u003e\\n\\n\\u003cfinal_answers\\u003e\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\u003c/final_answers\\u003e\\n\\n\\u003cenv\\u003e\\nWorking directory: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/grep_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\u003c/env\\u003e\\n\\n\\n\\n\\n\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"tool_choice\":{\"disable_parallel_tool_use\":false,\"type\":\"auto\"},\"tools\":[{\"input_schema\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"},\"name\":\"bash\",\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\u003ccross_platform\\u003e\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\u003c/cross_platform\\u003e\\n\\n\\u003cexecution_steps\\u003e\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with \\u003ccwd\\u003e\\u003c/cwd\\u003e tags\\n\\u003c/execution_steps\\u003e\\n\\n\\u003cusage_notes\\u003e\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '\\u0026\\u0026', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\u003c/usage_notes\\u003e\\n\\n\\u003cbackground_execution\\u003e\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `\\u0026` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\u003c/background_execution\\u003e\\n\\n\\u003cgit_commits\\u003e\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in \\u003ccommit_analysis\\u003e tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\u003ccrush@charm.land\\u003e\\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\u003c/git_commits\\u003e\\n\\n\\u003cpull_requests\\u003e\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in \\u003cpr_analysis\\u003e tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n\\n ## Summary\\n\\n \\u0026lt;1-3 bullet points\\u003e\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\u003c/pull_requests\\u003e\\n\\n\\u003cexamples\\u003e\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar \\u0026\\u0026 pytest tests\\n\\u003c/examples\\u003e\\n\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"},\"name\":\"download\",\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"},\"name\":\"edit\",\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\"},{\"input_schema\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"},\"name\":\"multiedit\",\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\"},{\"input_schema\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"},\"name\":\"fetch\",\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\"},{\"input_schema\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"glob\",\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\"},{\"input_schema\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"grep\",\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\"},{\"input_schema\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"},\"name\":\"ls\",\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\"},{\"input_schema\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"},\"name\":\"sourcegraph\",\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"},\"name\":\"view\",\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\"},{\"input_schema\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"},\"name\":\"write\",\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-sonnet-4-6","id":"msg_01KGADT4CBzheKtZvtuZyXCD","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":3,"cache_creation_input_tokens":4547,"cache_read_input_tokens":3733,"cache_creation":{"ephemeral_5m_input_tokens":4547,"ephemeral_1h_input_tokens":0},"output_tokens":54,"service_tier":"standard","inference_geo":"global"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"tool_use","id":"toolu_01TdBHCNqQsYHG8DBQcwwAyU","name":"grep","input":{},"caller":{"type":"direct"}} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"{\"patt"}}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"ern\": \"pac"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"kage\""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":", \"inclu"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"de\": \"*."} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"go\"}"}}
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"tool_use","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":3,"cache_creation_input_tokens":4547,"cache_read_input_tokens":3733,"output_tokens":70} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 2.821156834s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31407
- host: ""
- body: "{\"max_tokens\":10000,\"messages\":[{\"content\":[{\"text\":\"\\u003csystem_reminder\\u003eThis is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\\u003c/system_reminder\\u003e\",\"type\":\"text\"},{\"text\":\"use grep to search for the word 'package' in go files\",\"type\":\"text\"}],\"role\":\"user\"},{\"content\":[{\"id\":\"toolu_01TdBHCNqQsYHG8DBQcwwAyU\",\"input\":{\"include\":\"*.go\",\"pattern\":\"package\"},\"name\":\"grep\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"tool_use\"}],\"role\":\"assistant\"},{\"content\":[{\"tool_use_id\":\"toolu_01TdBHCNqQsYHG8DBQcwwAyU\",\"cache_control\":{\"type\":\"ephemeral\"},\"content\":[{\"text\":\"Found 1 matches\\n/tmp/crush-test/TestCoderAgent/anthropic-sonnet/grep_tool/main.go:\\n Line 1, Char 1: package main\\n\",\"type\":\"text\"}],\"type\":\"tool_result\"}],\"role\":\"user\"}],\"model\":\"claude-sonnet-4-6\",\"system\":[{\"text\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\u003ccritical_rules\\u003e\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default \\u003c4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\u003c/critical_rules\\u003e\\n\\n\\u003ccommunication_style\\u003e\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\u003c/communication_style\\u003e\\n\\n\\u003ccode_references\\u003e\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\u003c/code_references\\u003e\\n\\n\\u003cworkflow\\u003e\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\u003c/workflow\\u003e\\n\\n\\u003cdecision_making\\u003e\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\u003c/decision_making\\u003e\\n\\n\\u003cediting_files\\u003e\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\u003c/editing_files\\u003e\\n\\n\\u003cwhitespace_and_exact_matching\\u003e\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\u003c/whitespace_and_exact_matching\\u003e\\n\\n\\u003ctask_completion\\u003e\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\u003c/task_completion\\u003e\\n\\n\\u003cerror_handling\\u003e\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\u003c/error_handling\\u003e\\n\\n\\u003cmemory_instructions\\u003e\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\u003c/memory_instructions\\u003e\\n\\n\\u003ccode_conventions\\u003e\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\u003c/code_conventions\\u003e\\n\\n\\u003ctesting\\u003e\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\u003c/testing\\u003e\\n\\n\\u003ctool_usage\\u003e\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\u003cbash_commands\\u003e\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `\\u0026` for background processes that won't stop on their own (e.g., `node server.js \\u0026`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status \\u0026\\u0026 git diff HEAD \\u0026\\u0026 git log -n 3`)\\n\\u003c/bash_commands\\u003e\\n\\u003c/tool_usage\\u003e\\n\\n\\u003cproactiveness\\u003e\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\u003c/proactiveness\\u003e\\n\\n\\u003cfinal_answers\\u003e\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\u003c/final_answers\\u003e\\n\\n\\u003cenv\\u003e\\nWorking directory: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/grep_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\u003c/env\\u003e\\n\\n\\n\\n\\n\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"tool_choice\":{\"disable_parallel_tool_use\":false,\"type\":\"auto\"},\"tools\":[{\"input_schema\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"},\"name\":\"bash\",\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\u003ccross_platform\\u003e\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\u003c/cross_platform\\u003e\\n\\n\\u003cexecution_steps\\u003e\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with \\u003ccwd\\u003e\\u003c/cwd\\u003e tags\\n\\u003c/execution_steps\\u003e\\n\\n\\u003cusage_notes\\u003e\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '\\u0026\\u0026', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\u003c/usage_notes\\u003e\\n\\n\\u003cbackground_execution\\u003e\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `\\u0026` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\u003c/background_execution\\u003e\\n\\n\\u003cgit_commits\\u003e\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in \\u003ccommit_analysis\\u003e tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\u003ccrush@charm.land\\u003e\\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\u003c/git_commits\\u003e\\n\\n\\u003cpull_requests\\u003e\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in \\u003cpr_analysis\\u003e tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n\\n ## Summary\\n\\n \\u0026lt;1-3 bullet points\\u003e\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\u003c/pull_requests\\u003e\\n\\n\\u003cexamples\\u003e\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar \\u0026\\u0026 pytest tests\\n\\u003c/examples\\u003e\\n\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"},\"name\":\"download\",\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"},\"name\":\"edit\",\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\"},{\"input_schema\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"},\"name\":\"multiedit\",\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\"},{\"input_schema\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"},\"name\":\"fetch\",\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\"},{\"input_schema\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"glob\",\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\"},{\"input_schema\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"grep\",\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\"},{\"input_schema\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"},\"name\":\"ls\",\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\"},{\"input_schema\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"},\"name\":\"sourcegraph\",\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"},\"name\":\"view\",\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\"},{\"input_schema\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"},\"name\":\"write\",\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-sonnet-4-6","id":"msg_015Xu2T43K2gg1SQDFYJV222","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":1,"cache_creation_input_tokens":130,"cache_read_input_tokens":8280,"cache_creation":{"ephemeral_5m_input_tokens":130,"ephemeral_1h_input_tokens":0},"output_tokens":1,"service_tier":"standard","inference_geo":"global"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"Found"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" 1 match: `main.go:1` — `"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"package main`"} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":1,"cache_creation_input_tokens":130,"cache_read_input_tokens":8280,"output_tokens":21} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 1.5494585s
diff --git a/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/ls_tool.yaml b/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/ls_tool.yaml
deleted file mode 100644
index c38df25b93f0fed404955a23cf1549dc23036445..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/ls_tool.yaml
+++ /dev/null
@@ -1,162 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 841
- host: ""
- body: '{"max_tokens":40,"messages":[{"content":[{"text":"Generate a concise title for the following content:\n\nuse ls to list the files in the current directory\n \u003cthink\u003e\n\n\u003c/think\u003e","type":"text"}],"role":"user"}],"model":"claude-haiku-4-5-20251001","system":[{"text":"You will generate a short title based on the first message a user begins a conversation with.\n\n\u003crules\u003e\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\u003c/rules\u003e\n\n /no_think","type":"text"}],"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-haiku-4-5-20251001","id":"msg_01524QdtZ125onKVyuqZxJuM","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":165,"cache_creation_input_tokens":0,"cache_read_input_tokens":0,"cache_creation":{"ephemeral_5m_input_tokens":0,"ephemeral_1h_input_tokens":0},"output_tokens":1,"service_tier":"standard","inference_geo":"not_available"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"Using"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" ls to list directory files"} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":165,"cache_creation_input_tokens":0,"cache_read_input_tokens":0,"output_tokens":9}}
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 671.612666ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30996
- host: ""
- body: "{\"max_tokens\":10000,\"messages\":[{\"content\":[{\"text\":\"\\u003csystem_reminder\\u003eThis is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\\u003c/system_reminder\\u003e\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"},{\"text\":\"use ls to list the files in the current directory\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"role\":\"user\"}],\"model\":\"claude-sonnet-4-6\",\"system\":[{\"text\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\u003ccritical_rules\\u003e\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default \\u003c4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\u003c/critical_rules\\u003e\\n\\n\\u003ccommunication_style\\u003e\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\u003c/communication_style\\u003e\\n\\n\\u003ccode_references\\u003e\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\u003c/code_references\\u003e\\n\\n\\u003cworkflow\\u003e\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\u003c/workflow\\u003e\\n\\n\\u003cdecision_making\\u003e\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\u003c/decision_making\\u003e\\n\\n\\u003cediting_files\\u003e\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\u003c/editing_files\\u003e\\n\\n\\u003cwhitespace_and_exact_matching\\u003e\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\u003c/whitespace_and_exact_matching\\u003e\\n\\n\\u003ctask_completion\\u003e\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\u003c/task_completion\\u003e\\n\\n\\u003cerror_handling\\u003e\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\u003c/error_handling\\u003e\\n\\n\\u003cmemory_instructions\\u003e\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\u003c/memory_instructions\\u003e\\n\\n\\u003ccode_conventions\\u003e\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\u003c/code_conventions\\u003e\\n\\n\\u003ctesting\\u003e\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\u003c/testing\\u003e\\n\\n\\u003ctool_usage\\u003e\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\u003cbash_commands\\u003e\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `\\u0026` for background processes that won't stop on their own (e.g., `node server.js \\u0026`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status \\u0026\\u0026 git diff HEAD \\u0026\\u0026 git log -n 3`)\\n\\u003c/bash_commands\\u003e\\n\\u003c/tool_usage\\u003e\\n\\n\\u003cproactiveness\\u003e\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\u003c/proactiveness\\u003e\\n\\n\\u003cfinal_answers\\u003e\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\u003c/final_answers\\u003e\\n\\n\\u003cenv\\u003e\\nWorking directory: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/ls_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\u003c/env\\u003e\\n\\n\\n\\n\\n\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"tool_choice\":{\"disable_parallel_tool_use\":false,\"type\":\"auto\"},\"tools\":[{\"input_schema\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"},\"name\":\"bash\",\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\u003ccross_platform\\u003e\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\u003c/cross_platform\\u003e\\n\\n\\u003cexecution_steps\\u003e\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with \\u003ccwd\\u003e\\u003c/cwd\\u003e tags\\n\\u003c/execution_steps\\u003e\\n\\n\\u003cusage_notes\\u003e\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '\\u0026\\u0026', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\u003c/usage_notes\\u003e\\n\\n\\u003cbackground_execution\\u003e\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `\\u0026` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\u003c/background_execution\\u003e\\n\\n\\u003cgit_commits\\u003e\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in \\u003ccommit_analysis\\u003e tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\u003ccrush@charm.land\\u003e\\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\u003c/git_commits\\u003e\\n\\n\\u003cpull_requests\\u003e\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in \\u003cpr_analysis\\u003e tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n\\n ## Summary\\n\\n \\u0026lt;1-3 bullet points\\u003e\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\u003c/pull_requests\\u003e\\n\\n\\u003cexamples\\u003e\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar \\u0026\\u0026 pytest tests\\n\\u003c/examples\\u003e\\n\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"},\"name\":\"download\",\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"},\"name\":\"edit\",\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\"},{\"input_schema\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"},\"name\":\"multiedit\",\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\"},{\"input_schema\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"},\"name\":\"fetch\",\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\"},{\"input_schema\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"glob\",\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\"},{\"input_schema\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"grep\",\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\"},{\"input_schema\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"},\"name\":\"ls\",\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\"},{\"input_schema\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"},\"name\":\"sourcegraph\",\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"},\"name\":\"view\",\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\"},{\"input_schema\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"},\"name\":\"write\",\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-sonnet-4-6","id":"msg_01B7zih181AcTZUE9L88Pmyq","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":3,"cache_creation_input_tokens":4543,"cache_read_input_tokens":3733,"cache_creation":{"ephemeral_5m_input_tokens":4543,"ephemeral_1h_input_tokens":0},"output_tokens":35,"service_tier":"standard","inference_geo":"global"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"tool_use","id":"toolu_012Xiw3b8UsWffbNPXcvBoPs","name":"ls","input":{},"caller":{"type":"direct"}} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":""} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"tool_use","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":3,"cache_creation_input_tokens":4543,"cache_read_input_tokens":3733,"output_tokens":35} }
-
- event: message_stop
- data: {"type":"message_stop"}
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 1.448616959s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31333
- host: ""
- body: "{\"max_tokens\":10000,\"messages\":[{\"content\":[{\"text\":\"\\u003csystem_reminder\\u003eThis is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\\u003c/system_reminder\\u003e\",\"type\":\"text\"},{\"text\":\"use ls to list the files in the current directory\",\"type\":\"text\"}],\"role\":\"user\"},{\"content\":[{\"id\":\"toolu_012Xiw3b8UsWffbNPXcvBoPs\",\"input\":{},\"name\":\"ls\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"tool_use\"}],\"role\":\"assistant\"},{\"content\":[{\"tool_use_id\":\"toolu_012Xiw3b8UsWffbNPXcvBoPs\",\"cache_control\":{\"type\":\"ephemeral\"},\"content\":[{\"text\":\"\\n- /tmp/crush-test/TestCoderAgent/anthropic-sonnet/ls_tool/\\n - go.mod\\n - main.go\\n\",\"type\":\"text\"}],\"type\":\"tool_result\"}],\"role\":\"user\"}],\"model\":\"claude-sonnet-4-6\",\"system\":[{\"text\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\u003ccritical_rules\\u003e\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default \\u003c4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\u003c/critical_rules\\u003e\\n\\n\\u003ccommunication_style\\u003e\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\u003c/communication_style\\u003e\\n\\n\\u003ccode_references\\u003e\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\u003c/code_references\\u003e\\n\\n\\u003cworkflow\\u003e\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\u003c/workflow\\u003e\\n\\n\\u003cdecision_making\\u003e\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\u003c/decision_making\\u003e\\n\\n\\u003cediting_files\\u003e\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\u003c/editing_files\\u003e\\n\\n\\u003cwhitespace_and_exact_matching\\u003e\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\u003c/whitespace_and_exact_matching\\u003e\\n\\n\\u003ctask_completion\\u003e\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\u003c/task_completion\\u003e\\n\\n\\u003cerror_handling\\u003e\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\u003c/error_handling\\u003e\\n\\n\\u003cmemory_instructions\\u003e\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\u003c/memory_instructions\\u003e\\n\\n\\u003ccode_conventions\\u003e\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\u003c/code_conventions\\u003e\\n\\n\\u003ctesting\\u003e\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\u003c/testing\\u003e\\n\\n\\u003ctool_usage\\u003e\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\u003cbash_commands\\u003e\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `\\u0026` for background processes that won't stop on their own (e.g., `node server.js \\u0026`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status \\u0026\\u0026 git diff HEAD \\u0026\\u0026 git log -n 3`)\\n\\u003c/bash_commands\\u003e\\n\\u003c/tool_usage\\u003e\\n\\n\\u003cproactiveness\\u003e\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\u003c/proactiveness\\u003e\\n\\n\\u003cfinal_answers\\u003e\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\u003c/final_answers\\u003e\\n\\n\\u003cenv\\u003e\\nWorking directory: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/ls_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\u003c/env\\u003e\\n\\n\\n\\n\\n\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"tool_choice\":{\"disable_parallel_tool_use\":false,\"type\":\"auto\"},\"tools\":[{\"input_schema\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"},\"name\":\"bash\",\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\u003ccross_platform\\u003e\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\u003c/cross_platform\\u003e\\n\\n\\u003cexecution_steps\\u003e\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with \\u003ccwd\\u003e\\u003c/cwd\\u003e tags\\n\\u003c/execution_steps\\u003e\\n\\n\\u003cusage_notes\\u003e\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '\\u0026\\u0026', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\u003c/usage_notes\\u003e\\n\\n\\u003cbackground_execution\\u003e\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `\\u0026` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\u003c/background_execution\\u003e\\n\\n\\u003cgit_commits\\u003e\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in \\u003ccommit_analysis\\u003e tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\u003ccrush@charm.land\\u003e\\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\u003c/git_commits\\u003e\\n\\n\\u003cpull_requests\\u003e\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in \\u003cpr_analysis\\u003e tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n\\n ## Summary\\n\\n \\u0026lt;1-3 bullet points\\u003e\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\u003c/pull_requests\\u003e\\n\\n\\u003cexamples\\u003e\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar \\u0026\\u0026 pytest tests\\n\\u003c/examples\\u003e\\n\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"},\"name\":\"download\",\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"},\"name\":\"edit\",\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\"},{\"input_schema\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"},\"name\":\"multiedit\",\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\"},{\"input_schema\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"},\"name\":\"fetch\",\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\"},{\"input_schema\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"glob\",\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\"},{\"input_schema\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"grep\",\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\"},{\"input_schema\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"},\"name\":\"ls\",\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\"},{\"input_schema\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"},\"name\":\"sourcegraph\",\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"},\"name\":\"view\",\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\"},{\"input_schema\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"},\"name\":\"write\",\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-sonnet-4-6","id":"msg_01ETM6drrkX7jwMNodMNYFBR","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":1,"cache_creation_input_tokens":4629,"cache_read_input_tokens":3733,"cache_creation":{"ephemeral_5m_input_tokens":4629,"ephemeral_1h_input_tokens":0},"output_tokens":1,"service_tier":"standard","inference_geo":"global"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"`"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"go.mod`, `main.go`"} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":1,"cache_creation_input_tokens":4629,"cache_read_input_tokens":3733,"output_tokens":13} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 806.586166ms
diff --git a/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/multiedit_tool.yaml b/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/multiedit_tool.yaml
deleted file mode 100644
index fc0901adb800549ccb1eabed5b960d75fd6cf66d..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/multiedit_tool.yaml
+++ /dev/null
@@ -1,346 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 920
- host: ""
- body: '{"max_tokens":40,"messages":[{"content":[{"text":"Generate a concise title for the following content:\n\nuse multiedit to change ''Hello, World!'' to ''Hello, Crush!'' and add a comment ''// Greeting'' above the fmt.Println line in main.go\n \u003cthink\u003e\n\n\u003c/think\u003e","type":"text"}],"role":"user"}],"model":"claude-haiku-4-5-20251001","system":[{"text":"You will generate a short title based on the first message a user begins a conversation with.\n\n\u003crules\u003e\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\u003c/rules\u003e\n\n /no_think","type":"text"}],"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-haiku-4-5-20251001","id":"msg_01RKM1qKvmwGiYc4aTLUrgpd","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":195,"cache_creation_input_tokens":0,"cache_read_input_tokens":0,"cache_creation":{"ephemeral_5m_input_tokens":0,"ephemeral_1h_input_tokens":0},"output_tokens":5,"service_tier":"standard","inference_geo":"not_available"}}}
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"Multiedit Hello"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" World to Hello Crush with comment"} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":195,"cache_creation_input_tokens":0,"cache_read_input_tokens":0,"output_tokens":15} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 642.710416ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31082
- host: ""
- body: "{\"max_tokens\":10000,\"messages\":[{\"content\":[{\"text\":\"\\u003csystem_reminder\\u003eThis is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\\u003c/system_reminder\\u003e\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"},{\"text\":\"use multiedit to change 'Hello, World!' to 'Hello, Crush!' and add a comment '// Greeting' above the fmt.Println line in main.go\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"role\":\"user\"}],\"model\":\"claude-sonnet-4-6\",\"system\":[{\"text\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\u003ccritical_rules\\u003e\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default \\u003c4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\u003c/critical_rules\\u003e\\n\\n\\u003ccommunication_style\\u003e\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\u003c/communication_style\\u003e\\n\\n\\u003ccode_references\\u003e\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\u003c/code_references\\u003e\\n\\n\\u003cworkflow\\u003e\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\u003c/workflow\\u003e\\n\\n\\u003cdecision_making\\u003e\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\u003c/decision_making\\u003e\\n\\n\\u003cediting_files\\u003e\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\u003c/editing_files\\u003e\\n\\n\\u003cwhitespace_and_exact_matching\\u003e\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\u003c/whitespace_and_exact_matching\\u003e\\n\\n\\u003ctask_completion\\u003e\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\u003c/task_completion\\u003e\\n\\n\\u003cerror_handling\\u003e\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\u003c/error_handling\\u003e\\n\\n\\u003cmemory_instructions\\u003e\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\u003c/memory_instructions\\u003e\\n\\n\\u003ccode_conventions\\u003e\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\u003c/code_conventions\\u003e\\n\\n\\u003ctesting\\u003e\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\u003c/testing\\u003e\\n\\n\\u003ctool_usage\\u003e\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\u003cbash_commands\\u003e\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `\\u0026` for background processes that won't stop on their own (e.g., `node server.js \\u0026`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status \\u0026\\u0026 git diff HEAD \\u0026\\u0026 git log -n 3`)\\n\\u003c/bash_commands\\u003e\\n\\u003c/tool_usage\\u003e\\n\\n\\u003cproactiveness\\u003e\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\u003c/proactiveness\\u003e\\n\\n\\u003cfinal_answers\\u003e\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\u003c/final_answers\\u003e\\n\\n\\u003cenv\\u003e\\nWorking directory: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/multiedit_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\u003c/env\\u003e\\n\\n\\n\\n\\n\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"tool_choice\":{\"disable_parallel_tool_use\":false,\"type\":\"auto\"},\"tools\":[{\"input_schema\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"},\"name\":\"bash\",\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\u003ccross_platform\\u003e\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\u003c/cross_platform\\u003e\\n\\n\\u003cexecution_steps\\u003e\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with \\u003ccwd\\u003e\\u003c/cwd\\u003e tags\\n\\u003c/execution_steps\\u003e\\n\\n\\u003cusage_notes\\u003e\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '\\u0026\\u0026', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\u003c/usage_notes\\u003e\\n\\n\\u003cbackground_execution\\u003e\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `\\u0026` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\u003c/background_execution\\u003e\\n\\n\\u003cgit_commits\\u003e\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in \\u003ccommit_analysis\\u003e tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\u003ccrush@charm.land\\u003e\\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\u003c/git_commits\\u003e\\n\\n\\u003cpull_requests\\u003e\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in \\u003cpr_analysis\\u003e tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n\\n ## Summary\\n\\n \\u0026lt;1-3 bullet points\\u003e\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\u003c/pull_requests\\u003e\\n\\n\\u003cexamples\\u003e\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar \\u0026\\u0026 pytest tests\\n\\u003c/examples\\u003e\\n\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"},\"name\":\"download\",\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"},\"name\":\"edit\",\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\"},{\"input_schema\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"},\"name\":\"multiedit\",\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\"},{\"input_schema\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"},\"name\":\"fetch\",\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\"},{\"input_schema\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"glob\",\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\"},{\"input_schema\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"grep\",\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\"},{\"input_schema\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"},\"name\":\"ls\",\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\"},{\"input_schema\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"},\"name\":\"sourcegraph\",\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"},\"name\":\"view\",\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\"},{\"input_schema\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"},\"name\":\"write\",\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-sonnet-4-6","id":"msg_01TrcV3LWtSfbqBzfRLzdWip","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":3,"cache_creation_input_tokens":4575,"cache_read_input_tokens":3733,"cache_creation":{"ephemeral_5m_input_tokens":4575,"ephemeral_1h_input_tokens":0},"output_tokens":54,"service_tier":"standard","inference_geo":"global"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"tool_use","id":"toolu_01RsJ95C6xpH8RzXbzcrpMu4","name":"view","input":{},"caller":{"type":"direct"}} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"{\"file_"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"path\""}}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":": \"/tmp/cr"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"ush-test/T"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"est"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"CoderAgen"}}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"t/anthropi"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"c-son"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"net/mu"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"lti"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"edit_tool/ma"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"in."} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"go\"}"} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"tool_use","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":3,"cache_creation_input_tokens":4575,"cache_read_input_tokens":3733,"output_tokens":80} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 1.362773167s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31590
- host: ""
- body: "{\"max_tokens\":10000,\"messages\":[{\"content\":[{\"text\":\"\\u003csystem_reminder\\u003eThis is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\\u003c/system_reminder\\u003e\",\"type\":\"text\"},{\"text\":\"use multiedit to change 'Hello, World!' to 'Hello, Crush!' and add a comment '// Greeting' above the fmt.Println line in main.go\",\"type\":\"text\"}],\"role\":\"user\"},{\"content\":[{\"id\":\"toolu_01RsJ95C6xpH8RzXbzcrpMu4\",\"input\":{\"file_path\":\"/tmp/crush-test/TestCoderAgent/anthropic-sonnet/multiedit_tool/main.go\"},\"name\":\"view\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"tool_use\"}],\"role\":\"assistant\"},{\"content\":[{\"tool_use_id\":\"toolu_01RsJ95C6xpH8RzXbzcrpMu4\",\"cache_control\":{\"type\":\"ephemeral\"},\"content\":[{\"text\":\"\\u003cfile\\u003e\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\tfmt.Println(\\\"Hello, World!\\\")\\n 7|}\\n\\u003c/file\\u003e\\n\",\"type\":\"text\"}],\"type\":\"tool_result\"}],\"role\":\"user\"}],\"model\":\"claude-sonnet-4-6\",\"system\":[{\"text\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\u003ccritical_rules\\u003e\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default \\u003c4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\u003c/critical_rules\\u003e\\n\\n\\u003ccommunication_style\\u003e\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\u003c/communication_style\\u003e\\n\\n\\u003ccode_references\\u003e\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\u003c/code_references\\u003e\\n\\n\\u003cworkflow\\u003e\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\u003c/workflow\\u003e\\n\\n\\u003cdecision_making\\u003e\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\u003c/decision_making\\u003e\\n\\n\\u003cediting_files\\u003e\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\u003c/editing_files\\u003e\\n\\n\\u003cwhitespace_and_exact_matching\\u003e\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\u003c/whitespace_and_exact_matching\\u003e\\n\\n\\u003ctask_completion\\u003e\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\u003c/task_completion\\u003e\\n\\n\\u003cerror_handling\\u003e\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\u003c/error_handling\\u003e\\n\\n\\u003cmemory_instructions\\u003e\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\u003c/memory_instructions\\u003e\\n\\n\\u003ccode_conventions\\u003e\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\u003c/code_conventions\\u003e\\n\\n\\u003ctesting\\u003e\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\u003c/testing\\u003e\\n\\n\\u003ctool_usage\\u003e\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\u003cbash_commands\\u003e\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `\\u0026` for background processes that won't stop on their own (e.g., `node server.js \\u0026`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status \\u0026\\u0026 git diff HEAD \\u0026\\u0026 git log -n 3`)\\n\\u003c/bash_commands\\u003e\\n\\u003c/tool_usage\\u003e\\n\\n\\u003cproactiveness\\u003e\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\u003c/proactiveness\\u003e\\n\\n\\u003cfinal_answers\\u003e\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\u003c/final_answers\\u003e\\n\\n\\u003cenv\\u003e\\nWorking directory: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/multiedit_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\u003c/env\\u003e\\n\\n\\n\\n\\n\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"tool_choice\":{\"disable_parallel_tool_use\":false,\"type\":\"auto\"},\"tools\":[{\"input_schema\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"},\"name\":\"bash\",\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\u003ccross_platform\\u003e\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\u003c/cross_platform\\u003e\\n\\n\\u003cexecution_steps\\u003e\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with \\u003ccwd\\u003e\\u003c/cwd\\u003e tags\\n\\u003c/execution_steps\\u003e\\n\\n\\u003cusage_notes\\u003e\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '\\u0026\\u0026', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\u003c/usage_notes\\u003e\\n\\n\\u003cbackground_execution\\u003e\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `\\u0026` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\u003c/background_execution\\u003e\\n\\n\\u003cgit_commits\\u003e\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in \\u003ccommit_analysis\\u003e tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\u003ccrush@charm.land\\u003e\\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\u003c/git_commits\\u003e\\n\\n\\u003cpull_requests\\u003e\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in \\u003cpr_analysis\\u003e tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n\\n ## Summary\\n\\n \\u0026lt;1-3 bullet points\\u003e\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\u003c/pull_requests\\u003e\\n\\n\\u003cexamples\\u003e\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar \\u0026\\u0026 pytest tests\\n\\u003c/examples\\u003e\\n\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"},\"name\":\"download\",\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"},\"name\":\"edit\",\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\"},{\"input_schema\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"},\"name\":\"multiedit\",\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\"},{\"input_schema\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"},\"name\":\"fetch\",\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\"},{\"input_schema\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"glob\",\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\"},{\"input_schema\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"grep\",\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\"},{\"input_schema\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"},\"name\":\"ls\",\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\"},{\"input_schema\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"},\"name\":\"sourcegraph\",\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"},\"name\":\"view\",\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\"},{\"input_schema\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"},\"name\":\"write\",\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-sonnet-4-6","id":"msg_01CPRPJ7bFThq2BTkVzMT25x","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":1,"cache_creation_input_tokens":149,"cache_read_input_tokens":8308,"cache_creation":{"ephemeral_5m_input_tokens":149,"ephemeral_1h_input_tokens":0},"output_tokens":57,"service_tier":"standard","inference_geo":"global"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"tool_use","id":"toolu_01Xe9Mwv1Pkqn26MehAPTikm","name":"multiedit","input":{},"caller":{"type":"direct"}} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"{\""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"file_path\""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":": \"/tmp"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"/crush"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"-te"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"st/"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"TestC"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"oderAgen"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"t/anthr"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"opic"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"-sonnet/mult"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"iedit_too"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"l/m"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"ain.go\""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":", \"edits\":"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":" [{\"old_str"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"ing\":\"\\tfmt."} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"Println(\\\"H"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"el"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"lo,"}}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":" World!"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"\\\")\","} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"\"new_"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"string\""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":":\""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"\\t// Gree"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"ting\\"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"n\\tfmt."} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"Println"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"(\\\"Hello, C"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"ru"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"sh!\\\""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":")\"}]}"} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"tool_use","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":1,"cache_creation_input_tokens":149,"cache_read_input_tokens":8308,"output_tokens":157} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 1.654217458s
-- id: 3
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 32192
- host: ""
- body: "{\"max_tokens\":10000,\"messages\":[{\"content\":[{\"text\":\"\\u003csystem_reminder\\u003eThis is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\\u003c/system_reminder\\u003e\",\"type\":\"text\"},{\"text\":\"use multiedit to change 'Hello, World!' to 'Hello, Crush!' and add a comment '// Greeting' above the fmt.Println line in main.go\",\"type\":\"text\"}],\"role\":\"user\"},{\"content\":[{\"id\":\"toolu_01RsJ95C6xpH8RzXbzcrpMu4\",\"input\":{\"file_path\":\"/tmp/crush-test/TestCoderAgent/anthropic-sonnet/multiedit_tool/main.go\"},\"name\":\"view\",\"type\":\"tool_use\"}],\"role\":\"assistant\"},{\"content\":[{\"tool_use_id\":\"toolu_01RsJ95C6xpH8RzXbzcrpMu4\",\"content\":[{\"text\":\"\\u003cfile\\u003e\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\tfmt.Println(\\\"Hello, World!\\\")\\n 7|}\\n\\u003c/file\\u003e\\n\",\"type\":\"text\"}],\"type\":\"tool_result\"}],\"role\":\"user\"},{\"content\":[{\"id\":\"toolu_01Xe9Mwv1Pkqn26MehAPTikm\",\"input\":{\"edits\":[{\"new_string\":\"\\t// Greeting\\n\\tfmt.Println(\\\"Hello, Crush!\\\")\",\"old_string\":\"\\tfmt.Println(\\\"Hello, World!\\\")\"}],\"file_path\":\"/tmp/crush-test/TestCoderAgent/anthropic-sonnet/multiedit_tool/main.go\"},\"name\":\"multiedit\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"tool_use\"}],\"role\":\"assistant\"},{\"content\":[{\"tool_use_id\":\"toolu_01Xe9Mwv1Pkqn26MehAPTikm\",\"cache_control\":{\"type\":\"ephemeral\"},\"content\":[{\"text\":\"\\u003cresult\\u003e\\nApplied 1 edits to file: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/multiedit_tool/main.go\\n\\u003c/result\\u003e\\n\",\"type\":\"text\"}],\"type\":\"tool_result\"}],\"role\":\"user\"}],\"model\":\"claude-sonnet-4-6\",\"system\":[{\"text\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\u003ccritical_rules\\u003e\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default \\u003c4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\u003c/critical_rules\\u003e\\n\\n\\u003ccommunication_style\\u003e\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\u003c/communication_style\\u003e\\n\\n\\u003ccode_references\\u003e\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\u003c/code_references\\u003e\\n\\n\\u003cworkflow\\u003e\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\u003c/workflow\\u003e\\n\\n\\u003cdecision_making\\u003e\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\u003c/decision_making\\u003e\\n\\n\\u003cediting_files\\u003e\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\u003c/editing_files\\u003e\\n\\n\\u003cwhitespace_and_exact_matching\\u003e\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\u003c/whitespace_and_exact_matching\\u003e\\n\\n\\u003ctask_completion\\u003e\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\u003c/task_completion\\u003e\\n\\n\\u003cerror_handling\\u003e\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\u003c/error_handling\\u003e\\n\\n\\u003cmemory_instructions\\u003e\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\u003c/memory_instructions\\u003e\\n\\n\\u003ccode_conventions\\u003e\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\u003c/code_conventions\\u003e\\n\\n\\u003ctesting\\u003e\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\u003c/testing\\u003e\\n\\n\\u003ctool_usage\\u003e\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\u003cbash_commands\\u003e\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `\\u0026` for background processes that won't stop on their own (e.g., `node server.js \\u0026`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status \\u0026\\u0026 git diff HEAD \\u0026\\u0026 git log -n 3`)\\n\\u003c/bash_commands\\u003e\\n\\u003c/tool_usage\\u003e\\n\\n\\u003cproactiveness\\u003e\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\u003c/proactiveness\\u003e\\n\\n\\u003cfinal_answers\\u003e\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\u003c/final_answers\\u003e\\n\\n\\u003cenv\\u003e\\nWorking directory: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/multiedit_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\u003c/env\\u003e\\n\\n\\n\\n\\n\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"tool_choice\":{\"disable_parallel_tool_use\":false,\"type\":\"auto\"},\"tools\":[{\"input_schema\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"},\"name\":\"bash\",\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\u003ccross_platform\\u003e\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\u003c/cross_platform\\u003e\\n\\n\\u003cexecution_steps\\u003e\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with \\u003ccwd\\u003e\\u003c/cwd\\u003e tags\\n\\u003c/execution_steps\\u003e\\n\\n\\u003cusage_notes\\u003e\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '\\u0026\\u0026', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\u003c/usage_notes\\u003e\\n\\n\\u003cbackground_execution\\u003e\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `\\u0026` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\u003c/background_execution\\u003e\\n\\n\\u003cgit_commits\\u003e\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in \\u003ccommit_analysis\\u003e tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\u003ccrush@charm.land\\u003e\\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\u003c/git_commits\\u003e\\n\\n\\u003cpull_requests\\u003e\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in \\u003cpr_analysis\\u003e tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n\\n ## Summary\\n\\n \\u0026lt;1-3 bullet points\\u003e\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\u003c/pull_requests\\u003e\\n\\n\\u003cexamples\\u003e\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar \\u0026\\u0026 pytest tests\\n\\u003c/examples\\u003e\\n\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"},\"name\":\"download\",\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"},\"name\":\"edit\",\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\"},{\"input_schema\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"},\"name\":\"multiedit\",\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\"},{\"input_schema\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"},\"name\":\"fetch\",\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\"},{\"input_schema\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"glob\",\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\"},{\"input_schema\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"grep\",\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\"},{\"input_schema\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"},\"name\":\"ls\",\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\"},{\"input_schema\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"},\"name\":\"sourcegraph\",\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"},\"name\":\"view\",\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\"},{\"input_schema\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"},\"name\":\"write\",\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-sonnet-4-6","id":"msg_016wdCpaYaWaniVPFHBzaxEg","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":1,"cache_creation_input_tokens":351,"cache_read_input_tokens":8308,"cache_creation":{"ephemeral_5m_input_tokens":351,"ephemeral_1h_input_tokens":0},"output_tokens":2,"service_tier":"standard","inference_geo":"global"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"Done."} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":1,"cache_creation_input_tokens":351,"cache_read_input_tokens":8308,"output_tokens":5} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 1.380185292s
diff --git a/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/parallel_tool_calls.yaml b/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/parallel_tool_calls.yaml
deleted file mode 100644
index a8ad86056e05d574845b2d168a073355a60875c4..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/parallel_tool_calls.yaml
+++ /dev/null
@@ -1,216 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 926
- host: ""
- body: '{"max_tokens":40,"messages":[{"content":[{"text":"Generate a concise title for the following content:\n\nuse glob to find all .go files and use ls to list the current directory, it is very important that you run both tool calls in parallel\n \u003cthink\u003e\n\n\u003c/think\u003e","type":"text"}],"role":"user"}],"model":"claude-haiku-4-5-20251001","system":[{"text":"You will generate a short title based on the first message a user begins a conversation with.\n\n\u003crules\u003e\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\u003c/rules\u003e\n\n /no_think","type":"text"}],"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-haiku-4-5-20251001","id":"msg_01HKqZKuLTTyE6hXkLYkJCug","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":184,"cache_creation_input_tokens":0,"cache_read_input_tokens":0,"cache_creation":{"ephemeral_5m_input_tokens":0,"ephemeral_1h_input_tokens":0},"output_tokens":1,"service_tier":"standard","inference_geo":"not_available"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"Finding"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" Go files and listing directory in parallel"}}
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":184,"cache_creation_input_tokens":0,"cache_read_input_tokens":0,"output_tokens":11} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 644.135083ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31093
- host: ""
- body: "{\"max_tokens\":10000,\"messages\":[{\"content\":[{\"text\":\"\\u003csystem_reminder\\u003eThis is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\\u003c/system_reminder\\u003e\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"},{\"text\":\"use glob to find all .go files and use ls to list the current directory, it is very important that you run both tool calls in parallel\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"role\":\"user\"}],\"model\":\"claude-sonnet-4-6\",\"system\":[{\"text\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\u003ccritical_rules\\u003e\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default \\u003c4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\u003c/critical_rules\\u003e\\n\\n\\u003ccommunication_style\\u003e\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\u003c/communication_style\\u003e\\n\\n\\u003ccode_references\\u003e\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\u003c/code_references\\u003e\\n\\n\\u003cworkflow\\u003e\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\u003c/workflow\\u003e\\n\\n\\u003cdecision_making\\u003e\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\u003c/decision_making\\u003e\\n\\n\\u003cediting_files\\u003e\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\u003c/editing_files\\u003e\\n\\n\\u003cwhitespace_and_exact_matching\\u003e\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\u003c/whitespace_and_exact_matching\\u003e\\n\\n\\u003ctask_completion\\u003e\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\u003c/task_completion\\u003e\\n\\n\\u003cerror_handling\\u003e\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\u003c/error_handling\\u003e\\n\\n\\u003cmemory_instructions\\u003e\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\u003c/memory_instructions\\u003e\\n\\n\\u003ccode_conventions\\u003e\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\u003c/code_conventions\\u003e\\n\\n\\u003ctesting\\u003e\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\u003c/testing\\u003e\\n\\n\\u003ctool_usage\\u003e\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\u003cbash_commands\\u003e\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `\\u0026` for background processes that won't stop on their own (e.g., `node server.js \\u0026`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status \\u0026\\u0026 git diff HEAD \\u0026\\u0026 git log -n 3`)\\n\\u003c/bash_commands\\u003e\\n\\u003c/tool_usage\\u003e\\n\\n\\u003cproactiveness\\u003e\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\u003c/proactiveness\\u003e\\n\\n\\u003cfinal_answers\\u003e\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\u003c/final_answers\\u003e\\n\\n\\u003cenv\\u003e\\nWorking directory: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/parallel_tool_calls\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\u003c/env\\u003e\\n\\n\\n\\n\\n\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"tool_choice\":{\"disable_parallel_tool_use\":false,\"type\":\"auto\"},\"tools\":[{\"input_schema\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"},\"name\":\"bash\",\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\u003ccross_platform\\u003e\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\u003c/cross_platform\\u003e\\n\\n\\u003cexecution_steps\\u003e\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with \\u003ccwd\\u003e\\u003c/cwd\\u003e tags\\n\\u003c/execution_steps\\u003e\\n\\n\\u003cusage_notes\\u003e\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '\\u0026\\u0026', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\u003c/usage_notes\\u003e\\n\\n\\u003cbackground_execution\\u003e\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `\\u0026` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\u003c/background_execution\\u003e\\n\\n\\u003cgit_commits\\u003e\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in \\u003ccommit_analysis\\u003e tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\u003ccrush@charm.land\\u003e\\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\u003c/git_commits\\u003e\\n\\n\\u003cpull_requests\\u003e\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in \\u003cpr_analysis\\u003e tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n\\n ## Summary\\n\\n \\u0026lt;1-3 bullet points\\u003e\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\u003c/pull_requests\\u003e\\n\\n\\u003cexamples\\u003e\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar \\u0026\\u0026 pytest tests\\n\\u003c/examples\\u003e\\n\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"},\"name\":\"download\",\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"},\"name\":\"edit\",\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\"},{\"input_schema\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"},\"name\":\"multiedit\",\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\"},{\"input_schema\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"},\"name\":\"fetch\",\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\"},{\"input_schema\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"glob\",\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\"},{\"input_schema\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"grep\",\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\"},{\"input_schema\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"},\"name\":\"ls\",\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\"},{\"input_schema\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"},\"name\":\"sourcegraph\",\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"},\"name\":\"view\",\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\"},{\"input_schema\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"},\"name\":\"write\",\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-sonnet-4-6","id":"msg_019ZytTuTmQsvhM8mPUc8dKt","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":3,"cache_creation_input_tokens":4564,"cache_read_input_tokens":3733,"cache_creation":{"ephemeral_5m_input_tokens":4564,"ephemeral_1h_input_tokens":0},"output_tokens":62,"service_tier":"standard","inference_geo":"global"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"tool_use","id":"toolu_01QYv5XJsq3fV1NnHGH5qE1j","name":"glob","input":{},"caller":{"type":"direct"}} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"{\"patte"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"rn\""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":": \"**"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"/*.g"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"o\"}"} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: content_block_start
- data: {"type":"content_block_start","index":1,"content_block":{"type":"tool_use","id":"toolu_01RZ13pSVWhjBKJJQfzAcEmP","name":"ls","input":{},"caller":{"type":"direct"}} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":1,"delta":{"type":"input_json_delta","partial_json":""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":1,"delta":{"type":"input_json_delta","partial_json":"{\"path\": \"/"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":1,"delta":{"type":"input_json_delta","partial_json":"tmp/"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":1,"delta":{"type":"input_json_delta","partial_json":"crush"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":1,"delta":{"type":"input_json_delta","partial_json":"-test/Tes"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":1,"delta":{"type":"input_json_delta","partial_json":"tCoderAge"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":1,"delta":{"type":"input_json_delta","partial_json":"nt/anthrop"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":1,"delta":{"type":"input_json_delta","partial_json":"ic-son"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":1,"delta":{"type":"input_json_delta","partial_json":"net/paralle"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":1,"delta":{"type":"input_json_delta","partial_json":"l_tool_cal"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":1,"delta":{"type":"input_json_delta","partial_json":"ls\"}"} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":1 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"tool_use","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":3,"cache_creation_input_tokens":4564,"cache_read_input_tokens":3733,"output_tokens":110} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 1.778554625s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31803
- host: ""
- body: "{\"max_tokens\":10000,\"messages\":[{\"content\":[{\"text\":\"\\u003csystem_reminder\\u003eThis is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\\u003c/system_reminder\\u003e\",\"type\":\"text\"},{\"text\":\"use glob to find all .go files and use ls to list the current directory, it is very important that you run both tool calls in parallel\",\"type\":\"text\"}],\"role\":\"user\"},{\"content\":[{\"id\":\"toolu_01QYv5XJsq3fV1NnHGH5qE1j\",\"input\":{\"pattern\":\"**/*.go\"},\"name\":\"glob\",\"type\":\"tool_use\"},{\"id\":\"toolu_01RZ13pSVWhjBKJJQfzAcEmP\",\"input\":{\"path\":\"/tmp/crush-test/TestCoderAgent/anthropic-sonnet/parallel_tool_calls\"},\"name\":\"ls\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"tool_use\"}],\"role\":\"assistant\"},{\"content\":[{\"tool_use_id\":\"toolu_01QYv5XJsq3fV1NnHGH5qE1j\",\"content\":[{\"text\":\"/tmp/crush-test/TestCoderAgent/anthropic-sonnet/parallel_tool_calls/main.go\",\"type\":\"text\"}],\"type\":\"tool_result\"},{\"tool_use_id\":\"toolu_01RZ13pSVWhjBKJJQfzAcEmP\",\"cache_control\":{\"type\":\"ephemeral\"},\"content\":[{\"text\":\"\\n- /tmp/crush-test/TestCoderAgent/anthropic-sonnet/parallel_tool_calls/\\n - go.mod\\n - main.go\\n\",\"type\":\"text\"}],\"type\":\"tool_result\"}],\"role\":\"user\"}],\"model\":\"claude-sonnet-4-6\",\"system\":[{\"text\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\u003ccritical_rules\\u003e\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default \\u003c4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\u003c/critical_rules\\u003e\\n\\n\\u003ccommunication_style\\u003e\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\u003c/communication_style\\u003e\\n\\n\\u003ccode_references\\u003e\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\u003c/code_references\\u003e\\n\\n\\u003cworkflow\\u003e\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\u003c/workflow\\u003e\\n\\n\\u003cdecision_making\\u003e\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\u003c/decision_making\\u003e\\n\\n\\u003cediting_files\\u003e\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\u003c/editing_files\\u003e\\n\\n\\u003cwhitespace_and_exact_matching\\u003e\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\u003c/whitespace_and_exact_matching\\u003e\\n\\n\\u003ctask_completion\\u003e\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\u003c/task_completion\\u003e\\n\\n\\u003cerror_handling\\u003e\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\u003c/error_handling\\u003e\\n\\n\\u003cmemory_instructions\\u003e\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\u003c/memory_instructions\\u003e\\n\\n\\u003ccode_conventions\\u003e\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\u003c/code_conventions\\u003e\\n\\n\\u003ctesting\\u003e\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\u003c/testing\\u003e\\n\\n\\u003ctool_usage\\u003e\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\u003cbash_commands\\u003e\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `\\u0026` for background processes that won't stop on their own (e.g., `node server.js \\u0026`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status \\u0026\\u0026 git diff HEAD \\u0026\\u0026 git log -n 3`)\\n\\u003c/bash_commands\\u003e\\n\\u003c/tool_usage\\u003e\\n\\n\\u003cproactiveness\\u003e\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\u003c/proactiveness\\u003e\\n\\n\\u003cfinal_answers\\u003e\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\u003c/final_answers\\u003e\\n\\n\\u003cenv\\u003e\\nWorking directory: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/parallel_tool_calls\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\u003c/env\\u003e\\n\\n\\n\\n\\n\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"tool_choice\":{\"disable_parallel_tool_use\":false,\"type\":\"auto\"},\"tools\":[{\"input_schema\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"},\"name\":\"bash\",\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\u003ccross_platform\\u003e\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\u003c/cross_platform\\u003e\\n\\n\\u003cexecution_steps\\u003e\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with \\u003ccwd\\u003e\\u003c/cwd\\u003e tags\\n\\u003c/execution_steps\\u003e\\n\\n\\u003cusage_notes\\u003e\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '\\u0026\\u0026', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\u003c/usage_notes\\u003e\\n\\n\\u003cbackground_execution\\u003e\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `\\u0026` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\u003c/background_execution\\u003e\\n\\n\\u003cgit_commits\\u003e\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in \\u003ccommit_analysis\\u003e tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\u003ccrush@charm.land\\u003e\\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\u003c/git_commits\\u003e\\n\\n\\u003cpull_requests\\u003e\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in \\u003cpr_analysis\\u003e tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n\\n ## Summary\\n\\n \\u0026lt;1-3 bullet points\\u003e\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\u003c/pull_requests\\u003e\\n\\n\\u003cexamples\\u003e\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar \\u0026\\u0026 pytest tests\\n\\u003c/examples\\u003e\\n\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"},\"name\":\"download\",\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"},\"name\":\"edit\",\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\"},{\"input_schema\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"},\"name\":\"multiedit\",\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\"},{\"input_schema\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"},\"name\":\"fetch\",\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\"},{\"input_schema\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"glob\",\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\"},{\"input_schema\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"grep\",\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\"},{\"input_schema\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"},\"name\":\"ls\",\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\"},{\"input_schema\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"},\"name\":\"sourcegraph\",\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"},\"name\":\"view\",\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\"},{\"input_schema\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"},\"name\":\"write\",\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-sonnet-4-6","id":"msg_01D4xaMw2CyQ5LeiJT7RB792","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":1,"cache_creation_input_tokens":4800,"cache_read_input_tokens":3733,"cache_creation":{"ephemeral_5m_input_tokens":4800,"ephemeral_1h_input_tokens":0},"output_tokens":1,"service_tier":"standard","inference_geo":"global"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"**"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"Go files:** `main.go`\n\n**Directory contents:** `go.mod`, `main.go`"} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":1,"cache_creation_input_tokens":4800,"cache_read_input_tokens":3733,"output_tokens":27} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 1.255951083s
diff --git a/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/read_a_file.yaml b/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/read_a_file.yaml
deleted file mode 100644
index ad5bd60d714942af47f1ac837921d16c69789ede..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/read_a_file.yaml
+++ /dev/null
@@ -1,260 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 807
- host: ""
- body: '{"max_tokens":40,"messages":[{"content":[{"text":"Generate a concise title for the following content:\n\nRead the go mod\n \u003cthink\u003e\n\n\u003c/think\u003e","type":"text"}],"role":"user"}],"model":"claude-haiku-4-5-20251001","system":[{"text":"You will generate a short title based on the first message a user begins a conversation with.\n\n\u003crules\u003e\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\u003c/rules\u003e\n\n /no_think","type":"text"}],"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-haiku-4-5-20251001","id":"msg_01LhuXKez6yFTEEV385d1cux","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":159,"cache_creation_input_tokens":0,"cache_read_input_tokens":0,"cache_creation":{"ephemeral_5m_input_tokens":0,"ephemeral_1h_input_tokens":0},"output_tokens":1,"service_tier":"standard","inference_geo":"not_available"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"Go"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" mod file reading"} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":159,"cache_creation_input_tokens":0,"cache_read_input_tokens":0,"output_tokens":7} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 1.1756565s
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30966
- host: ""
- body: "{\"max_tokens\":10000,\"messages\":[{\"content\":[{\"text\":\"\\u003csystem_reminder\\u003eThis is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\\u003c/system_reminder\\u003e\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"},{\"text\":\"Read the go mod\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"role\":\"user\"}],\"model\":\"claude-sonnet-4-6\",\"system\":[{\"text\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\u003ccritical_rules\\u003e\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default \\u003c4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\u003c/critical_rules\\u003e\\n\\n\\u003ccommunication_style\\u003e\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\u003c/communication_style\\u003e\\n\\n\\u003ccode_references\\u003e\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\u003c/code_references\\u003e\\n\\n\\u003cworkflow\\u003e\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\u003c/workflow\\u003e\\n\\n\\u003cdecision_making\\u003e\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\u003c/decision_making\\u003e\\n\\n\\u003cediting_files\\u003e\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\u003c/editing_files\\u003e\\n\\n\\u003cwhitespace_and_exact_matching\\u003e\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\u003c/whitespace_and_exact_matching\\u003e\\n\\n\\u003ctask_completion\\u003e\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\u003c/task_completion\\u003e\\n\\n\\u003cerror_handling\\u003e\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\u003c/error_handling\\u003e\\n\\n\\u003cmemory_instructions\\u003e\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\u003c/memory_instructions\\u003e\\n\\n\\u003ccode_conventions\\u003e\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\u003c/code_conventions\\u003e\\n\\n\\u003ctesting\\u003e\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\u003c/testing\\u003e\\n\\n\\u003ctool_usage\\u003e\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\u003cbash_commands\\u003e\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `\\u0026` for background processes that won't stop on their own (e.g., `node server.js \\u0026`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status \\u0026\\u0026 git diff HEAD \\u0026\\u0026 git log -n 3`)\\n\\u003c/bash_commands\\u003e\\n\\u003c/tool_usage\\u003e\\n\\n\\u003cproactiveness\\u003e\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\u003c/proactiveness\\u003e\\n\\n\\u003cfinal_answers\\u003e\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\u003c/final_answers\\u003e\\n\\n\\u003cenv\\u003e\\nWorking directory: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/read_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\u003c/env\\u003e\\n\\n\\n\\n\\n\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"tool_choice\":{\"disable_parallel_tool_use\":false,\"type\":\"auto\"},\"tools\":[{\"input_schema\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"},\"name\":\"bash\",\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\u003ccross_platform\\u003e\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\u003c/cross_platform\\u003e\\n\\n\\u003cexecution_steps\\u003e\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with \\u003ccwd\\u003e\\u003c/cwd\\u003e tags\\n\\u003c/execution_steps\\u003e\\n\\n\\u003cusage_notes\\u003e\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '\\u0026\\u0026', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\u003c/usage_notes\\u003e\\n\\n\\u003cbackground_execution\\u003e\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `\\u0026` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\u003c/background_execution\\u003e\\n\\n\\u003cgit_commits\\u003e\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in \\u003ccommit_analysis\\u003e tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\u003ccrush@charm.land\\u003e\\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\u003c/git_commits\\u003e\\n\\n\\u003cpull_requests\\u003e\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in \\u003cpr_analysis\\u003e tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n\\n ## Summary\\n\\n \\u0026lt;1-3 bullet points\\u003e\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\u003c/pull_requests\\u003e\\n\\n\\u003cexamples\\u003e\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar \\u0026\\u0026 pytest tests\\n\\u003c/examples\\u003e\\n\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"},\"name\":\"download\",\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"},\"name\":\"edit\",\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\"},{\"input_schema\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"},\"name\":\"multiedit\",\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\"},{\"input_schema\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"},\"name\":\"fetch\",\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\"},{\"input_schema\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"glob\",\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\"},{\"input_schema\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"grep\",\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\"},{\"input_schema\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"},\"name\":\"ls\",\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\"},{\"input_schema\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"},\"name\":\"sourcegraph\",\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"},\"name\":\"view\",\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\"},{\"input_schema\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"},\"name\":\"write\",\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-sonnet-4-6","id":"msg_01Vwm9FpTV9G35n3sRU6FRxg","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":3,"cache_creation_input_tokens":4539,"cache_read_input_tokens":3733,"cache_creation":{"ephemeral_5m_input_tokens":4539,"ephemeral_1h_input_tokens":0},"output_tokens":54,"service_tier":"standard","inference_geo":"global"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"tool_use","id":"toolu_01Evrr81C7EXFWSF5d44wap4","name":"glob","input":{},"caller":{"type":"direct"}} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"{\"pa"}}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"ttern\": \"go"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":".mod\"}"} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"tool_use","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":3,"cache_creation_input_tokens":4539,"cache_read_input_tokens":3733,"output_tokens":54} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 1.907795291s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31302
- host: ""
- body: "{\"max_tokens\":10000,\"messages\":[{\"content\":[{\"text\":\"\\u003csystem_reminder\\u003eThis is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\\u003c/system_reminder\\u003e\",\"type\":\"text\"},{\"text\":\"Read the go mod\",\"type\":\"text\"}],\"role\":\"user\"},{\"content\":[{\"id\":\"toolu_01Evrr81C7EXFWSF5d44wap4\",\"input\":{\"pattern\":\"go.mod\"},\"name\":\"glob\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"tool_use\"}],\"role\":\"assistant\"},{\"content\":[{\"tool_use_id\":\"toolu_01Evrr81C7EXFWSF5d44wap4\",\"cache_control\":{\"type\":\"ephemeral\"},\"content\":[{\"text\":\"/tmp/crush-test/TestCoderAgent/anthropic-sonnet/read_a_file/go.mod\",\"type\":\"text\"}],\"type\":\"tool_result\"}],\"role\":\"user\"}],\"model\":\"claude-sonnet-4-6\",\"system\":[{\"text\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\u003ccritical_rules\\u003e\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default \\u003c4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\u003c/critical_rules\\u003e\\n\\n\\u003ccommunication_style\\u003e\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\u003c/communication_style\\u003e\\n\\n\\u003ccode_references\\u003e\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\u003c/code_references\\u003e\\n\\n\\u003cworkflow\\u003e\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\u003c/workflow\\u003e\\n\\n\\u003cdecision_making\\u003e\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\u003c/decision_making\\u003e\\n\\n\\u003cediting_files\\u003e\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\u003c/editing_files\\u003e\\n\\n\\u003cwhitespace_and_exact_matching\\u003e\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\u003c/whitespace_and_exact_matching\\u003e\\n\\n\\u003ctask_completion\\u003e\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\u003c/task_completion\\u003e\\n\\n\\u003cerror_handling\\u003e\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\u003c/error_handling\\u003e\\n\\n\\u003cmemory_instructions\\u003e\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\u003c/memory_instructions\\u003e\\n\\n\\u003ccode_conventions\\u003e\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\u003c/code_conventions\\u003e\\n\\n\\u003ctesting\\u003e\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\u003c/testing\\u003e\\n\\n\\u003ctool_usage\\u003e\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\u003cbash_commands\\u003e\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `\\u0026` for background processes that won't stop on their own (e.g., `node server.js \\u0026`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status \\u0026\\u0026 git diff HEAD \\u0026\\u0026 git log -n 3`)\\n\\u003c/bash_commands\\u003e\\n\\u003c/tool_usage\\u003e\\n\\n\\u003cproactiveness\\u003e\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\u003c/proactiveness\\u003e\\n\\n\\u003cfinal_answers\\u003e\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\u003c/final_answers\\u003e\\n\\n\\u003cenv\\u003e\\nWorking directory: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/read_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\u003c/env\\u003e\\n\\n\\n\\n\\n\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"tool_choice\":{\"disable_parallel_tool_use\":false,\"type\":\"auto\"},\"tools\":[{\"input_schema\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"},\"name\":\"bash\",\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\u003ccross_platform\\u003e\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\u003c/cross_platform\\u003e\\n\\n\\u003cexecution_steps\\u003e\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with \\u003ccwd\\u003e\\u003c/cwd\\u003e tags\\n\\u003c/execution_steps\\u003e\\n\\n\\u003cusage_notes\\u003e\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '\\u0026\\u0026', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\u003c/usage_notes\\u003e\\n\\n\\u003cbackground_execution\\u003e\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `\\u0026` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\u003c/background_execution\\u003e\\n\\n\\u003cgit_commits\\u003e\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in \\u003ccommit_analysis\\u003e tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\u003ccrush@charm.land\\u003e\\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\u003c/git_commits\\u003e\\n\\n\\u003cpull_requests\\u003e\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in \\u003cpr_analysis\\u003e tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n\\n ## Summary\\n\\n \\u0026lt;1-3 bullet points\\u003e\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\u003c/pull_requests\\u003e\\n\\n\\u003cexamples\\u003e\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar \\u0026\\u0026 pytest tests\\n\\u003c/examples\\u003e\\n\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"},\"name\":\"download\",\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"},\"name\":\"edit\",\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\"},{\"input_schema\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"},\"name\":\"multiedit\",\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\"},{\"input_schema\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"},\"name\":\"fetch\",\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\"},{\"input_schema\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"glob\",\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\"},{\"input_schema\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"grep\",\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\"},{\"input_schema\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"},\"name\":\"ls\",\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\"},{\"input_schema\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"},\"name\":\"sourcegraph\",\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"},\"name\":\"view\",\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\"},{\"input_schema\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"},\"name\":\"write\",\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-sonnet-4-6","id":"msg_0152uCnUSyQnUCoZZUFxcr32","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":1,"cache_creation_input_tokens":4633,"cache_read_input_tokens":3733,"cache_creation":{"ephemeral_5m_input_tokens":4633,"ephemeral_1h_input_tokens":0},"output_tokens":53,"service_tier":"standard","inference_geo":"global"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"tool_use","id":"toolu_01HxoxYPFDDn5V5wY4eJoe6q","name":"view","input":{},"caller":{"type":"direct"}} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"{\"file_"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"pa"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"th\": \"/"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"tm"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"p/crush-t"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"est/TestCo"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"de"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"rAgent/an"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"thropic"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"-sonnet/"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"read_"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"a_file/g"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"o.mod\"}"} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"tool_use","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":1,"cache_creation_input_tokens":4633,"cache_read_input_tokens":3733,"output_tokens":79} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 2.1693505s
-- id: 3
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31735
- host: ""
- body: "{\"max_tokens\":10000,\"messages\":[{\"content\":[{\"text\":\"\\u003csystem_reminder\\u003eThis is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\\u003c/system_reminder\\u003e\",\"type\":\"text\"},{\"text\":\"Read the go mod\",\"type\":\"text\"}],\"role\":\"user\"},{\"content\":[{\"id\":\"toolu_01Evrr81C7EXFWSF5d44wap4\",\"input\":{\"pattern\":\"go.mod\"},\"name\":\"glob\",\"type\":\"tool_use\"}],\"role\":\"assistant\"},{\"content\":[{\"tool_use_id\":\"toolu_01Evrr81C7EXFWSF5d44wap4\",\"content\":[{\"text\":\"/tmp/crush-test/TestCoderAgent/anthropic-sonnet/read_a_file/go.mod\",\"type\":\"text\"}],\"type\":\"tool_result\"}],\"role\":\"user\"},{\"content\":[{\"id\":\"toolu_01HxoxYPFDDn5V5wY4eJoe6q\",\"input\":{\"file_path\":\"/tmp/crush-test/TestCoderAgent/anthropic-sonnet/read_a_file/go.mod\"},\"name\":\"view\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"tool_use\"}],\"role\":\"assistant\"},{\"content\":[{\"tool_use_id\":\"toolu_01HxoxYPFDDn5V5wY4eJoe6q\",\"cache_control\":{\"type\":\"ephemeral\"},\"content\":[{\"text\":\"\\u003cfile\\u003e\\n 1|module example.com/testproject\\n 2|\\n 3|go 1.23\\n\\u003c/file\\u003e\\n\",\"type\":\"text\"}],\"type\":\"tool_result\"}],\"role\":\"user\"}],\"model\":\"claude-sonnet-4-6\",\"system\":[{\"text\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\u003ccritical_rules\\u003e\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default \\u003c4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\u003c/critical_rules\\u003e\\n\\n\\u003ccommunication_style\\u003e\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\u003c/communication_style\\u003e\\n\\n\\u003ccode_references\\u003e\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\u003c/code_references\\u003e\\n\\n\\u003cworkflow\\u003e\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\u003c/workflow\\u003e\\n\\n\\u003cdecision_making\\u003e\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\u003c/decision_making\\u003e\\n\\n\\u003cediting_files\\u003e\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\u003c/editing_files\\u003e\\n\\n\\u003cwhitespace_and_exact_matching\\u003e\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\u003c/whitespace_and_exact_matching\\u003e\\n\\n\\u003ctask_completion\\u003e\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\u003c/task_completion\\u003e\\n\\n\\u003cerror_handling\\u003e\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\u003c/error_handling\\u003e\\n\\n\\u003cmemory_instructions\\u003e\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\u003c/memory_instructions\\u003e\\n\\n\\u003ccode_conventions\\u003e\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\u003c/code_conventions\\u003e\\n\\n\\u003ctesting\\u003e\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\u003c/testing\\u003e\\n\\n\\u003ctool_usage\\u003e\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\u003cbash_commands\\u003e\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `\\u0026` for background processes that won't stop on their own (e.g., `node server.js \\u0026`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status \\u0026\\u0026 git diff HEAD \\u0026\\u0026 git log -n 3`)\\n\\u003c/bash_commands\\u003e\\n\\u003c/tool_usage\\u003e\\n\\n\\u003cproactiveness\\u003e\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\u003c/proactiveness\\u003e\\n\\n\\u003cfinal_answers\\u003e\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\u003c/final_answers\\u003e\\n\\n\\u003cenv\\u003e\\nWorking directory: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/read_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\u003c/env\\u003e\\n\\n\\n\\n\\n\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"tool_choice\":{\"disable_parallel_tool_use\":false,\"type\":\"auto\"},\"tools\":[{\"input_schema\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"},\"name\":\"bash\",\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\u003ccross_platform\\u003e\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\u003c/cross_platform\\u003e\\n\\n\\u003cexecution_steps\\u003e\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with \\u003ccwd\\u003e\\u003c/cwd\\u003e tags\\n\\u003c/execution_steps\\u003e\\n\\n\\u003cusage_notes\\u003e\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '\\u0026\\u0026', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\u003c/usage_notes\\u003e\\n\\n\\u003cbackground_execution\\u003e\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `\\u0026` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\u003c/background_execution\\u003e\\n\\n\\u003cgit_commits\\u003e\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in \\u003ccommit_analysis\\u003e tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\u003ccrush@charm.land\\u003e\\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\u003c/git_commits\\u003e\\n\\n\\u003cpull_requests\\u003e\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in \\u003cpr_analysis\\u003e tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n\\n ## Summary\\n\\n \\u0026lt;1-3 bullet points\\u003e\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\u003c/pull_requests\\u003e\\n\\n\\u003cexamples\\u003e\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar \\u0026\\u0026 pytest tests\\n\\u003c/examples\\u003e\\n\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"},\"name\":\"download\",\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"},\"name\":\"edit\",\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\"},{\"input_schema\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"},\"name\":\"multiedit\",\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\"},{\"input_schema\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"},\"name\":\"fetch\",\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\"},{\"input_schema\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"glob\",\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\"},{\"input_schema\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"grep\",\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\"},{\"input_schema\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"},\"name\":\"ls\",\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\"},{\"input_schema\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"},\"name\":\"sourcegraph\",\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"},\"name\":\"view\",\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\"},{\"input_schema\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"},\"name\":\"write\",\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-sonnet-4-6","id":"msg_01SVo1zkEqqJJbdzrMJGAcEW","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":1,"cache_creation_input_tokens":216,"cache_read_input_tokens":8272,"cache_creation":{"ephemeral_5m_input_tokens":216,"ephemeral_1h_input_tokens":0},"output_tokens":1,"service_tier":"standard","inference_geo":"global"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"Module"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" `example.com/testproject`, Go 1.23, no external dependencies."} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0}
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":1,"cache_creation_input_tokens":216,"cache_read_input_tokens":8272,"output_tokens":22} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 1.538647875s
diff --git a/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/simple_test.yaml b/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/simple_test.yaml
deleted file mode 100644
index 9073906a5a2ae2ef8d667d09f5cea976d987a5c4..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/simple_test.yaml
+++ /dev/null
@@ -1,107 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 797
- host: ""
- body: '{"max_tokens":40,"messages":[{"content":[{"text":"Generate a concise title for the following content:\n\nHello\n \u003cthink\u003e\n\n\u003c/think\u003e","type":"text"}],"role":"user"}],"model":"claude-haiku-4-5-20251001","system":[{"text":"You will generate a short title based on the first message a user begins a conversation with.\n\n\u003crules\u003e\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\u003c/rules\u003e\n\n /no_think","type":"text"}],"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-haiku-4-5-20251001","id":"msg_019QZqWubF2xvU8AxWu51CHS","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":156,"cache_creation_input_tokens":0,"cache_read_input_tokens":0,"cache_creation":{"ephemeral_5m_input_tokens":0,"ephemeral_1h_input_tokens":0},"output_tokens":1,"service_tier":"standard","inference_geo":"not_available"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"Hello"} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":156,"cache_creation_input_tokens":0,"cache_read_input_tokens":0,"output_tokens":4} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 755.416208ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30956
- host: ""
- body: "{\"max_tokens\":10000,\"messages\":[{\"content\":[{\"text\":\"\\u003csystem_reminder\\u003eThis is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\\u003c/system_reminder\\u003e\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"},{\"text\":\"Hello\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"role\":\"user\"}],\"model\":\"claude-sonnet-4-6\",\"system\":[{\"text\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\u003ccritical_rules\\u003e\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default \\u003c4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\u003c/critical_rules\\u003e\\n\\n\\u003ccommunication_style\\u003e\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\u003c/communication_style\\u003e\\n\\n\\u003ccode_references\\u003e\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\u003c/code_references\\u003e\\n\\n\\u003cworkflow\\u003e\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\u003c/workflow\\u003e\\n\\n\\u003cdecision_making\\u003e\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\u003c/decision_making\\u003e\\n\\n\\u003cediting_files\\u003e\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\u003c/editing_files\\u003e\\n\\n\\u003cwhitespace_and_exact_matching\\u003e\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\u003c/whitespace_and_exact_matching\\u003e\\n\\n\\u003ctask_completion\\u003e\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\u003c/task_completion\\u003e\\n\\n\\u003cerror_handling\\u003e\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\u003c/error_handling\\u003e\\n\\n\\u003cmemory_instructions\\u003e\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\u003c/memory_instructions\\u003e\\n\\n\\u003ccode_conventions\\u003e\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\u003c/code_conventions\\u003e\\n\\n\\u003ctesting\\u003e\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\u003c/testing\\u003e\\n\\n\\u003ctool_usage\\u003e\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\u003cbash_commands\\u003e\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `\\u0026` for background processes that won't stop on their own (e.g., `node server.js \\u0026`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status \\u0026\\u0026 git diff HEAD \\u0026\\u0026 git log -n 3`)\\n\\u003c/bash_commands\\u003e\\n\\u003c/tool_usage\\u003e\\n\\n\\u003cproactiveness\\u003e\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\u003c/proactiveness\\u003e\\n\\n\\u003cfinal_answers\\u003e\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\u003c/final_answers\\u003e\\n\\n\\u003cenv\\u003e\\nWorking directory: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/simple_test\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\u003c/env\\u003e\\n\\n\\n\\n\\n\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"tool_choice\":{\"disable_parallel_tool_use\":false,\"type\":\"auto\"},\"tools\":[{\"input_schema\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"},\"name\":\"bash\",\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\u003ccross_platform\\u003e\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\u003c/cross_platform\\u003e\\n\\n\\u003cexecution_steps\\u003e\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with \\u003ccwd\\u003e\\u003c/cwd\\u003e tags\\n\\u003c/execution_steps\\u003e\\n\\n\\u003cusage_notes\\u003e\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '\\u0026\\u0026', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\u003c/usage_notes\\u003e\\n\\n\\u003cbackground_execution\\u003e\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `\\u0026` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\u003c/background_execution\\u003e\\n\\n\\u003cgit_commits\\u003e\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in \\u003ccommit_analysis\\u003e tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\u003ccrush@charm.land\\u003e\\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\u003c/git_commits\\u003e\\n\\n\\u003cpull_requests\\u003e\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in \\u003cpr_analysis\\u003e tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n\\n ## Summary\\n\\n \\u0026lt;1-3 bullet points\\u003e\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\u003c/pull_requests\\u003e\\n\\n\\u003cexamples\\u003e\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar \\u0026\\u0026 pytest tests\\n\\u003c/examples\\u003e\\n\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"},\"name\":\"download\",\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"},\"name\":\"edit\",\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\"},{\"input_schema\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"},\"name\":\"multiedit\",\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\"},{\"input_schema\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"},\"name\":\"fetch\",\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\"},{\"input_schema\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"glob\",\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\"},{\"input_schema\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"grep\",\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\"},{\"input_schema\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"},\"name\":\"ls\",\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\"},{\"input_schema\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"},\"name\":\"sourcegraph\",\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"},\"name\":\"view\",\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\"},{\"input_schema\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"},\"name\":\"write\",\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-sonnet-4-6","id":"msg_0172mX38eGNfBfzSWdyVhYZA","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":3,"cache_creation_input_tokens":8267,"cache_read_input_tokens":0,"cache_creation":{"ephemeral_5m_input_tokens":8267,"ephemeral_1h_input_tokens":0},"output_tokens":1,"service_tier":"standard","inference_geo":"global"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"Hello"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"! How can I help?"}}
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":3,"cache_creation_input_tokens":8267,"cache_read_input_tokens":0,"output_tokens":10} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 1.658727958s
diff --git a/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/sourcegraph_tool.yaml b/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/sourcegraph_tool.yaml
deleted file mode 100644
index 20a57ba45179306fc997a26bdeb368e3d777a3df..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/sourcegraph_tool.yaml
+++ /dev/null
@@ -1,229 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 852
- host: ""
- body: '{"max_tokens":40,"messages":[{"content":[{"text":"Generate a concise title for the following content:\n\nuse sourcegraph to search for ''func main'' in Go repositories\n \u003cthink\u003e\n\n\u003c/think\u003e","type":"text"}],"role":"user"}],"model":"claude-haiku-4-5-20251001","system":[{"text":"You will generate a short title based on the first message a user begins a conversation with.\n\n\u003crules\u003e\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\u003c/rules\u003e\n\n /no_think","type":"text"}],"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-haiku-4-5-20251001","id":"msg_016XyqvjsUrPpyWxdMpxne7i","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":170,"cache_creation_input_tokens":0,"cache_read_input_tokens":0,"cache_creation":{"ephemeral_5m_input_tokens":0,"ephemeral_1h_input_tokens":0},"output_tokens":5,"service_tier":"standard","inference_geo":"not_available"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"Sourcegraph Go"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" func main Search"} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":170,"cache_creation_input_tokens":0,"cache_read_input_tokens":0,"output_tokens":11} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 692.93075ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31016
- host: ""
- body: "{\"max_tokens\":10000,\"messages\":[{\"content\":[{\"text\":\"\\u003csystem_reminder\\u003eThis is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\\u003c/system_reminder\\u003e\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"},{\"text\":\"use sourcegraph to search for 'func main' in Go repositories\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"role\":\"user\"}],\"model\":\"claude-sonnet-4-6\",\"system\":[{\"text\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\u003ccritical_rules\\u003e\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default \\u003c4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\u003c/critical_rules\\u003e\\n\\n\\u003ccommunication_style\\u003e\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\u003c/communication_style\\u003e\\n\\n\\u003ccode_references\\u003e\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\u003c/code_references\\u003e\\n\\n\\u003cworkflow\\u003e\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\u003c/workflow\\u003e\\n\\n\\u003cdecision_making\\u003e\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\u003c/decision_making\\u003e\\n\\n\\u003cediting_files\\u003e\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\u003c/editing_files\\u003e\\n\\n\\u003cwhitespace_and_exact_matching\\u003e\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\u003c/whitespace_and_exact_matching\\u003e\\n\\n\\u003ctask_completion\\u003e\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\u003c/task_completion\\u003e\\n\\n\\u003cerror_handling\\u003e\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\u003c/error_handling\\u003e\\n\\n\\u003cmemory_instructions\\u003e\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\u003c/memory_instructions\\u003e\\n\\n\\u003ccode_conventions\\u003e\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\u003c/code_conventions\\u003e\\n\\n\\u003ctesting\\u003e\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\u003c/testing\\u003e\\n\\n\\u003ctool_usage\\u003e\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\u003cbash_commands\\u003e\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `\\u0026` for background processes that won't stop on their own (e.g., `node server.js \\u0026`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status \\u0026\\u0026 git diff HEAD \\u0026\\u0026 git log -n 3`)\\n\\u003c/bash_commands\\u003e\\n\\u003c/tool_usage\\u003e\\n\\n\\u003cproactiveness\\u003e\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\u003c/proactiveness\\u003e\\n\\n\\u003cfinal_answers\\u003e\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\u003c/final_answers\\u003e\\n\\n\\u003cenv\\u003e\\nWorking directory: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/sourcegraph_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\u003c/env\\u003e\\n\\n\\n\\n\\n\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"tool_choice\":{\"disable_parallel_tool_use\":false,\"type\":\"auto\"},\"tools\":[{\"input_schema\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"},\"name\":\"bash\",\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\u003ccross_platform\\u003e\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\u003c/cross_platform\\u003e\\n\\n\\u003cexecution_steps\\u003e\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with \\u003ccwd\\u003e\\u003c/cwd\\u003e tags\\n\\u003c/execution_steps\\u003e\\n\\n\\u003cusage_notes\\u003e\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '\\u0026\\u0026', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\u003c/usage_notes\\u003e\\n\\n\\u003cbackground_execution\\u003e\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `\\u0026` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\u003c/background_execution\\u003e\\n\\n\\u003cgit_commits\\u003e\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in \\u003ccommit_analysis\\u003e tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\u003ccrush@charm.land\\u003e\\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\u003c/git_commits\\u003e\\n\\n\\u003cpull_requests\\u003e\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in \\u003cpr_analysis\\u003e tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n\\n ## Summary\\n\\n \\u0026lt;1-3 bullet points\\u003e\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\u003c/pull_requests\\u003e\\n\\n\\u003cexamples\\u003e\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar \\u0026\\u0026 pytest tests\\n\\u003c/examples\\u003e\\n\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"},\"name\":\"download\",\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"},\"name\":\"edit\",\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\"},{\"input_schema\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"},\"name\":\"multiedit\",\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\"},{\"input_schema\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"},\"name\":\"fetch\",\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\"},{\"input_schema\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"glob\",\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\"},{\"input_schema\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"grep\",\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\"},{\"input_schema\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"},\"name\":\"ls\",\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\"},{\"input_schema\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"},\"name\":\"sourcegraph\",\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"},\"name\":\"view\",\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\"},{\"input_schema\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"},\"name\":\"write\",\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-sonnet-4-6","id":"msg_019qrShkWbJ2thQQAxM6pNrG","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":3,"cache_creation_input_tokens":4550,"cache_read_input_tokens":3733,"cache_creation":{"ephemeral_5m_input_tokens":4550,"ephemeral_1h_input_tokens":0},"output_tokens":58,"service_tier":"standard","inference_geo":"global"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"tool_use","id":"toolu_012pWRXFz2ahxJw8gWbJXU1s","name":"sourcegraph","input":{},"caller":{"type":"direct"}} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"{\"qu"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"ery\": "} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"\"func "} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"main l"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"ang:go\"}"} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"tool_use","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":3,"cache_creation_input_tokens":4550,"cache_read_input_tokens":3733,"output_tokens":58} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 2.294361041s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 424
- host: sourcegraph.com
- body: '{"query":"query Search($query: String!) { search(query: $query, version: V2, patternType: keyword ) { results { matchCount, limitHit, resultCount, approximateResultCount, missing { name }, timedout { name }, indexUnavailable, results { __typename, ... on FileMatch { repository { name }, file { path, url, content }, lineMatches { preview, lineNumber, offsetAndLengths } } } } } }","variables":{"query":"func main lang:go"}}'
- headers:
- Content-Type:
- - application/json
- User-Agent:
- - crush/1.0
- url: https://sourcegraph.com/.api/graphql
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: '{"data":{"search":{"results":{"matchCount":30,"limitHit":true,"resultCount":30,"approximateResultCount":"30+","missing":[],"timedout":[],"indexUnavailable":false,"results":[{"__typename":"FileMatch","repository":{"name":"github.com/GoesToEleven/GolangTraining"},"file":{"path":"14_functions/10_func-expression/01_before-func-expression/main.go","url":"/r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/01_before-func-expression/main.go","content":"package main\n\nimport \"fmt\"\n\nfunc greeting() {\n\tfmt.Println(\"Hello world!\")\n}\n\nfunc main() {\n\tgreeting()\n}\n"},"lineMatches":[{"preview":"func main() {","lineNumber":8,"offsetAndLengths":[[0,9]]},{"preview":"package main","lineNumber":0,"offsetAndLengths":[[8,4]]},{"preview":"func greeting() {","lineNumber":4,"offsetAndLengths":[[0,4]]}]},{"__typename":"FileMatch","repository":{"name":"github.com/GoesToEleven/GolangTraining"},"file":{"path":"14_functions/10_func-expression/02_func-expression/main.go","url":"/r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/02_func-expression/main.go","content":"package main\n\nimport \"fmt\"\n\nfunc main() {\n\n\tgreeting := func() {\n\t\tfmt.Println(\"Hello world!\")\n\t}\n\n\tgreeting()\n}\n"},"lineMatches":[{"preview":"func main() {","lineNumber":4,"offsetAndLengths":[[0,9]]},{"preview":"\tgreeting := func() {","lineNumber":6,"offsetAndLengths":[[13,4]]},{"preview":"package main","lineNumber":0,"offsetAndLengths":[[8,4]]}]},{"__typename":"FileMatch","repository":{"name":"github.com/GoesToEleven/GolangTraining"},"file":{"path":"14_functions/10_func-expression/03_func-expression_shows-type/main.go","url":"/r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/03_func-expression_shows-type/main.go","content":"package main\n\nimport \"fmt\"\n\nfunc main() {\n\n\tgreeting := func() {\n\t\tfmt.Println(\"Hello world!\")\n\t}\n\n\tgreeting()\n\tfmt.Printf(\"%T\\n\", greeting)\n}\n"},"lineMatches":[{"preview":"func main() {","lineNumber":4,"offsetAndLengths":[[0,9]]},{"preview":"\tgreeting := func() {","lineNumber":6,"offsetAndLengths":[[13,4]]},{"preview":"package main","lineNumber":0,"offsetAndLengths":[[8,4]]}]},{"__typename":"FileMatch","repository":{"name":"github.com/GoesToEleven/GolangTraining"},"file":{"path":"14_functions/10_func-expression/04_another-way_func-expression/main.go","url":"/r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/04_another-way_func-expression/main.go","content":"package main\n\nimport \"fmt\"\n\nfunc makeGreeter() func() string {\n\treturn func() string {\n\t\treturn \"Hello world!\"\n\t}\n}\n\nfunc main() {\n\tgreet := makeGreeter()\n\tfmt.Println(greet())\n}\n"},"lineMatches":[{"preview":"func main() {","lineNumber":10,"offsetAndLengths":[[0,9]]},{"preview":"package main","lineNumber":0,"offsetAndLengths":[[8,4]]},{"preview":"func makeGreeter() func() string {","lineNumber":4,"offsetAndLengths":[[0,4],[19,4]]},{"preview":"\treturn func() string {","lineNumber":5,"offsetAndLengths":[[8,4]]}]},{"__typename":"FileMatch","repository":{"name":"github.com/jmbharathram/executeoncommand"},"file":{"path":"go/channel/main.go","url":"/r/github.com/jmbharathram/executeoncommand/-/blob/go/channel/main.go","content":"package main\n\nimport \"fmt\"\n\n// Example 0\nfunc dummy_func() {\n\tfmt.Println(\"dummy_func\")\n}\nfunc main() {\n\tgo dummy_func()\n fmt.Println(\"end of main\")\n}\n\n\n// Example 1\n\npackage main\n\nimport \"fmt\"\n\nfunc dummy_func(ch chan int) {\n // receive a value from channel\n val := \u003c-ch\n\tfmt.Println(val)\n}\nfunc main() {\n // create a channel\n\tch := make(chan int)\n // trigger a parallel go routine\n\tgo dummy_func(ch)\n // send a value into the channel\n\tch \u003c- 10\n}\n\n// Example 2\n\npackage main\n\nimport (\n \"fmt\"\n //\"flag\"\n)\n\nfunc dummy_func(ch chan int) {\n // receive a value from channel\n val := \u003c-ch\n\tfmt.Println(val)\n\n //send a value again\n ch \u003c- val*2\n}\nfunc main() {\n fmt.Println(\"Welcome to main function\")\n \n //flag statements for user input\n var input_value int\n input_value = 10\n //flag.IntVar(\u0026input_value, \"v\", 10, \"Input value to send/receive via channel.\" )\n //flag.Parse()\n\n // create a channel\n\tch := make(chan int)\n\n // trigger a parallel go routine\n\tgo dummy_func(ch)\n \n // send a value into the channel\n\tch \u003c- input_value\n\n // receive a value from channel\n response := \u003c- ch\n fmt.Println(response)\n\n}\n\n// example 3\n\n\npackage main\n\nimport(\n\"fmt\"\n\"time\"\n\"math/rand\"\n)\n\nfunc routine1(ch chan int, min int, max int) {\n\n rand.Seed(time.Now().UnixNano())\n t := rand.Intn(max - min) + min\n fmt.Println(\"routine 1 is waiting: \", t, \"secs\")\n time.Sleep(time.Duration(t)*time.Second)\n\n ch \u003c- t\n}\n\nfunc routine2(ch chan int, min int, max int) {\n\n rand.Seed(time.Now().UnixNano())\n t := rand.Intn(max - min) + min\n fmt.Println(\"routine 2 is waiting: \", t, \"secs\")\n time.Sleep(time.Duration(t)*time.Second)\n\n ch \u003c- t\n}\n\nfunc main(){\n\nmin := 1\nmax := 10\n\nch1:= make(chan int)\nch2:= make(chan int)\n\ngo routine1(ch1, min, max)\ngo routine2(ch2, min, max)\n\nselect{\n\ncase \u003c- ch1:\n fmt.Println(\"routine 1 finished\")\n\ncase \u003c- ch2:\n fmt.Println(\"routine 2 finished\")\n}\n\nfmt.Println(\"end of main function\")\n\n}\n"},"lineMatches":[{"preview":"func main(){","lineNumber":106,"offsetAndLengths":[[0,9]]},{"preview":"func dummy_func() {","lineNumber":5,"offsetAndLengths":[[0,4],[11,4]]},{"preview":"\tfmt.Println(\"dummy_func\")","lineNumber":6,"offsetAndLengths":[[20,4]]},{"preview":"func main() {","lineNumber":8,"offsetAndLengths":[[0,9]]},{"preview":"\tgo dummy_func()","lineNumber":9,"offsetAndLengths":[[10,4]]},{"preview":" fmt.Println(\"end of main\")","lineNumber":10,"offsetAndLengths":[[24,4]]},{"preview":"func main() {","lineNumber":25,"offsetAndLengths":[[0,9]]},{"preview":"func main() {","lineNumber":51,"offsetAndLengths":[[0,9]]},{"preview":" fmt.Println(\"Welcome to main function\")","lineNumber":52,"offsetAndLengths":[[28,4],[33,4]]},{"preview":"fmt.Println(\"end of main function\")","lineNumber":126,"offsetAndLengths":[[20,4],[25,4]]},{"preview":"package main","lineNumber":78,"offsetAndLengths":[[8,4]]},{"preview":"package main","lineNumber":16,"offsetAndLengths":[[8,4]]},{"preview":"package main","lineNumber":0,"offsetAndLengths":[[8,4]]}]}]}}}}'
- headers:
- Content-Type:
- - application/json
- status: 200 OK
- code: 200
- duration: 1.520166542s
-- id: 3
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 40418
- host: ""
- body: "{\"max_tokens\":10000,\"messages\":[{\"content\":[{\"text\":\"\\u003csystem_reminder\\u003eThis is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\\u003c/system_reminder\\u003e\",\"type\":\"text\"},{\"text\":\"use sourcegraph to search for 'func main' in Go repositories\",\"type\":\"text\"}],\"role\":\"user\"},{\"content\":[{\"id\":\"toolu_012pWRXFz2ahxJw8gWbJXU1s\",\"input\":{\"query\":\"func main lang:go\"},\"name\":\"sourcegraph\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"tool_use\"}],\"role\":\"assistant\"},{\"content\":[{\"tool_use_id\":\"toolu_012pWRXFz2ahxJw8gWbJXU1s\",\"cache_control\":{\"type\":\"ephemeral\"},\"content\":[{\"text\":\"# Sourcegraph Search Results\\n\\nFound 30 matches across 30 results\\n(Result limit reached, try a more specific query)\\n\\n## Result 1: github.com/GoesToEleven/GolangTraining/14_functions/10_func-expression/01_before-func-expression/main.go\\n\\nURL: /r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/01_before-func-expression/main.go\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func greeting() {\\n6| \\tfmt.Println(\\\"Hello world!\\\")\\n7| }\\n8| func main() {\\n9| func main() {\\n10| \\tgreeting()\\n11| }\\n12| \\n```\\n\\n```\\n0| package main\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func greeting() {\\n6| \\tfmt.Println(\\\"Hello world!\\\")\\n7| }\\n8| \\n9| func main() {\\n10| \\tgreeting()\\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| func greeting() {\\n5| func greeting() {\\n6| \\tfmt.Println(\\\"Hello world!\\\")\\n7| }\\n8| \\n9| func main() {\\n10| \\tgreeting()\\n11| }\\n12| \\n```\\n\\n## Result 2: github.com/GoesToEleven/GolangTraining/14_functions/10_func-expression/02_func-expression/main.go\\n\\nURL: /r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/02_func-expression/main.go\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| func main() {\\n5| func main() {\\n6| \\n7| \\tgreeting := func() {\\n8| \\t\\tfmt.Println(\\\"Hello world!\\\")\\n9| \\t}\\n10| \\n11| \\tgreeting()\\n12| }\\n13| \\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func main() {\\n6| \\tgreeting := func() {\\n7| \\tgreeting := func() {\\n8| \\t\\tfmt.Println(\\\"Hello world!\\\")\\n9| \\t}\\n10| \\n11| \\tgreeting()\\n12| }\\n13| \\n```\\n\\n```\\n0| package main\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func main() {\\n6| \\n7| \\tgreeting := func() {\\n8| \\t\\tfmt.Println(\\\"Hello world!\\\")\\n9| \\t}\\n10| \\n```\\n\\n## Result 3: github.com/GoesToEleven/GolangTraining/14_functions/10_func-expression/03_func-expression_shows-type/main.go\\n\\nURL: /r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/03_func-expression_shows-type/main.go\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| func main() {\\n5| func main() {\\n6| \\n7| \\tgreeting := func() {\\n8| \\t\\tfmt.Println(\\\"Hello world!\\\")\\n9| \\t}\\n10| \\n11| \\tgreeting()\\n12| \\tfmt.Printf(\\\"%T\\\\n\\\", greeting)\\n13| }\\n14| \\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func main() {\\n6| \\tgreeting := func() {\\n7| \\tgreeting := func() {\\n8| \\t\\tfmt.Println(\\\"Hello world!\\\")\\n9| \\t}\\n10| \\n11| \\tgreeting()\\n12| \\tfmt.Printf(\\\"%T\\\\n\\\", greeting)\\n13| }\\n14| \\n```\\n\\n```\\n0| package main\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func main() {\\n6| \\n7| \\tgreeting := func() {\\n8| \\t\\tfmt.Println(\\\"Hello world!\\\")\\n9| \\t}\\n10| \\n```\\n\\n## Result 4: github.com/GoesToEleven/GolangTraining/14_functions/10_func-expression/04_another-way_func-expression/main.go\\n\\nURL: /r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/04_another-way_func-expression/main.go\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func makeGreeter() func() string {\\n6| \\treturn func() string {\\n7| \\t\\treturn \\\"Hello world!\\\"\\n8| \\t}\\n9| }\\n10| func main() {\\n11| func main() {\\n12| \\tgreet := makeGreeter()\\n13| \\tfmt.Println(greet())\\n14| }\\n15| \\n```\\n\\n```\\n0| package main\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func makeGreeter() func() string {\\n6| \\treturn func() string {\\n7| \\t\\treturn \\\"Hello world!\\\"\\n8| \\t}\\n9| }\\n10| \\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| func makeGreeter() func() string {\\n5| func makeGreeter() func() string {\\n6| \\treturn func() string {\\n7| \\t\\treturn \\\"Hello world!\\\"\\n8| \\t}\\n9| }\\n10| \\n11| func main() {\\n12| \\tgreet := makeGreeter()\\n13| \\tfmt.Println(greet())\\n14| }\\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| \\treturn func() string {\\n6| \\treturn func() string {\\n7| \\t\\treturn \\\"Hello world!\\\"\\n8| \\t}\\n9| }\\n10| \\n11| func main() {\\n12| \\tgreet := makeGreeter()\\n13| \\tfmt.Println(greet())\\n14| }\\n15| \\n```\\n\\n## Result 5: github.com/jmbharathram/executeoncommand/go/channel/main.go\\n\\nURL: /r/github.com/jmbharathram/executeoncommand/-/blob/go/channel/main.go\\n\\n```\\n96| \\n97| func routine2(ch chan int, min int, max int) {\\n98| \\n99| rand.Seed(time.Now().UnixNano())\\n100| t := rand.Intn(max - min) + min\\n101| fmt.Println(\\\"routine 2 is waiting: \\\", t, \\\"secs\\\")\\n102| time.Sleep(time.Duration(t)*time.Second)\\n103| \\n104| ch \\u003c- t\\n105| }\\n106| func main(){\\n107| func main(){\\n108| \\n109| min := 1\\n110| max := 10\\n111| \\n112| ch1:= make(chan int)\\n113| ch2:= make(chan int)\\n114| \\n115| go routine1(ch1, min, max)\\n116| go routine2(ch2, min, max)\\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func dummy_func() {\\n6| func dummy_func() {\\n7| \\tfmt.Println(\\\"dummy_func\\\")\\n8| }\\n9| func main() {\\n10| \\tgo dummy_func()\\n11| fmt.Println(\\\"end of main\\\")\\n12| }\\n13| \\n14| \\n15| // Example 1\\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| // Example 0\\n6| \\tfmt.Println(\\\"dummy_func\\\")\\n7| \\tfmt.Println(\\\"dummy_func\\\")\\n8| }\\n9| func main() {\\n10| \\tgo dummy_func()\\n11| fmt.Println(\\\"end of main\\\")\\n12| }\\n13| \\n14| \\n15| // Example 1\\n16| \\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| // Example 0\\n6| func dummy_func() {\\n7| \\tfmt.Println(\\\"dummy_func\\\")\\n8| func main() {\\n9| func main() {\\n10| \\tgo dummy_func()\\n11| fmt.Println(\\\"end of main\\\")\\n12| }\\n13| \\n14| \\n15| // Example 1\\n16| \\n17| package main\\n18| \\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| // Example 0\\n6| func dummy_func() {\\n7| \\tfmt.Println(\\\"dummy_func\\\")\\n8| }\\n9| \\tgo dummy_func()\\n10| \\tgo dummy_func()\\n11| fmt.Println(\\\"end of main\\\")\\n12| }\\n13| \\n14| \\n15| // Example 1\\n16| \\n17| package main\\n18| \\n19| import \\\"fmt\\\"\\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| // Example 0\\n6| func dummy_func() {\\n7| \\tfmt.Println(\\\"dummy_func\\\")\\n8| }\\n9| func main() {\\n10| fmt.Println(\\\"end of main\\\")\\n11| fmt.Println(\\\"end of main\\\")\\n12| }\\n13| \\n14| \\n15| // Example 1\\n16| \\n17| package main\\n18| \\n19| import \\\"fmt\\\"\\n20| \\n```\\n\\n```\\n15| // Example 1\\n16| \\n17| package main\\n18| \\n19| import \\\"fmt\\\"\\n20| \\n21| func dummy_func(ch chan int) {\\n22| // receive a value from channel\\n23| val := \\u003c-ch\\n24| \\tfmt.Println(val)\\n25| func main() {\\n26| func main() {\\n27| // create a channel\\n28| \\tch := make(chan int)\\n29| // trigger a parallel go routine\\n30| \\tgo dummy_func(ch)\\n31| // send a value into the channel\\n32| \\tch \\u003c- 10\\n33| }\\n34| \\n35| // Example 2\\n```\\n\\n```\\n41| //\\\"flag\\\"\\n42| )\\n43| \\n44| func dummy_func(ch chan int) {\\n45| // receive a value from channel\\n46| val := \\u003c-ch\\n47| \\tfmt.Println(val)\\n48| \\n49| //send a value again\\n50| ch \\u003c- val*2\\n51| func main() {\\n52| func main() {\\n53| fmt.Println(\\\"Welcome to main function\\\")\\n54| \\n55| //flag statements for user input\\n56| var input_value int\\n57| input_value = 10\\n58| //flag.IntVar(\\u0026input_value, \\\"v\\\", 10, \\\"Input value to send/receive via channel.\\\" )\\n59| //flag.Parse()\\n60| \\n61| // create a channel\\n```\\n\\n```\\n42| )\\n43| \\n44| func dummy_func(ch chan int) {\\n45| // receive a value from channel\\n46| val := \\u003c-ch\\n47| \\tfmt.Println(val)\\n48| \\n49| //send a value again\\n50| ch \\u003c- val*2\\n51| }\\n52| fmt.Println(\\\"Welcome to main function\\\")\\n53| fmt.Println(\\\"Welcome to main function\\\")\\n54| \\n55| //flag statements for user input\\n56| var input_value int\\n57| input_value = 10\\n58| //flag.IntVar(\\u0026input_value, \\\"v\\\", 10, \\\"Input value to send/receive via channel.\\\" )\\n59| //flag.Parse()\\n60| \\n61| // create a channel\\n62| \\tch := make(chan int)\\n```\\n\\n```\\n116| go routine2(ch2, min, max)\\n117| \\n118| select{\\n119| \\n120| case \\u003c- ch1:\\n121| fmt.Println(\\\"routine 1 finished\\\")\\n122| \\n123| case \\u003c- ch2:\\n124| fmt.Println(\\\"routine 2 finished\\\")\\n125| }\\n126| fmt.Println(\\\"end of main function\\\")\\n127| fmt.Println(\\\"end of main function\\\")\\n128| \\n129| }\\n130| \\n```\\n\\n```\\n68| \\tch \\u003c- input_value\\n69| \\n70| // receive a value from channel\\n71| response := \\u003c- ch\\n72| fmt.Println(response)\\n73| \\n74| }\\n75| \\n76| // example 3\\n77| \\n78| package main\\n79| package main\\n80| \\n81| import(\\n82| \\\"fmt\\\"\\n83| \\\"time\\\"\\n84| \\\"math/rand\\\"\\n85| )\\n86| \\n87| func routine1(ch chan int, min int, max int) {\\n88| \\n```\\n\\n```\\n6| func dummy_func() {\\n7| \\tfmt.Println(\\\"dummy_func\\\")\\n8| }\\n9| func main() {\\n10| \\tgo dummy_func()\\n11| fmt.Println(\\\"end of main\\\")\\n12| }\\n13| \\n14| \\n15| // Example 1\\n16| package main\\n17| package main\\n18| \\n19| import \\\"fmt\\\"\\n20| \\n21| func dummy_func(ch chan int) {\\n22| // receive a value from channel\\n23| val := \\u003c-ch\\n24| \\tfmt.Println(val)\\n25| }\\n26| func main() {\\n```\\n\\n```\\n0| package main\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| // Example 0\\n6| func dummy_func() {\\n7| \\tfmt.Println(\\\"dummy_func\\\")\\n8| }\\n9| func main() {\\n10| \\tgo dummy_func()\\n```\\n\\n\",\"type\":\"text\"}],\"type\":\"tool_result\"}],\"role\":\"user\"}],\"model\":\"claude-sonnet-4-6\",\"system\":[{\"text\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\u003ccritical_rules\\u003e\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default \\u003c4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\u003c/critical_rules\\u003e\\n\\n\\u003ccommunication_style\\u003e\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\u003c/communication_style\\u003e\\n\\n\\u003ccode_references\\u003e\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\u003c/code_references\\u003e\\n\\n\\u003cworkflow\\u003e\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\u003c/workflow\\u003e\\n\\n\\u003cdecision_making\\u003e\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\u003c/decision_making\\u003e\\n\\n\\u003cediting_files\\u003e\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\u003c/editing_files\\u003e\\n\\n\\u003cwhitespace_and_exact_matching\\u003e\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\u003c/whitespace_and_exact_matching\\u003e\\n\\n\\u003ctask_completion\\u003e\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\u003c/task_completion\\u003e\\n\\n\\u003cerror_handling\\u003e\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\u003c/error_handling\\u003e\\n\\n\\u003cmemory_instructions\\u003e\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\u003c/memory_instructions\\u003e\\n\\n\\u003ccode_conventions\\u003e\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\u003c/code_conventions\\u003e\\n\\n\\u003ctesting\\u003e\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\u003c/testing\\u003e\\n\\n\\u003ctool_usage\\u003e\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\u003cbash_commands\\u003e\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `\\u0026` for background processes that won't stop on their own (e.g., `node server.js \\u0026`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status \\u0026\\u0026 git diff HEAD \\u0026\\u0026 git log -n 3`)\\n\\u003c/bash_commands\\u003e\\n\\u003c/tool_usage\\u003e\\n\\n\\u003cproactiveness\\u003e\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\u003c/proactiveness\\u003e\\n\\n\\u003cfinal_answers\\u003e\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\u003c/final_answers\\u003e\\n\\n\\u003cenv\\u003e\\nWorking directory: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/sourcegraph_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\u003c/env\\u003e\\n\\n\\n\\n\\n\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"tool_choice\":{\"disable_parallel_tool_use\":false,\"type\":\"auto\"},\"tools\":[{\"input_schema\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"},\"name\":\"bash\",\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\u003ccross_platform\\u003e\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\u003c/cross_platform\\u003e\\n\\n\\u003cexecution_steps\\u003e\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with \\u003ccwd\\u003e\\u003c/cwd\\u003e tags\\n\\u003c/execution_steps\\u003e\\n\\n\\u003cusage_notes\\u003e\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '\\u0026\\u0026', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\u003c/usage_notes\\u003e\\n\\n\\u003cbackground_execution\\u003e\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `\\u0026` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\u003c/background_execution\\u003e\\n\\n\\u003cgit_commits\\u003e\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in \\u003ccommit_analysis\\u003e tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\u003ccrush@charm.land\\u003e\\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\u003c/git_commits\\u003e\\n\\n\\u003cpull_requests\\u003e\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in \\u003cpr_analysis\\u003e tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n\\n ## Summary\\n\\n \\u0026lt;1-3 bullet points\\u003e\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\u003c/pull_requests\\u003e\\n\\n\\u003cexamples\\u003e\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar \\u0026\\u0026 pytest tests\\n\\u003c/examples\\u003e\\n\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"},\"name\":\"download\",\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"},\"name\":\"edit\",\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\"},{\"input_schema\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"},\"name\":\"multiedit\",\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\"},{\"input_schema\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"},\"name\":\"fetch\",\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\"},{\"input_schema\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"glob\",\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\"},{\"input_schema\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"grep\",\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\"},{\"input_schema\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"},\"name\":\"ls\",\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\"},{\"input_schema\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"},\"name\":\"sourcegraph\",\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"},\"name\":\"view\",\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\"},{\"input_schema\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"},\"name\":\"write\",\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-sonnet-4-6","id":"msg_01Wk9W4jkeD1U8bVHPLH7eeY","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":1,"cache_creation_input_tokens":3595,"cache_read_input_tokens":8283,"cache_creation":{"ephemeral_5m_input_tokens":3595,"ephemeral_1h_input_tokens":0},"output_tokens":1,"service_tier":"standard","inference_geo":"global"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""}}
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"Here"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"'s a summary of results for `func main` in Go repos:\n\n-"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" **30 matches** across 30 files (result limit"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" reached)\n- Top repos found:\n - `GoesToEleven"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"/GolangTraining` — multiple training examples showing"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" different function expression patterns\n - `jmbharathram/executeoncommand` — Go channel examples"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" with goroutines\n\nAll matches show standard `"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"func main()` entry points in `package main` files. To narrow results"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":", try a more specific query (e.g., add a repo"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" filter or additional context around `func main`)."} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":1,"cache_creation_input_tokens":3595,"cache_read_input_tokens":8283,"output_tokens":134} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 1.27399475s
diff --git a/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/update_a_file.yaml b/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/update_a_file.yaml
deleted file mode 100644
index 2b073b3584becd877b5014453a3ab16183a9e662..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/update_a_file.yaml
+++ /dev/null
@@ -1,381 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 861
- host: ""
- body: '{"max_tokens":40,"messages":[{"content":[{"text":"Generate a concise title for the following content:\n\nupdate the main.go file by changing the print to say hello from crush\n \u003cthink\u003e\n\n\u003c/think\u003e","type":"text"}],"role":"user"}],"model":"claude-haiku-4-5-20251001","system":[{"text":"You will generate a short title based on the first message a user begins a conversation with.\n\n\u003crules\u003e\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\u003c/rules\u003e\n\n /no_think","type":"text"}],"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-haiku-4-5-20251001","id":"msg_01UR3MBvvH5GXuLq66hf7uDj","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":170,"cache_creation_input_tokens":0,"cache_read_input_tokens":0,"cache_creation":{"ephemeral_5m_input_tokens":0,"ephemeral_1h_input_tokens":0},"output_tokens":1,"service_tier":"standard","inference_geo":"not_available"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""}}
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"Update"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" main.go print statement to say hello from crush"} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":170,"cache_creation_input_tokens":0,"cache_read_input_tokens":0,"output_tokens":14} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 681.551083ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31022
- host: ""
- body: "{\"max_tokens\":10000,\"messages\":[{\"content\":[{\"text\":\"\\u003csystem_reminder\\u003eThis is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\\u003c/system_reminder\\u003e\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"},{\"text\":\"update the main.go file by changing the print to say hello from crush\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"role\":\"user\"}],\"model\":\"claude-sonnet-4-6\",\"system\":[{\"text\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\u003ccritical_rules\\u003e\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default \\u003c4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\u003c/critical_rules\\u003e\\n\\n\\u003ccommunication_style\\u003e\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\u003c/communication_style\\u003e\\n\\n\\u003ccode_references\\u003e\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\u003c/code_references\\u003e\\n\\n\\u003cworkflow\\u003e\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\u003c/workflow\\u003e\\n\\n\\u003cdecision_making\\u003e\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\u003c/decision_making\\u003e\\n\\n\\u003cediting_files\\u003e\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\u003c/editing_files\\u003e\\n\\n\\u003cwhitespace_and_exact_matching\\u003e\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\u003c/whitespace_and_exact_matching\\u003e\\n\\n\\u003ctask_completion\\u003e\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\u003c/task_completion\\u003e\\n\\n\\u003cerror_handling\\u003e\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\u003c/error_handling\\u003e\\n\\n\\u003cmemory_instructions\\u003e\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\u003c/memory_instructions\\u003e\\n\\n\\u003ccode_conventions\\u003e\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\u003c/code_conventions\\u003e\\n\\n\\u003ctesting\\u003e\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\u003c/testing\\u003e\\n\\n\\u003ctool_usage\\u003e\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\u003cbash_commands\\u003e\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `\\u0026` for background processes that won't stop on their own (e.g., `node server.js \\u0026`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status \\u0026\\u0026 git diff HEAD \\u0026\\u0026 git log -n 3`)\\n\\u003c/bash_commands\\u003e\\n\\u003c/tool_usage\\u003e\\n\\n\\u003cproactiveness\\u003e\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\u003c/proactiveness\\u003e\\n\\n\\u003cfinal_answers\\u003e\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\u003c/final_answers\\u003e\\n\\n\\u003cenv\\u003e\\nWorking directory: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/update_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\u003c/env\\u003e\\n\\n\\n\\n\\n\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"tool_choice\":{\"disable_parallel_tool_use\":false,\"type\":\"auto\"},\"tools\":[{\"input_schema\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"},\"name\":\"bash\",\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\u003ccross_platform\\u003e\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\u003c/cross_platform\\u003e\\n\\n\\u003cexecution_steps\\u003e\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with \\u003ccwd\\u003e\\u003c/cwd\\u003e tags\\n\\u003c/execution_steps\\u003e\\n\\n\\u003cusage_notes\\u003e\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '\\u0026\\u0026', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\u003c/usage_notes\\u003e\\n\\n\\u003cbackground_execution\\u003e\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `\\u0026` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\u003c/background_execution\\u003e\\n\\n\\u003cgit_commits\\u003e\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in \\u003ccommit_analysis\\u003e tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\u003ccrush@charm.land\\u003e\\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\u003c/git_commits\\u003e\\n\\n\\u003cpull_requests\\u003e\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in \\u003cpr_analysis\\u003e tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n\\n ## Summary\\n\\n \\u0026lt;1-3 bullet points\\u003e\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\u003c/pull_requests\\u003e\\n\\n\\u003cexamples\\u003e\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar \\u0026\\u0026 pytest tests\\n\\u003c/examples\\u003e\\n\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"},\"name\":\"download\",\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"},\"name\":\"edit\",\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\"},{\"input_schema\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"},\"name\":\"multiedit\",\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\"},{\"input_schema\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"},\"name\":\"fetch\",\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\"},{\"input_schema\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"glob\",\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\"},{\"input_schema\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"grep\",\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\"},{\"input_schema\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"},\"name\":\"ls\",\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\"},{\"input_schema\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"},\"name\":\"sourcegraph\",\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"},\"name\":\"view\",\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\"},{\"input_schema\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"},\"name\":\"write\",\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-sonnet-4-6","id":"msg_01UtAG5a5ZmSgV6WuQbUDzUS","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":3,"cache_creation_input_tokens":4550,"cache_read_input_tokens":3733,"cache_creation":{"ephemeral_5m_input_tokens":4550,"ephemeral_1h_input_tokens":0},"output_tokens":35,"service_tier":"standard","inference_geo":"global"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"tool_use","id":"toolu_01C5th3Xy2LJY4R51J9ATgpd","name":"ls","input":{},"caller":{"type":"direct"}} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":""} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"tool_use","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":3,"cache_creation_input_tokens":4550,"cache_read_input_tokens":3733,"output_tokens":35} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 2.652786958s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31365
- host: ""
- body: "{\"max_tokens\":10000,\"messages\":[{\"content\":[{\"text\":\"\\u003csystem_reminder\\u003eThis is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\\u003c/system_reminder\\u003e\",\"type\":\"text\"},{\"text\":\"update the main.go file by changing the print to say hello from crush\",\"type\":\"text\"}],\"role\":\"user\"},{\"content\":[{\"id\":\"toolu_01C5th3Xy2LJY4R51J9ATgpd\",\"input\":{},\"name\":\"ls\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"tool_use\"}],\"role\":\"assistant\"},{\"content\":[{\"tool_use_id\":\"toolu_01C5th3Xy2LJY4R51J9ATgpd\",\"cache_control\":{\"type\":\"ephemeral\"},\"content\":[{\"text\":\"\\n- /tmp/crush-test/TestCoderAgent/anthropic-sonnet/update_a_file/\\n - go.mod\\n - main.go\\n\",\"type\":\"text\"}],\"type\":\"tool_result\"}],\"role\":\"user\"}],\"model\":\"claude-sonnet-4-6\",\"system\":[{\"text\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\u003ccritical_rules\\u003e\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default \\u003c4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\u003c/critical_rules\\u003e\\n\\n\\u003ccommunication_style\\u003e\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\u003c/communication_style\\u003e\\n\\n\\u003ccode_references\\u003e\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\u003c/code_references\\u003e\\n\\n\\u003cworkflow\\u003e\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\u003c/workflow\\u003e\\n\\n\\u003cdecision_making\\u003e\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\u003c/decision_making\\u003e\\n\\n\\u003cediting_files\\u003e\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\u003c/editing_files\\u003e\\n\\n\\u003cwhitespace_and_exact_matching\\u003e\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\u003c/whitespace_and_exact_matching\\u003e\\n\\n\\u003ctask_completion\\u003e\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\u003c/task_completion\\u003e\\n\\n\\u003cerror_handling\\u003e\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\u003c/error_handling\\u003e\\n\\n\\u003cmemory_instructions\\u003e\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\u003c/memory_instructions\\u003e\\n\\n\\u003ccode_conventions\\u003e\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\u003c/code_conventions\\u003e\\n\\n\\u003ctesting\\u003e\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\u003c/testing\\u003e\\n\\n\\u003ctool_usage\\u003e\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\u003cbash_commands\\u003e\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `\\u0026` for background processes that won't stop on their own (e.g., `node server.js \\u0026`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status \\u0026\\u0026 git diff HEAD \\u0026\\u0026 git log -n 3`)\\n\\u003c/bash_commands\\u003e\\n\\u003c/tool_usage\\u003e\\n\\n\\u003cproactiveness\\u003e\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\u003c/proactiveness\\u003e\\n\\n\\u003cfinal_answers\\u003e\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\u003c/final_answers\\u003e\\n\\n\\u003cenv\\u003e\\nWorking directory: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/update_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\u003c/env\\u003e\\n\\n\\n\\n\\n\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"tool_choice\":{\"disable_parallel_tool_use\":false,\"type\":\"auto\"},\"tools\":[{\"input_schema\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"},\"name\":\"bash\",\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\u003ccross_platform\\u003e\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\u003c/cross_platform\\u003e\\n\\n\\u003cexecution_steps\\u003e\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with \\u003ccwd\\u003e\\u003c/cwd\\u003e tags\\n\\u003c/execution_steps\\u003e\\n\\n\\u003cusage_notes\\u003e\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '\\u0026\\u0026', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\u003c/usage_notes\\u003e\\n\\n\\u003cbackground_execution\\u003e\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `\\u0026` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\u003c/background_execution\\u003e\\n\\n\\u003cgit_commits\\u003e\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in \\u003ccommit_analysis\\u003e tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\u003ccrush@charm.land\\u003e\\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\u003c/git_commits\\u003e\\n\\n\\u003cpull_requests\\u003e\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in \\u003cpr_analysis\\u003e tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n\\n ## Summary\\n\\n \\u0026lt;1-3 bullet points\\u003e\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\u003c/pull_requests\\u003e\\n\\n\\u003cexamples\\u003e\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar \\u0026\\u0026 pytest tests\\n\\u003c/examples\\u003e\\n\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"},\"name\":\"download\",\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"},\"name\":\"edit\",\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\"},{\"input_schema\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"},\"name\":\"multiedit\",\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\"},{\"input_schema\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"},\"name\":\"fetch\",\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\"},{\"input_schema\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"glob\",\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\"},{\"input_schema\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"grep\",\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\"},{\"input_schema\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"},\"name\":\"ls\",\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\"},{\"input_schema\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"},\"name\":\"sourcegraph\",\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"},\"name\":\"view\",\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\"},{\"input_schema\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"},\"name\":\"write\",\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-sonnet-4-6","id":"msg_015HMzdwfeqyVqJaz34efKzH","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":1,"cache_creation_input_tokens":88,"cache_read_input_tokens":8283,"cache_creation":{"ephemeral_5m_input_tokens":88,"ephemeral_1h_input_tokens":0},"output_tokens":57,"service_tier":"standard","inference_geo":"global"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"tool_use","id":"toolu_01RAYHD98P3FQq8dMgNoY4gA","name":"view","input":{},"caller":{"type":"direct"}} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"{\"file_pat"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"h\": \"/t"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"mp"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"/cru"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"sh-"}}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"test/TestC"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"oderAgent/a"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"nthrop"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"ic-son"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"net/update_"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"a_file"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"/main.go\"}"} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0}
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"tool_use","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":1,"cache_creation_input_tokens":88,"cache_read_input_tokens":8283,"output_tokens":79} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 1.872657625s
-- id: 3
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31872
- host: ""
- body: "{\"max_tokens\":10000,\"messages\":[{\"content\":[{\"text\":\"\\u003csystem_reminder\\u003eThis is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\\u003c/system_reminder\\u003e\",\"type\":\"text\"},{\"text\":\"update the main.go file by changing the print to say hello from crush\",\"type\":\"text\"}],\"role\":\"user\"},{\"content\":[{\"id\":\"toolu_01C5th3Xy2LJY4R51J9ATgpd\",\"input\":{},\"name\":\"ls\",\"type\":\"tool_use\"}],\"role\":\"assistant\"},{\"content\":[{\"tool_use_id\":\"toolu_01C5th3Xy2LJY4R51J9ATgpd\",\"content\":[{\"text\":\"\\n- /tmp/crush-test/TestCoderAgent/anthropic-sonnet/update_a_file/\\n - go.mod\\n - main.go\\n\",\"type\":\"text\"}],\"type\":\"tool_result\"}],\"role\":\"user\"},{\"content\":[{\"id\":\"toolu_01RAYHD98P3FQq8dMgNoY4gA\",\"input\":{\"file_path\":\"/tmp/crush-test/TestCoderAgent/anthropic-sonnet/update_a_file/main.go\"},\"name\":\"view\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"tool_use\"}],\"role\":\"assistant\"},{\"content\":[{\"tool_use_id\":\"toolu_01RAYHD98P3FQq8dMgNoY4gA\",\"cache_control\":{\"type\":\"ephemeral\"},\"content\":[{\"text\":\"\\u003cfile\\u003e\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\tfmt.Println(\\\"Hello, World!\\\")\\n 7|}\\n\\u003c/file\\u003e\\n\",\"type\":\"text\"}],\"type\":\"tool_result\"}],\"role\":\"user\"}],\"model\":\"claude-sonnet-4-6\",\"system\":[{\"text\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\u003ccritical_rules\\u003e\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default \\u003c4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\u003c/critical_rules\\u003e\\n\\n\\u003ccommunication_style\\u003e\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\u003c/communication_style\\u003e\\n\\n\\u003ccode_references\\u003e\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\u003c/code_references\\u003e\\n\\n\\u003cworkflow\\u003e\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\u003c/workflow\\u003e\\n\\n\\u003cdecision_making\\u003e\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\u003c/decision_making\\u003e\\n\\n\\u003cediting_files\\u003e\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\u003c/editing_files\\u003e\\n\\n\\u003cwhitespace_and_exact_matching\\u003e\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\u003c/whitespace_and_exact_matching\\u003e\\n\\n\\u003ctask_completion\\u003e\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\u003c/task_completion\\u003e\\n\\n\\u003cerror_handling\\u003e\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\u003c/error_handling\\u003e\\n\\n\\u003cmemory_instructions\\u003e\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\u003c/memory_instructions\\u003e\\n\\n\\u003ccode_conventions\\u003e\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\u003c/code_conventions\\u003e\\n\\n\\u003ctesting\\u003e\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\u003c/testing\\u003e\\n\\n\\u003ctool_usage\\u003e\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\u003cbash_commands\\u003e\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `\\u0026` for background processes that won't stop on their own (e.g., `node server.js \\u0026`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status \\u0026\\u0026 git diff HEAD \\u0026\\u0026 git log -n 3`)\\n\\u003c/bash_commands\\u003e\\n\\u003c/tool_usage\\u003e\\n\\n\\u003cproactiveness\\u003e\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\u003c/proactiveness\\u003e\\n\\n\\u003cfinal_answers\\u003e\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\u003c/final_answers\\u003e\\n\\n\\u003cenv\\u003e\\nWorking directory: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/update_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\u003c/env\\u003e\\n\\n\\n\\n\\n\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"tool_choice\":{\"disable_parallel_tool_use\":false,\"type\":\"auto\"},\"tools\":[{\"input_schema\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"},\"name\":\"bash\",\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\u003ccross_platform\\u003e\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\u003c/cross_platform\\u003e\\n\\n\\u003cexecution_steps\\u003e\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with \\u003ccwd\\u003e\\u003c/cwd\\u003e tags\\n\\u003c/execution_steps\\u003e\\n\\n\\u003cusage_notes\\u003e\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '\\u0026\\u0026', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\u003c/usage_notes\\u003e\\n\\n\\u003cbackground_execution\\u003e\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `\\u0026` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\u003c/background_execution\\u003e\\n\\n\\u003cgit_commits\\u003e\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in \\u003ccommit_analysis\\u003e tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\u003ccrush@charm.land\\u003e\\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\u003c/git_commits\\u003e\\n\\n\\u003cpull_requests\\u003e\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in \\u003cpr_analysis\\u003e tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n\\n ## Summary\\n\\n \\u0026lt;1-3 bullet points\\u003e\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\u003c/pull_requests\\u003e\\n\\n\\u003cexamples\\u003e\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar \\u0026\\u0026 pytest tests\\n\\u003c/examples\\u003e\\n\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"},\"name\":\"download\",\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"},\"name\":\"edit\",\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\"},{\"input_schema\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"},\"name\":\"multiedit\",\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\"},{\"input_schema\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"},\"name\":\"fetch\",\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\"},{\"input_schema\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"glob\",\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\"},{\"input_schema\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"grep\",\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\"},{\"input_schema\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"},\"name\":\"ls\",\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\"},{\"input_schema\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"},\"name\":\"sourcegraph\",\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"},\"name\":\"view\",\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\"},{\"input_schema\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"},\"name\":\"write\",\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-sonnet-4-6","id":"msg_01RwxDGtPPuKs3u3Br7oSK87","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":1,"cache_creation_input_tokens":146,"cache_read_input_tokens":8371,"cache_creation":{"ephemeral_5m_input_tokens":146,"ephemeral_1h_input_tokens":0},"output_tokens":57,"service_tier":"standard","inference_geo":"global"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"tool_use","id":"toolu_01CeMWmSYU2GhDe3EebTJCke","name":"edit","input":{},"caller":{"type":"direct"}} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"{\"file_pat"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"h\": \"/tmp"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"/cru"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"sh-"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"test/TestCo"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"der"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"Agent/anth"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"ropic-sonn"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"et/upda"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"te_a_file"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"/ma"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"in.go\""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":", \"old_"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"string"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"\": \"\\tf"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"mt.Printl"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"n(\\\"He"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"llo, World!"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"\\\")\""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":", \"ne"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"w_string\""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":": \""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"\\tfmt.P"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"rintln(\\\"Hel"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"lo "} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"from"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":" Crush!"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"\\\")\"}"} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"tool_use","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":1,"cache_creation_input_tokens":146,"cache_read_input_tokens":8371,"output_tokens":136} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 2.1133855s
-- id: 4
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 32445
- host: ""
- body: "{\"max_tokens\":10000,\"messages\":[{\"content\":[{\"text\":\"\\u003csystem_reminder\\u003eThis is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\\u003c/system_reminder\\u003e\",\"type\":\"text\"},{\"text\":\"update the main.go file by changing the print to say hello from crush\",\"type\":\"text\"}],\"role\":\"user\"},{\"content\":[{\"id\":\"toolu_01C5th3Xy2LJY4R51J9ATgpd\",\"input\":{},\"name\":\"ls\",\"type\":\"tool_use\"}],\"role\":\"assistant\"},{\"content\":[{\"tool_use_id\":\"toolu_01C5th3Xy2LJY4R51J9ATgpd\",\"content\":[{\"text\":\"\\n- /tmp/crush-test/TestCoderAgent/anthropic-sonnet/update_a_file/\\n - go.mod\\n - main.go\\n\",\"type\":\"text\"}],\"type\":\"tool_result\"}],\"role\":\"user\"},{\"content\":[{\"id\":\"toolu_01RAYHD98P3FQq8dMgNoY4gA\",\"input\":{\"file_path\":\"/tmp/crush-test/TestCoderAgent/anthropic-sonnet/update_a_file/main.go\"},\"name\":\"view\",\"type\":\"tool_use\"}],\"role\":\"assistant\"},{\"content\":[{\"tool_use_id\":\"toolu_01RAYHD98P3FQq8dMgNoY4gA\",\"content\":[{\"text\":\"\\u003cfile\\u003e\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\tfmt.Println(\\\"Hello, World!\\\")\\n 7|}\\n\\u003c/file\\u003e\\n\",\"type\":\"text\"}],\"type\":\"tool_result\"}],\"role\":\"user\"},{\"content\":[{\"id\":\"toolu_01CeMWmSYU2GhDe3EebTJCke\",\"input\":{\"file_path\":\"/tmp/crush-test/TestCoderAgent/anthropic-sonnet/update_a_file/main.go\",\"new_string\":\"\\tfmt.Println(\\\"Hello from Crush!\\\")\",\"old_string\":\"\\tfmt.Println(\\\"Hello, World!\\\")\"},\"name\":\"edit\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"tool_use\"}],\"role\":\"assistant\"},{\"content\":[{\"tool_use_id\":\"toolu_01CeMWmSYU2GhDe3EebTJCke\",\"cache_control\":{\"type\":\"ephemeral\"},\"content\":[{\"text\":\"\\u003cresult\\u003e\\nContent replaced in file: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/update_a_file/main.go\\n\\u003c/result\\u003e\\n\",\"type\":\"text\"}],\"type\":\"tool_result\"}],\"role\":\"user\"}],\"model\":\"claude-sonnet-4-6\",\"system\":[{\"text\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\u003ccritical_rules\\u003e\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default \\u003c4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\u003c/critical_rules\\u003e\\n\\n\\u003ccommunication_style\\u003e\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\u003c/communication_style\\u003e\\n\\n\\u003ccode_references\\u003e\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\u003c/code_references\\u003e\\n\\n\\u003cworkflow\\u003e\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\u003c/workflow\\u003e\\n\\n\\u003cdecision_making\\u003e\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\u003c/decision_making\\u003e\\n\\n\\u003cediting_files\\u003e\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\u003c/editing_files\\u003e\\n\\n\\u003cwhitespace_and_exact_matching\\u003e\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\u003c/whitespace_and_exact_matching\\u003e\\n\\n\\u003ctask_completion\\u003e\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\u003c/task_completion\\u003e\\n\\n\\u003cerror_handling\\u003e\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\u003c/error_handling\\u003e\\n\\n\\u003cmemory_instructions\\u003e\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\u003c/memory_instructions\\u003e\\n\\n\\u003ccode_conventions\\u003e\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\u003c/code_conventions\\u003e\\n\\n\\u003ctesting\\u003e\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\u003c/testing\\u003e\\n\\n\\u003ctool_usage\\u003e\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\u003cbash_commands\\u003e\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `\\u0026` for background processes that won't stop on their own (e.g., `node server.js \\u0026`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status \\u0026\\u0026 git diff HEAD \\u0026\\u0026 git log -n 3`)\\n\\u003c/bash_commands\\u003e\\n\\u003c/tool_usage\\u003e\\n\\n\\u003cproactiveness\\u003e\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\u003c/proactiveness\\u003e\\n\\n\\u003cfinal_answers\\u003e\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\u003c/final_answers\\u003e\\n\\n\\u003cenv\\u003e\\nWorking directory: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/update_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\u003c/env\\u003e\\n\\n\\n\\n\\n\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"tool_choice\":{\"disable_parallel_tool_use\":false,\"type\":\"auto\"},\"tools\":[{\"input_schema\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"},\"name\":\"bash\",\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\u003ccross_platform\\u003e\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\u003c/cross_platform\\u003e\\n\\n\\u003cexecution_steps\\u003e\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with \\u003ccwd\\u003e\\u003c/cwd\\u003e tags\\n\\u003c/execution_steps\\u003e\\n\\n\\u003cusage_notes\\u003e\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '\\u0026\\u0026', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\u003c/usage_notes\\u003e\\n\\n\\u003cbackground_execution\\u003e\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `\\u0026` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\u003c/background_execution\\u003e\\n\\n\\u003cgit_commits\\u003e\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in \\u003ccommit_analysis\\u003e tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\u003ccrush@charm.land\\u003e\\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\u003c/git_commits\\u003e\\n\\n\\u003cpull_requests\\u003e\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in \\u003cpr_analysis\\u003e tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n\\n ## Summary\\n\\n \\u0026lt;1-3 bullet points\\u003e\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\u003c/pull_requests\\u003e\\n\\n\\u003cexamples\\u003e\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar \\u0026\\u0026 pytest tests\\n\\u003c/examples\\u003e\\n\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"},\"name\":\"download\",\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"},\"name\":\"edit\",\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\"},{\"input_schema\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"},\"name\":\"multiedit\",\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\"},{\"input_schema\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"},\"name\":\"fetch\",\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\"},{\"input_schema\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"glob\",\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\"},{\"input_schema\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"grep\",\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\"},{\"input_schema\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"},\"name\":\"ls\",\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\"},{\"input_schema\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"},\"name\":\"sourcegraph\",\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"},\"name\":\"view\",\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\"},{\"input_schema\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"},\"name\":\"write\",\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-sonnet-4-6","id":"msg_01TajZeTPcAgjrC4MHQKoJD3","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":1,"cache_creation_input_tokens":188,"cache_read_input_tokens":8517,"cache_creation":{"ephemeral_5m_input_tokens":188,"ephemeral_1h_input_tokens":0},"output_tokens":2,"service_tier":"standard","inference_geo":"global"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"Done."} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":1,"cache_creation_input_tokens":188,"cache_read_input_tokens":8517,"output_tokens":5} }
-
- event: message_stop
- data: {"type":"message_stop"}
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 1.8818025s
diff --git a/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/write_tool.yaml b/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/write_tool.yaml
deleted file mode 100644
index 3bf40260fc5857b21d934a6202cd1064ec30e0bb..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/anthropic-sonnet/write_tool.yaml
+++ /dev/null
@@ -1,225 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 901
- host: ""
- body: '{"max_tokens":40,"messages":[{"content":[{"text":"Generate a concise title for the following content:\n\nuse write to create a new file called config.json with content ''{\"name\": \"test\", \"version\": \"1.0.0\"}''\n \u003cthink\u003e\n\n\u003c/think\u003e","type":"text"}],"role":"user"}],"model":"claude-haiku-4-5-20251001","system":[{"text":"You will generate a short title based on the first message a user begins a conversation with.\n\n\u003crules\u003e\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\u003c/rules\u003e\n\n /no_think","type":"text"}],"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-haiku-4-5-20251001","id":"msg_01B97T62rPotpjpBoqwYEGhV","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":186,"cache_creation_input_tokens":0,"cache_read_input_tokens":0,"cache_creation":{"ephemeral_5m_input_tokens":0,"ephemeral_1h_input_tokens":0},"output_tokens":1,"service_tier":"standard","inference_geo":"not_available"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"Create"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" config.json file with JSON content"} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":186,"cache_creation_input_tokens":0,"cache_read_input_tokens":0,"output_tokens":11} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 1.171629833s
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31059
- host: ""
- body: "{\"max_tokens\":10000,\"messages\":[{\"content\":[{\"text\":\"\\u003csystem_reminder\\u003eThis is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\\u003c/system_reminder\\u003e\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"},{\"text\":\"use write to create a new file called config.json with content '{\\\"name\\\": \\\"test\\\", \\\"version\\\": \\\"1.0.0\\\"}'\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"role\":\"user\"}],\"model\":\"claude-sonnet-4-6\",\"system\":[{\"text\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\u003ccritical_rules\\u003e\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default \\u003c4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\u003c/critical_rules\\u003e\\n\\n\\u003ccommunication_style\\u003e\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\u003c/communication_style\\u003e\\n\\n\\u003ccode_references\\u003e\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\u003c/code_references\\u003e\\n\\n\\u003cworkflow\\u003e\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\u003c/workflow\\u003e\\n\\n\\u003cdecision_making\\u003e\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\u003c/decision_making\\u003e\\n\\n\\u003cediting_files\\u003e\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\u003c/editing_files\\u003e\\n\\n\\u003cwhitespace_and_exact_matching\\u003e\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\u003c/whitespace_and_exact_matching\\u003e\\n\\n\\u003ctask_completion\\u003e\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\u003c/task_completion\\u003e\\n\\n\\u003cerror_handling\\u003e\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\u003c/error_handling\\u003e\\n\\n\\u003cmemory_instructions\\u003e\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\u003c/memory_instructions\\u003e\\n\\n\\u003ccode_conventions\\u003e\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\u003c/code_conventions\\u003e\\n\\n\\u003ctesting\\u003e\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\u003c/testing\\u003e\\n\\n\\u003ctool_usage\\u003e\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\u003cbash_commands\\u003e\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `\\u0026` for background processes that won't stop on their own (e.g., `node server.js \\u0026`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status \\u0026\\u0026 git diff HEAD \\u0026\\u0026 git log -n 3`)\\n\\u003c/bash_commands\\u003e\\n\\u003c/tool_usage\\u003e\\n\\n\\u003cproactiveness\\u003e\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\u003c/proactiveness\\u003e\\n\\n\\u003cfinal_answers\\u003e\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\u003c/final_answers\\u003e\\n\\n\\u003cenv\\u003e\\nWorking directory: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/write_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\u003c/env\\u003e\\n\\n\\n\\n\\n\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"tool_choice\":{\"disable_parallel_tool_use\":false,\"type\":\"auto\"},\"tools\":[{\"input_schema\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"},\"name\":\"bash\",\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\u003ccross_platform\\u003e\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\u003c/cross_platform\\u003e\\n\\n\\u003cexecution_steps\\u003e\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with \\u003ccwd\\u003e\\u003c/cwd\\u003e tags\\n\\u003c/execution_steps\\u003e\\n\\n\\u003cusage_notes\\u003e\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '\\u0026\\u0026', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\u003c/usage_notes\\u003e\\n\\n\\u003cbackground_execution\\u003e\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `\\u0026` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\u003c/background_execution\\u003e\\n\\n\\u003cgit_commits\\u003e\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in \\u003ccommit_analysis\\u003e tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\u003ccrush@charm.land\\u003e\\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\u003c/git_commits\\u003e\\n\\n\\u003cpull_requests\\u003e\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in \\u003cpr_analysis\\u003e tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n\\n ## Summary\\n\\n \\u0026lt;1-3 bullet points\\u003e\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\u003c/pull_requests\\u003e\\n\\n\\u003cexamples\\u003e\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar \\u0026\\u0026 pytest tests\\n\\u003c/examples\\u003e\\n\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"},\"name\":\"download\",\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"},\"name\":\"edit\",\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\"},{\"input_schema\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"},\"name\":\"multiedit\",\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\"},{\"input_schema\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"},\"name\":\"fetch\",\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\"},{\"input_schema\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"glob\",\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\"},{\"input_schema\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"grep\",\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\"},{\"input_schema\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"},\"name\":\"ls\",\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\"},{\"input_schema\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"},\"name\":\"sourcegraph\",\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"},\"name\":\"view\",\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\"},{\"input_schema\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"},\"name\":\"write\",\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-sonnet-4-6","id":"msg_019PB81kZWdCFxjyvg91obPF","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":3,"cache_creation_input_tokens":4564,"cache_read_input_tokens":3733,"cache_creation":{"ephemeral_5m_input_tokens":4564,"ephemeral_1h_input_tokens":0},"output_tokens":62,"service_tier":"standard","inference_geo":"global"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"tool_use","id":"toolu_01SuvF9FfC2Gpa6z3tLZJvQf","name":"write","input":{},"caller":{"type":"direct"}} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"{\"file_path"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"\": \""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"/tmp/crush"}}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"-test/Test"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"Cod"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"erAgent/anth"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"ropic-"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"sonnet/wr"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"ite_t"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"ool/config."} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"js"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"on\""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":", \"conten"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"t\""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":": \"{\\\"na"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"me\\\": \\\"te"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"st\\"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"\","}}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":" \\\"versio"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"n\\\": \\\"1"} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":".0.0\\\""} }
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":"}\"}"} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0 }
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"tool_use","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":3,"cache_creation_input_tokens":4564,"cache_read_input_tokens":3733,"output_tokens":110} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 1.613104167s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31591
- host: ""
- body: "{\"max_tokens\":10000,\"messages\":[{\"content\":[{\"text\":\"\\u003csystem_reminder\\u003eThis is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\\u003c/system_reminder\\u003e\",\"type\":\"text\"},{\"text\":\"use write to create a new file called config.json with content '{\\\"name\\\": \\\"test\\\", \\\"version\\\": \\\"1.0.0\\\"}'\",\"type\":\"text\"}],\"role\":\"user\"},{\"content\":[{\"id\":\"toolu_01SuvF9FfC2Gpa6z3tLZJvQf\",\"input\":{\"content\":\"{\\\"name\\\": \\\"test\\\", \\\"version\\\": \\\"1.0.0\\\"}\",\"file_path\":\"/tmp/crush-test/TestCoderAgent/anthropic-sonnet/write_tool/config.json\"},\"name\":\"write\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"tool_use\"}],\"role\":\"assistant\"},{\"content\":[{\"tool_use_id\":\"toolu_01SuvF9FfC2Gpa6z3tLZJvQf\",\"cache_control\":{\"type\":\"ephemeral\"},\"content\":[{\"text\":\"\\u003cresult\\u003e\\nFile successfully written: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/write_tool/config.json\\n\\u003c/result\\u003e\",\"type\":\"text\"}],\"type\":\"tool_result\"}],\"role\":\"user\"}],\"model\":\"claude-sonnet-4-6\",\"system\":[{\"text\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\u003ccritical_rules\\u003e\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default \\u003c4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\u003c/critical_rules\\u003e\\n\\n\\u003ccommunication_style\\u003e\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\u003c/communication_style\\u003e\\n\\n\\u003ccode_references\\u003e\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\u003c/code_references\\u003e\\n\\n\\u003cworkflow\\u003e\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\u003c/workflow\\u003e\\n\\n\\u003cdecision_making\\u003e\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\u003c/decision_making\\u003e\\n\\n\\u003cediting_files\\u003e\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\u003c/editing_files\\u003e\\n\\n\\u003cwhitespace_and_exact_matching\\u003e\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\u003c/whitespace_and_exact_matching\\u003e\\n\\n\\u003ctask_completion\\u003e\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\u003c/task_completion\\u003e\\n\\n\\u003cerror_handling\\u003e\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\u003c/error_handling\\u003e\\n\\n\\u003cmemory_instructions\\u003e\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\u003c/memory_instructions\\u003e\\n\\n\\u003ccode_conventions\\u003e\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\u003c/code_conventions\\u003e\\n\\n\\u003ctesting\\u003e\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\u003c/testing\\u003e\\n\\n\\u003ctool_usage\\u003e\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\u003cbash_commands\\u003e\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `\\u0026` for background processes that won't stop on their own (e.g., `node server.js \\u0026`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status \\u0026\\u0026 git diff HEAD \\u0026\\u0026 git log -n 3`)\\n\\u003c/bash_commands\\u003e\\n\\u003c/tool_usage\\u003e\\n\\n\\u003cproactiveness\\u003e\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\u003c/proactiveness\\u003e\\n\\n\\u003cfinal_answers\\u003e\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\u003c/final_answers\\u003e\\n\\n\\u003cenv\\u003e\\nWorking directory: /tmp/crush-test/TestCoderAgent/anthropic-sonnet/write_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\u003c/env\\u003e\\n\\n\\n\\n\\n\",\"cache_control\":{\"type\":\"ephemeral\"},\"type\":\"text\"}],\"tool_choice\":{\"disable_parallel_tool_use\":false,\"type\":\"auto\"},\"tools\":[{\"input_schema\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"},\"name\":\"bash\",\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\u003ccross_platform\\u003e\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\u003c/cross_platform\\u003e\\n\\n\\u003cexecution_steps\\u003e\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with \\u003ccwd\\u003e\\u003c/cwd\\u003e tags\\n\\u003c/execution_steps\\u003e\\n\\n\\u003cusage_notes\\u003e\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '\\u0026\\u0026', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\u003c/usage_notes\\u003e\\n\\n\\u003cbackground_execution\\u003e\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `\\u0026` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\u003c/background_execution\\u003e\\n\\n\\u003cgit_commits\\u003e\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in \\u003ccommit_analysis\\u003e tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\u003ccrush@charm.land\\u003e\\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\u003c/git_commits\\u003e\\n\\n\\u003cpull_requests\\u003e\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in \\u003cpr_analysis\\u003e tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat \\u0026lt;\\u0026lt;'EOF'\\n\\n ## Summary\\n\\n \\u0026lt;1-3 bullet points\\u003e\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\u003c/pull_requests\\u003e\\n\\n\\u003cexamples\\u003e\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar \\u0026\\u0026 pytest tests\\n\\u003c/examples\\u003e\\n\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"},\"name\":\"download\",\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"},\"name\":\"edit\",\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\"},{\"input_schema\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"},\"name\":\"multiedit\",\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\"},{\"input_schema\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"},\"name\":\"fetch\",\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\"},{\"input_schema\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"glob\",\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\"},{\"input_schema\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"},\"name\":\"grep\",\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\"},{\"input_schema\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"},\"name\":\"ls\",\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\"},{\"input_schema\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"},\"name\":\"sourcegraph\",\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\"},{\"input_schema\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"},\"name\":\"view\",\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\"},{\"input_schema\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"},\"name\":\"write\",\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.anthropic.com/v1/messages
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |+
- event: message_start
- data: {"type":"message_start","message":{"model":"claude-sonnet-4-6","id":"msg_01KbyWVQNvRbmVHrLecvci6u","type":"message","role":"assistant","content":[],"stop_reason":null,"stop_sequence":null,"stop_details":null,"usage":{"input_tokens":1,"cache_creation_input_tokens":159,"cache_read_input_tokens":8297,"cache_creation":{"ephemeral_5m_input_tokens":159,"ephemeral_1h_input_tokens":0},"output_tokens":2,"service_tier":"standard","inference_geo":"global"}} }
-
- event: content_block_start
- data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""} }
-
- event: ping
- data: {"type": "ping"}
-
- event: content_block_delta
- data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"Done."} }
-
- event: content_block_stop
- data: {"type":"content_block_stop","index":0}
-
- event: message_delta
- data: {"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null,"stop_details":null},"usage":{"input_tokens":1,"cache_creation_input_tokens":159,"cache_read_input_tokens":8297,"output_tokens":5} }
-
- event: message_stop
- data: {"type":"message_stop" }
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 1.390672625s
diff --git a/internal/agent/testdata/TestCoderAgent/openai-gpt-5/parallel_tool_calls.yaml b/internal/agent/testdata/TestCoderAgent/glm-5.1/bash_tool.yaml
similarity index 60%
rename from internal/agent/testdata/TestCoderAgent/openai-gpt-5/parallel_tool_calls.yaml
rename to internal/agent/testdata/TestCoderAgent/glm-5.1/bash_tool.yaml
index 5d2a2a8bd957555f1db5c8841271b367be48294b..6dd6cae227a7f7b97649746898778fe21c815458 100644
--- a/internal/agent/testdata/TestCoderAgent/openai-gpt-5/parallel_tool_calls.yaml
+++ b/internal/agent/testdata/TestCoderAgent/glm-5.1/bash_tool.yaml
@@ -6,9 +6,9 @@ interactions:
proto: HTTP/1.1
proto_major: 1
proto_minor: 1
- content_length: 876
+ content_length: 853
host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse glob to find all .go files and use ls to list the current directory, it is very important that you run both tool calls in parallel\n \n\n","role":"user"}],"model":"gpt-4o","max_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
+ body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse bash to create a file named test.txt with content ''hello bash''. do not print its timestamp\n \n\n","role":"user"}],"model":"gpt-oss-120b","max_completion_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
headers:
Accept:
- application/json
@@ -16,50 +16,55 @@ interactions:
- application/json
User-Agent:
- Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
+ url: https://hyper.charm.land/v1/chat/completions
method: POST
response:
proto: HTTP/2.0
proto_major: 2
proto_minor: 0
content_length: -1
+ uncompressed: true
body: |+
- data: {"id":"chatcmpl-DUu90yHFiqANxMC539sFFf3Ox94Lf","object":"chat.completion.chunk","created":1776258030,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"role":"assistant","content":"","refusal":null},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"P4Pcx2SHsL1Ht0"}
+ data: {"id":"chatcmpl-0d035875a345ee3e89e04348","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu90yHFiqANxMC539sFFf3Ox94Lf","object":"chat.completion.chunk","created":1776258030,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":"Running"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"I23AxfXaa"}
+ data: {"id":"chatcmpl-0d035875a345ee3e89e04348","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":"B"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu90yHFiqANxMC539sFFf3Ox94Lf","object":"chat.completion.chunk","created":1776258030,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":" Glob"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"ig87JsfJZjB"}
+ data: {"id":"chatcmpl-0d035875a345ee3e89e04348","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":"ash"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu90yHFiqANxMC539sFFf3Ox94Lf","object":"chat.completion.chunk","created":1776258030,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":" and"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"SyFictSzn3CU"}
+ data: {"id":"chatcmpl-0d035875a345ee3e89e04348","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" command"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu90yHFiqANxMC539sFFf3Ox94Lf","object":"chat.completion.chunk","created":1776258030,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":" LS"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"ZWMk2qQZPUUII"}
+ data: {"id":"chatcmpl-0d035875a345ee3e89e04348","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" to"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu90yHFiqANxMC539sFFf3Ox94Lf","object":"chat.completion.chunk","created":1776258030,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":" Commands"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"KsOhZOQ"}
+ data: {"id":"chatcmpl-0d035875a345ee3e89e04348","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" create"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu90yHFiqANxMC539sFFf3Ox94Lf","object":"chat.completion.chunk","created":1776258030,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":" in"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"QvBWJsGovNPVh"}
+ data: {"id":"chatcmpl-0d035875a345ee3e89e04348","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" test"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu90yHFiqANxMC539sFFf3Ox94Lf","object":"chat.completion.chunk","created":1776258030,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":" Parallel"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"iwptu0d"}
+ data: {"id":"chatcmpl-0d035875a345ee3e89e04348","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":".txt"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu90yHFiqANxMC539sFFf3Ox94Lf","object":"chat.completion.chunk","created":1776258030,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{},"logprobs":null,"finish_reason":"stop"}],"usage":null,"obfuscation":"0hKCRMiXL4"}
+ data: {"id":"chatcmpl-0d035875a345ee3e89e04348","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" with"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu90yHFiqANxMC539sFFf3Ox94Lf","object":"chat.completion.chunk","created":1776258030,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[],"usage":{"prompt_tokens":170,"completion_tokens":7,"total_tokens":177,"prompt_tokens_details":{"cached_tokens":0,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":0,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"XQUDNF0k7Xd1jN"}
+ data: {"id":"chatcmpl-0d035875a345ee3e89e04348","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" hello"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-0d035875a345ee3e89e04348","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" bash"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-0d035875a345ee3e89e04348","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{"prompt_tokens":43,"completion_tokens":238,"total_tokens":457}}
data: [DONE]
headers:
Content-Type:
- - text/event-stream; charset=utf-8
+ - text/event-stream
status: 200 OK
code: 200
- duration: 552.086291ms
+ duration: 2.910817083s
- id: 1
request:
proto: HTTP/1.1
proto_major: 1
proto_minor: 1
- content_length: 30753
+ content_length: 30689
host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/parallel_tool_calls\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use glob to find all .go files and use ls to list the current directory, it is very important that you run both tool calls in parallel\",\"role\":\"user\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
+ body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/glm-5.1/bash_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use bash to create a file named test.txt with content 'hello bash'. do not print its timestamp\",\"role\":\"user\"}],\"model\":\"glm-5.1\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
headers:
Accept:
- application/json
@@ -67,134 +72,134 @@ interactions:
- application/json
User-Agent:
- Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
+ url: https://hyper.charm.land/v1/chat/completions
method: POST
response:
proto: HTTP/2.0
proto_major: 2
proto_minor: 0
content_length: -1
+ uncompressed: true
body: |+
- data: {"id":"chatcmpl-DUu90iWyBNAEsyVhXKrnfY13ew9Bn","object":"chat.completion.chunk","created":1776258030,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"usage":null,"choices":[{"index":0,"delta":{"role":"assistant","content":null},"finish_reason":null}],"obfuscation":"WXiEdR1"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu90iWyBNAEsyVhXKrnfY13ew9Bn","object":"chat.completion.chunk","created":1776258030,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"usage":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"id":"call_XNJBmJ3UIKrBN8hBqcEaU9yu","type":"function","function":{"name":"glob","arguments":""}}]},"finish_reason":null}],"obfuscation":"4msvnFmf8MaHV"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_UpwQiAEBdqIY9WrpTH2xGyYP","type":"function","index":0,"function":{"name":"bash","arguments":""}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu90iWyBNAEsyVhXKrnfY13ew9Bn","object":"chat.completion.chunk","created":1776258030,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"usage":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\"pa"}}]},"finish_reason":null}],"obfuscation":"A4CIByP5UkeIL"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu90iWyBNAEsyVhXKrnfY13ew9Bn","object":"chat.completion.chunk","created":1776258030,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"usage":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ttern"}}]},"finish_reason":null}],"obfuscation":"Op9k4DDNls82Z"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"command\": \""}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu90iWyBNAEsyVhXKrnfY13ew9Bn","object":"chat.completion.chunk","created":1776258030,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"usage":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\": \"**"}}]},"finish_reason":null}],"obfuscation":"N4RJs5vdjp"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"echo"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu90iWyBNAEsyVhXKrnfY13ew9Bn","object":"chat.completion.chunk","created":1776258030,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"usage":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/*.g"}}]},"finish_reason":null}],"obfuscation":"Rl7YAPBXak21mX"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" -"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu90iWyBNAEsyVhXKrnfY13ew9Bn","object":"chat.completion.chunk","created":1776258030,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"usage":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"o\"}"}}]},"finish_reason":null}],"obfuscation":"MpeQhbgjVibTPl"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"n"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu90iWyBNAEsyVhXKrnfY13ew9Bn","object":"chat.completion.chunk","created":1776258030,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"usage":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":1,"id":"call_3gjn2wDOLXX3t9EvrY16cw5N","type":"function","function":{"name":"ls","arguments":""}}]},"finish_reason":null}],"obfuscation":"SlVk8YOOU9WFtJl"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" '"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu90iWyBNAEsyVhXKrnfY13ew9Bn","object":"chat.completion.chunk","created":1776258030,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"usage":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":1,"function":{"arguments":"{\"de"}}]},"finish_reason":null}],"obfuscation":"WXdl07EcpAIB8"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"hello"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu90iWyBNAEsyVhXKrnfY13ew9Bn","object":"chat.completion.chunk","created":1776258030,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"usage":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":1,"function":{"arguments":"pth\":"}}]},"finish_reason":null}],"obfuscation":"9DPCoH8CEbc4"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" bash"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu90iWyBNAEsyVhXKrnfY13ew9Bn","object":"chat.completion.chunk","created":1776258030,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"usage":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":1,"function":{"arguments":" 1}"}}]},"finish_reason":null}],"obfuscation":"FF46aVe50QICdwt"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"'"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu90iWyBNAEsyVhXKrnfY13ew9Bn","object":"chat.completion.chunk","created":1776258030,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":null,"obfuscation":""}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" \u003e"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu90iWyBNAEsyVhXKrnfY13ew9Bn","object":"chat.completion.chunk","created":1776258030,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6395,"completion_tokens":630,"total_tokens":7025,"prompt_tokens_details":{"cached_tokens":3456,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":576,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"T"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" /"}}]},"finish_reason":null}]}
- data: [DONE]
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"tmp"}}]},"finish_reason":null}]}
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 12.959898583s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31343
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/parallel_tool_calls\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use glob to find all .go files and use ls to list the current directory, it is very important that you run both tool calls in parallel\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_XNJBmJ3UIKrBN8hBqcEaU9yu\",\"function\":{\"arguments\":\"{\\\"pattern\\\": \\\"**/*.go\\\"}\",\"name\":\"glob\"},\"type\":\"function\"},{\"id\":\"call_3gjn2wDOLXX3t9EvrY16cw5N\",\"function\":{\"arguments\":\"{\\\"depth\\\": 1}\",\"name\":\"ls\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"/tmp/crush-test/TestCoderAgent/openai-gpt-5/parallel_tool_calls/main.go\",\"tool_call_id\":\"call_XNJBmJ3UIKrBN8hBqcEaU9yu\",\"role\":\"tool\"},{\"content\":\"\\n- /tmp/crush-test/TestCoderAgent/openai-gpt-5/parallel_tool_calls/\\n - go.mod\\n - main.go\\n\",\"tool_call_id\":\"call_3gjn2wDOLXX3t9EvrY16cw5N\",\"role\":\"tool\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu9DRispUUiqJA9NI0fMzXJsCpva","object":"chat.completion.chunk","created":1776258043,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":"","refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"sY5nn2Yt4Z"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/c"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu9DRispUUiqJA9NI0fMzXJsCpva","object":"chat.completion.chunk","created":1776258043,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/"},"finish_reason":null}],"usage":null,"obfuscation":"5OyZivpf7xH"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"rush"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu9DRispUUiqJA9NI0fMzXJsCpva","object":"chat.completion.chunk","created":1776258043,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"tmp"},"finish_reason":null}],"usage":null,"obfuscation":"xsxD7RctN"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-test"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu9DRispUUiqJA9NI0fMzXJsCpva","object":"chat.completion.chunk","created":1776258043,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/cr"},"finish_reason":null}],"usage":null,"obfuscation":"5JYaZia0Q"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/Test"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu9DRispUUiqJA9NI0fMzXJsCpva","object":"chat.completion.chunk","created":1776258043,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"ush"},"finish_reason":null}],"usage":null,"obfuscation":"BrFLhbZJQ"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Coder"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu9DRispUUiqJA9NI0fMzXJsCpva","object":"chat.completion.chunk","created":1776258043,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"-test"},"finish_reason":null}],"usage":null,"obfuscation":"r6sYdxA"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Agent"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu9DRispUUiqJA9NI0fMzXJsCpva","object":"chat.completion.chunk","created":1776258043,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/Test"},"finish_reason":null}],"usage":null,"obfuscation":"YyAO7cQ"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/g"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu9DRispUUiqJA9NI0fMzXJsCpva","object":"chat.completion.chunk","created":1776258043,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"Coder"},"finish_reason":null}],"usage":null,"obfuscation":"0smHvsS"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"lm"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu9DRispUUiqJA9NI0fMzXJsCpva","object":"chat.completion.chunk","created":1776258043,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"Agent"},"finish_reason":null}],"usage":null,"obfuscation":"RSOHhpJ"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu9DRispUUiqJA9NI0fMzXJsCpva","object":"chat.completion.chunk","created":1776258043,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/open"},"finish_reason":null}],"usage":null,"obfuscation":"haIrX95"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"5"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu9DRispUUiqJA9NI0fMzXJsCpva","object":"chat.completion.chunk","created":1776258043,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"ai"},"finish_reason":null}],"usage":null,"obfuscation":"OBokp8WYtE"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"."}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu9DRispUUiqJA9NI0fMzXJsCpva","object":"chat.completion.chunk","created":1776258043,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"-g"},"finish_reason":null}],"usage":null,"obfuscation":"N1VeHzoh32"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"1"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu9DRispUUiqJA9NI0fMzXJsCpva","object":"chat.completion.chunk","created":1776258043,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"pt"},"finish_reason":null}],"usage":null,"obfuscation":"JWj8yG75f4"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/bash"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu9DRispUUiqJA9NI0fMzXJsCpva","object":"chat.completion.chunk","created":1776258043,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"-"},"finish_reason":null}],"usage":null,"obfuscation":"ZezhCBfxYKc"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_tool"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu9DRispUUiqJA9NI0fMzXJsCpva","object":"chat.completion.chunk","created":1776258043,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"5"},"finish_reason":null}],"usage":null,"obfuscation":"LVb33OvphXD"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/test"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu9DRispUUiqJA9NI0fMzXJsCpva","object":"chat.completion.chunk","created":1776258043,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/"},"finish_reason":null}],"usage":null,"obfuscation":"kKRCwYC4K8g"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".txt"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu9DRispUUiqJA9NI0fMzXJsCpva","object":"chat.completion.chunk","created":1776258043,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"parallel"},"finish_reason":null}],"usage":null,"obfuscation":"bXhY"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\", \"description\": \""}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu9DRispUUiqJA9NI0fMzXJsCpva","object":"chat.completion.chunk","created":1776258043,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"_tool"},"finish_reason":null}],"usage":null,"obfuscation":"wlbfkkG"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Create"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu9DRispUUiqJA9NI0fMzXJsCpva","object":"chat.completion.chunk","created":1776258043,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"_calls"},"finish_reason":null}],"usage":null,"obfuscation":"6bKDoo"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" test"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu9DRispUUiqJA9NI0fMzXJsCpva","object":"chat.completion.chunk","created":1776258043,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/main"},"finish_reason":null}],"usage":null,"obfuscation":"ZUqsC3c"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".txt"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu9DRispUUiqJA9NI0fMzXJsCpva","object":"chat.completion.chunk","created":1776258043,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":".go"},"finish_reason":null}],"usage":null,"obfuscation":"phOHnCBOV"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" with"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu9DRispUUiqJA9NI0fMzXJsCpva","object":"chat.completion.chunk","created":1776258043,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"\n"},"finish_reason":null}],"usage":null,"obfuscation":"jbPMzETkpH"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" content"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu9DRispUUiqJA9NI0fMzXJsCpva","object":"chat.completion.chunk","created":1776258043,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"go"},"finish_reason":null}],"usage":null,"obfuscation":"qyEqmFolU6"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"}"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu9DRispUUiqJA9NI0fMzXJsCpva","object":"chat.completion.chunk","created":1776258043,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":".mod"},"finish_reason":null}],"usage":null,"obfuscation":"F23Kd6aU"}
+ data: {"id":"chatcmpl-778f668283f7c6cab8f8ed46","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":{"prompt_tokens":166,"completion_tokens":56,"total_tokens":7134}}
- data: {"id":"chatcmpl-DUu9DRispUUiqJA9NI0fMzXJsCpva","object":"chat.completion.chunk","created":1776258043,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":","},"finish_reason":null}],"usage":null,"obfuscation":"7keRgKP2QUT"}
-
- data: {"id":"chatcmpl-DUu9DRispUUiqJA9NI0fMzXJsCpva","object":"chat.completion.chunk","created":1776258043,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":" main"},"finish_reason":null}],"usage":null,"obfuscation":"Howv0S1"}
+ data: [DONE]
- data: {"id":"chatcmpl-DUu9DRispUUiqJA9NI0fMzXJsCpva","object":"chat.completion.chunk","created":1776258043,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":".go"},"finish_reason":null}],"usage":null,"obfuscation":"cEpKMFHnm"}
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 3.881508541s
+- id: 2
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 31059
+ host: ""
+ body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/glm-5.1/bash_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use bash to create a file named test.txt with content 'hello bash'. do not print its timestamp\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_UpwQiAEBdqIY9WrpTH2xGyYP\",\"function\":{\"arguments\":\"{\\\"command\\\": \\\"echo -n 'hello bash' > /tmp/crush-test/TestCoderAgent/glm-5.1/bash_tool/test.txt\\\", \\\"description\\\": \\\"Create test.txt with content\\\"}\",\"name\":\"bash\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"no output\",\"tool_call_id\":\"call_UpwQiAEBdqIY9WrpTH2xGyYP\",\"role\":\"tool\"}],\"model\":\"glm-5.1\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
+ headers:
+ Accept:
+ - application/json
+ Content-Type:
+ - application/json
+ User-Agent:
+ - Charm-Crush/devel (https://charm.land/crush)
+ url: https://hyper.charm.land/v1/chat/completions
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |+
+ data: {"id":"chatcmpl-0f87d56b24e7cd9fe0e7aabc","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu9DRispUUiqJA9NI0fMzXJsCpva","object":"chat.completion.chunk","created":1776258043,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":null,"obfuscation":"fLCgGp"}
+ data: {"id":"chatcmpl-0f87d56b24e7cd9fe0e7aabc","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"Done."},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu9DRispUUiqJA9NI0fMzXJsCpva","object":"chat.completion.chunk","created":1776258043,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6512,"completion_tokens":29,"total_tokens":6541,"prompt_tokens_details":{"cached_tokens":6272,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":0,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"knl0"}
+ data: {"id":"chatcmpl-0f87d56b24e7cd9fe0e7aabc","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{"prompt_tokens":7099,"completion_tokens":3,"total_tokens":7134}}
data: [DONE]
headers:
Content-Type:
- - text/event-stream; charset=utf-8
+ - text/event-stream
status: 200 OK
code: 200
- duration: 789.747792ms
+ duration: 1.363214542s
diff --git a/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/write_tool.yaml b/internal/agent/testdata/TestCoderAgent/glm-5.1/download_tool.yaml
similarity index 56%
rename from internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/write_tool.yaml
rename to internal/agent/testdata/TestCoderAgent/glm-5.1/download_tool.yaml
index 8d03393077a4347d2ec0b55165ac0f35efa7d5ca..ef73614a7a07e17411f326510aeee579bbed0735 100644
--- a/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/write_tool.yaml
+++ b/internal/agent/testdata/TestCoderAgent/glm-5.1/download_tool.yaml
@@ -6,9 +6,9 @@ interactions:
proto: HTTP/1.1
proto_major: 1
proto_minor: 1
- content_length: 902
+ content_length: 874
host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse write to create a new file called config.json with content ''{\"name\": \"test\", \"version\": \"1.0.0\"}''\n \n\n","role":"user"}],"model":"qwen/qwen3-next-80b-a3b-instruct","max_tokens":40,"stream_options":{"include_usage":true},"usage":{"include":true},"stream":true}'
+ body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\ndownload the file from https://example-files.online-convert.com/document/txt/example.txt and save it as example.txt\n \n\n","role":"user"}],"model":"gpt-oss-120b","max_completion_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
headers:
Accept:
- application/json
@@ -16,25 +16,32 @@ interactions:
- application/json
User-Agent:
- Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
+ url: https://hyper.charm.land/v1/chat/completions
method: POST
response:
proto: HTTP/2.0
proto_major: 2
proto_minor: 0
content_length: -1
+ uncompressed: true
body: |+
- : OPENROUTER PROCESSING
+ data: {"id":"chatcmpl-ecd95edeb312d122018c3715","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
- data: {"id":"gen-1776258146-WjvXwiaFi9NttWNKetR4","object":"chat.completion.chunk","created":1776258146,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":"Create","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
+ data: {"id":"chatcmpl-ecd95edeb312d122018c3715","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":"Download"},"finish_reason":null}]}
- data: {"id":"gen-1776258146-WjvXwiaFi9NttWNKetR4","object":"chat.completion.chunk","created":1776258146,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":" config.json with test","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
+ data: {"id":"chatcmpl-ecd95edeb312d122018c3715","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" and"},"finish_reason":null}]}
- data: {"id":"gen-1776258146-WjvXwiaFi9NttWNKetR4","object":"chat.completion.chunk","created":1776258146,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":" data","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
+ data: {"id":"chatcmpl-ecd95edeb312d122018c3715","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" Save"},"finish_reason":null}]}
- data: {"id":"gen-1776258146-WjvXwiaFi9NttWNKetR4","object":"chat.completion.chunk","created":1776258146,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}]}
+ data: {"id":"chatcmpl-ecd95edeb312d122018c3715","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" example"},"finish_reason":null}]}
- data: {"id":"gen-1776258146-WjvXwiaFi9NttWNKetR4","object":"chat.completion.chunk","created":1776258146,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}],"usage":{"prompt_tokens":175,"completion_tokens":6,"total_tokens":181,"cost":0.0000217425,"is_byok":false,"prompt_tokens_details":{"cached_tokens":0,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00003345,"upstream_inference_prompt_cost":0.00002625,"upstream_inference_completions_cost":0.0000072},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
+ data: {"id":"chatcmpl-ecd95edeb312d122018c3715","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":".txt"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-ecd95edeb312d122018c3715","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" from"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-ecd95edeb312d122018c3715","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" URL"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-ecd95edeb312d122018c3715","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{"prompt_tokens":46,"completion_tokens":177,"total_tokens":399}}
data: [DONE]
@@ -43,15 +50,15 @@ interactions:
- text/event-stream
status: 200 OK
code: 200
- duration: 656.330833ms
+ duration: 3.017650708s
- id: 1
request:
proto: HTTP/1.1
proto_major: 1
proto_minor: 1
- content_length: 30868
+ content_length: 30714
host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/write_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"use write to create a new file called config.json with content '{\\\"name\\\": \\\"test\\\", \\\"version\\\": \\\"1.0.0\\\"}'\",\"role\":\"user\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
+ body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/glm-5.1/download_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"download the file from https://example-files.online-convert.com/document/txt/example.txt and save it as example.txt\",\"role\":\"user\"}],\"model\":\"glm-5.1\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
headers:
Accept:
- application/json
@@ -59,144 +66,85 @@ interactions:
- application/json
User-Agent:
- Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
+ url: https://hyper.charm.land/v1/chat/completions
method: POST
response:
proto: HTTP/2.0
proto_major: 2
proto_minor: 0
content_length: -1
+ uncompressed: true
body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"id":"write:0","type":"function","function":{"name":"write","arguments":""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"file"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"_path"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \"/"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"tmp"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/c"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"rush"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-test"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/Test"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"Coder"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"Agent"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/open"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"router"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-k"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"imi"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-k"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"2"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/write"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"_tool"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/config"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":".json"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\","}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \""}}]},"finish_reason":null,"native_finish_reason":null}]}
+ data: {"id":"chatcmpl-b6256a830edad7acc0d27d45","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"content"}}]},"finish_reason":null,"native_finish_reason":null}]}
+ data: {"id":"chatcmpl-b6256a830edad7acc0d27d45","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_df59","type":"function","index":0,"function":{"name":"download","arguments":""}}]},"finish_reason":null}]}
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
+ data: {"id":"chatcmpl-b6256a830edad7acc0d27d45","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\"file_path\": \"/tmp/crush-test/TestCoderAgent/glm-5.1/download_tool/example.txt\", \"url\": \"https://example-files.online-convert.com/document/txt/example.txt\"}"}}]},"finish_reason":null}]}
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \"{\\\""}}]},"finish_reason":null,"native_finish_reason":null}]}
+ data: {"id":"chatcmpl-b6256a830edad7acc0d27d45","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":{"prompt_tokens":138,"completion_tokens":63,"total_tokens":7145}}
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"name"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\\\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \\\""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"test"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\\\","}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \\\""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"version"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\\\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \\\""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"1"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"."}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"0"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"."}}]},"finish_reason":null,"native_finish_reason":null}]}
+ data: [DONE]
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"0"}}]},"finish_reason":null,"native_finish_reason":null}]}
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 4.507119375s
+- id: 2
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 0
+ host: example-files.online-convert.com
+ headers:
+ User-Agent:
+ - crush/1.0
+ url: https://example-files.online-convert.com/document/txt/example.txt
+ method: GET
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |-
+ TXT test file
+ Purpose: Provide example of this file type
+ Document file type: TXT
+ Version: 1.0
+ Remark:
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\\\""}}]},"finish_reason":null,"native_finish_reason":null}]}
+ Example content:
+ The names "John Doe" for males, "Jane Doe" or "Jane Roe" for females, or "Jonnie Doe" and "Janie Doe" for children, or just "Doe" non-gender-specifically are used as placeholder names for a party whose true identity is unknown or must be withheld in a legal action, case, or discussion. The names are also used to refer to acorpse or hospital patient whose identity is unknown. This practice is widely used in the United States and Canada, but is rarely used in other English-speaking countries including the United Kingdom itself, from where the use of "John Doe" in a legal context originates. The names Joe Bloggs or John Smith are used in the UK instead, as well as in Australia and New Zealand.
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"}\""}}]},"finish_reason":null,"native_finish_reason":null}]}
+ John Doe is sometimes used to refer to a typical male in other contexts as well, in a similar manner to John Q. Public, known in Great Britain as Joe Public, John Smith or Joe Bloggs. For example, the first name listed on a form is often John Doe, along with a fictional address or other fictional information to provide an example of how to fill in the form. The name is also used frequently in popular culture, for example in the Frank Capra film Meet John Doe. John Doe was also the name of a 2002 American television series.
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"}"}}]},"finish_reason":null,"native_finish_reason":null}]}
+ Similarly, a child or baby whose identity is unknown may be referred to as Baby Doe. A notorious murder case in Kansas City, Missouri, referred to the baby victim as Precious Doe. Other unidentified female murder victims are Cali Doe and Princess Doe. Additional persons may be called James Doe, Judy Doe, etc. However, to avoid possible confusion, if two anonymous or unknown parties are cited in a specific case or action, the surnames Doe and Roe may be used simultaneously; for example, "John Doe v. Jane Roe". If several anonymous parties are referenced, they may simply be labelled John Doe #1, John Doe #2, etc. (the U.S. Operation Delego cited 21 (numbered) "John Doe"s) or labelled with other variants of Doe / Roe / Poe / etc. Other early alternatives such as John Stiles and Richard Miles are now rarely used, and Mary Major has been used in some American federal cases.
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}]}
- data: {"id":"gen-1776258146-uP60oPfOQQ9X9cX5zPQw","object":"chat.completion.chunk","created":1776258146,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}],"usage":{"prompt_tokens":6511,"completion_tokens":53,"total_tokens":6564,"cost":0.0040391,"is_byok":false,"prompt_tokens_details":{"cached_tokens":6144,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0040391,"upstream_inference_prompt_cost":0.0039066,"upstream_inference_completions_cost":0.0001325},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
- data: [DONE]
+ File created by https://www.online-convert.com
+ More example files: https://www.online-convert.com/file-type
+ Text of Example content: Wikipedia (https://en.wikipedia.org/wiki/John_Doe)
+ License: Attribution-ShareAlike 4.0 (https://creativecommons.org/licenses/by-sa/4.0/)
+ Feel free to use and share the file according to the license above.
headers:
Content-Type:
- - text/event-stream
+ - text/plain; charset=UTF-8
status: 200 OK
code: 200
- duration: 3.158994292s
-- id: 2
+ duration: 1.541516042s
+- id: 3
request:
proto: HTTP/1.1
proto_major: 1
proto_minor: 1
- content_length: 31291
+ content_length: 31145
host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/write_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use write to create a new file called config.json with content '{\\\"name\\\": \\\"test\\\", \\\"version\\\": \\\"1.0.0\\\"}'\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"write:0\",\"function\":{\"arguments\":\"{\\\"file_path\\\": \\\"/tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/write_tool/config.json\\\", \\\"content\\\": \\\"{\\\\\\\"name\\\\\\\": \\\\\\\"test\\\\\\\", \\\\\\\"version\\\\\\\": \\\\\\\"1.0.0\\\\\\\"}\\\"}\",\"name\":\"write\"},\"type\":\"function\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"role\":\"assistant\"},{\"content\":\"\\nFile successfully written: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/write_tool/config.json\\n\",\"tool_call_id\":\"write:0\",\"role\":\"tool\"}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
+ body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/glm-5.1/download_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"download the file from https://example-files.online-convert.com/document/txt/example.txt and save it as example.txt\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_df59\",\"function\":{\"arguments\":\"{\\\"file_path\\\": \\\"/tmp/crush-test/TestCoderAgent/glm-5.1/download_tool/example.txt\\\", \\\"url\\\": \\\"https://example-files.online-convert.com/document/txt/example.txt\\\"}\",\"name\":\"download\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"Successfully downloaded 2574 bytes to example.txt (Content-Type: text/plain; charset=UTF-8)\",\"tool_call_id\":\"call_df59\",\"role\":\"tool\"}],\"model\":\"glm-5.1\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
headers:
Accept:
- application/json
@@ -204,23 +152,24 @@ interactions:
- application/json
User-Agent:
- Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
+ url: https://hyper.charm.land/v1/chat/completions
method: POST
response:
proto: HTTP/2.0
proto_major: 2
proto_minor: 0
content_length: -1
+ uncompressed: true
body: |+
- : OPENROUTER PROCESSING
+ data: {"id":"chatcmpl-b8f9f29682f5a994a9d174f6","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
- : OPENROUTER PROCESSING
+ data: {"id":"chatcmpl-b8f9f29682f5a994a9d174f6","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"File downloaded and"},"finish_reason":null}]}
- data: {"id":"gen-1776258150-70tC1YgMP2olkAYDf7l1","object":"chat.completion.chunk","created":1776258150,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"Done","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
+ data: {"id":"chatcmpl-b8f9f29682f5a994a9d174f6","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" saved as `"},"finish_reason":null}]}
- data: {"id":"gen-1776258150-70tC1YgMP2olkAYDf7l1","object":"chat.completion.chunk","created":1776258150,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}]}
+ data: {"id":"chatcmpl-b8f9f29682f5a994a9d174f6","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"example.txt`."},"finish_reason":null}]}
- data: {"id":"gen-1776258150-70tC1YgMP2olkAYDf7l1","object":"chat.completion.chunk","created":1776258150,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}],"usage":{"prompt_tokens":6609,"completion_tokens":2,"total_tokens":6611,"cost":0.0039704,"is_byok":false,"prompt_tokens_details":{"cached_tokens":6400,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0039704,"upstream_inference_prompt_cost":0.0039654,"upstream_inference_completions_cost":0.000005},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
+ data: {"id":"chatcmpl-b8f9f29682f5a994a9d174f6","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{"prompt_tokens":82,"completion_tokens":10,"total_tokens":7164}}
data: [DONE]
@@ -229,4 +178,4 @@ interactions:
- text/event-stream
status: 200 OK
code: 200
- duration: 1.069518208s
+ duration: 1.073888166s
diff --git a/internal/agent/testdata/TestCoderAgent/openai-gpt-5/fetch_tool.yaml b/internal/agent/testdata/TestCoderAgent/glm-5.1/fetch_tool.yaml
similarity index 78%
rename from internal/agent/testdata/TestCoderAgent/openai-gpt-5/fetch_tool.yaml
rename to internal/agent/testdata/TestCoderAgent/glm-5.1/fetch_tool.yaml
index f55129a7abfa3046002769db4c68f59813abe8df..043f8aee8e68b79d190f7cd7c71494d5458c18c3 100644
--- a/internal/agent/testdata/TestCoderAgent/openai-gpt-5/fetch_tool.yaml
+++ b/internal/agent/testdata/TestCoderAgent/glm-5.1/fetch_tool.yaml
@@ -6,9 +6,9 @@ interactions:
proto: HTTP/1.1
proto_major: 1
proto_minor: 1
- content_length: 878
+ content_length: 895
host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nfetch the content from https://example-files.online-convert.com/website/html/example.html and tell me if it contains the word ''John Doe''\n \n\n","role":"user"}],"model":"gpt-4o","max_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
+ body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nfetch the content from https://example-files.online-convert.com/website/html/example.html and tell me if it contains the word ''John Doe''\n \n\n","role":"user"}],"model":"gpt-oss-120b","max_completion_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
headers:
Accept:
- application/json
@@ -16,56 +16,49 @@ interactions:
- application/json
User-Agent:
- Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
+ url: https://hyper.charm.land/v1/chat/completions
method: POST
response:
proto: HTTP/2.0
proto_major: 2
proto_minor: 0
content_length: -1
+ uncompressed: true
body: |+
- data: {"id":"chatcmpl-DUu7leRpegv7FTwppRdPxLoDutmRp","object":"chat.completion.chunk","created":1776257953,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d65b5b17fe","choices":[{"index":0,"delta":{"role":"assistant","content":"","refusal":null},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"jvf30qULGidN2V"}
-
- data: {"id":"chatcmpl-DUu7leRpegv7FTwppRdPxLoDutmRp","object":"chat.completion.chunk","created":1776257953,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d65b5b17fe","choices":[{"index":0,"delta":{"content":"Check"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"giDQMHsHsaK"}
-
- data: {"id":"chatcmpl-DUu7leRpegv7FTwppRdPxLoDutmRp","object":"chat.completion.chunk","created":1776257953,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d65b5b17fe","choices":[{"index":0,"delta":{"content":" for"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"JAxGYkIBjd9W"}
-
- data: {"id":"chatcmpl-DUu7leRpegv7FTwppRdPxLoDutmRp","object":"chat.completion.chunk","created":1776257953,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d65b5b17fe","choices":[{"index":0,"delta":{"content":" '"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"yr4WzGAPPot7Sf"}
-
- data: {"id":"chatcmpl-DUu7leRpegv7FTwppRdPxLoDutmRp","object":"chat.completion.chunk","created":1776257953,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d65b5b17fe","choices":[{"index":0,"delta":{"content":"John"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"oKUPSAS7l6kZ"}
+ data: {"id":"chatcmpl-5e5f8ab078fc8e890a00550d","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu7leRpegv7FTwppRdPxLoDutmRp","object":"chat.completion.chunk","created":1776257953,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d65b5b17fe","choices":[{"index":0,"delta":{"content":" Doe"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"LjGLjnAdEQUr"}
+ data: {"id":"chatcmpl-5e5f8ab078fc8e890a00550d","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":"Check"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu7leRpegv7FTwppRdPxLoDutmRp","object":"chat.completion.chunk","created":1776257953,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d65b5b17fe","choices":[{"index":0,"delta":{"content":"'"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"9SHF6RnjpLTAEsu"}
+ data: {"id":"chatcmpl-5e5f8ab078fc8e890a00550d","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" for"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu7leRpegv7FTwppRdPxLoDutmRp","object":"chat.completion.chunk","created":1776257953,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d65b5b17fe","choices":[{"index":0,"delta":{"content":" in"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"sBfZdecUfRFOd"}
+ data: {"id":"chatcmpl-5e5f8ab078fc8e890a00550d","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" John"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu7leRpegv7FTwppRdPxLoDutmRp","object":"chat.completion.chunk","created":1776257953,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d65b5b17fe","choices":[{"index":0,"delta":{"content":" Online"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"zvGABZYSl"}
+ data: {"id":"chatcmpl-5e5f8ab078fc8e890a00550d","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" Doe"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu7leRpegv7FTwppRdPxLoDutmRp","object":"chat.completion.chunk","created":1776257953,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d65b5b17fe","choices":[{"index":0,"delta":{"content":" HTML"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"zMm9h9k5cZ0"}
+ data: {"id":"chatcmpl-5e5f8ab078fc8e890a00550d","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" in"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu7leRpegv7FTwppRdPxLoDutmRp","object":"chat.completion.chunk","created":1776257953,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d65b5b17fe","choices":[{"index":0,"delta":{"content":" File"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"kubOtDlhaEs"}
+ data: {"id":"chatcmpl-5e5f8ab078fc8e890a00550d","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" example"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu7leRpegv7FTwppRdPxLoDutmRp","object":"chat.completion.chunk","created":1776257953,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d65b5b17fe","choices":[{"index":0,"delta":{},"logprobs":null,"finish_reason":"stop"}],"usage":null,"obfuscation":"UsjdBVfwJ0"}
+ data: {"id":"chatcmpl-5e5f8ab078fc8e890a00550d","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" HTML"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu7leRpegv7FTwppRdPxLoDutmRp","object":"chat.completion.chunk","created":1776257953,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d65b5b17fe","choices":[],"usage":{"prompt_tokens":169,"completion_tokens":10,"total_tokens":179,"prompt_tokens_details":{"cached_tokens":0,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":0,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"K4M1Jf0PUNp4O"}
+ data: {"id":"chatcmpl-5e5f8ab078fc8e890a00550d","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{"prompt_tokens":51,"completion_tokens":281,"total_tokens":508}}
data: [DONE]
headers:
Content-Type:
- - text/event-stream; charset=utf-8
+ - text/event-stream
status: 200 OK
code: 200
- duration: 1.854110209s
+ duration: 3.15393675s
- id: 1
request:
proto: HTTP/1.1
proto_major: 1
proto_minor: 1
- content_length: 30746
+ content_length: 30732
host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/fetch_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"fetch the content from https://example-files.online-convert.com/website/html/example.html and tell me if it contains the word 'John Doe'\",\"role\":\"user\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
+ body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/glm-5.1/fetch_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"fetch the content from https://example-files.online-convert.com/website/html/example.html and tell me if it contains the word 'John Doe'\",\"role\":\"user\"}],\"model\":\"glm-5.1\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
headers:
Accept:
- application/json
@@ -73,78 +66,31 @@ interactions:
- application/json
User-Agent:
- Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
+ url: https://hyper.charm.land/v1/chat/completions
method: POST
response:
proto: HTTP/2.0
proto_major: 2
proto_minor: 0
content_length: -1
+ uncompressed: true
body: |+
- data: {"id":"chatcmpl-DUu7ksqxlXggVdt7buj81RDNH5dRr","object":"chat.completion.chunk","created":1776257952,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":null,"tool_calls":[{"index":0,"id":"call_bv9KrGPJ9DPWDQ8eWLccEJ32","type":"function","function":{"name":"fetch","arguments":""}}],"refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"ALMUaeorv8L"}
-
- data: {"id":"chatcmpl-DUu7ksqxlXggVdt7buj81RDNH5dRr","object":"chat.completion.chunk","created":1776257952,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"LChLMPJuWxNt4bG"}
-
- data: {"id":"chatcmpl-DUu7ksqxlXggVdt7buj81RDNH5dRr","object":"chat.completion.chunk","created":1776257952,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"format"}}]},"finish_reason":null}],"usage":null,"obfuscation":"zCAY4N2Ja20v"}
-
- data: {"id":"chatcmpl-DUu7ksqxlXggVdt7buj81RDNH5dRr","object":"chat.completion.chunk","created":1776257952,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"p6iWeMPkhdb6g"}
-
- data: {"id":"chatcmpl-DUu7ksqxlXggVdt7buj81RDNH5dRr","object":"chat.completion.chunk","created":1776257952,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"text"}}]},"finish_reason":null}],"usage":null,"obfuscation":"9nW0v5CexJ8svD"}
-
- data: {"id":"chatcmpl-DUu7ksqxlXggVdt7buj81RDNH5dRr","object":"chat.completion.chunk","created":1776257952,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\",\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"rQorrQoUUssxR"}
-
- data: {"id":"chatcmpl-DUu7ksqxlXggVdt7buj81RDNH5dRr","object":"chat.completion.chunk","created":1776257952,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"timeout"}}]},"finish_reason":null}],"usage":null,"obfuscation":"ZrwYUK4AugI"}
-
- data: {"id":"chatcmpl-DUu7ksqxlXggVdt7buj81RDNH5dRr","object":"chat.completion.chunk","created":1776257952,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"CV3GEfZLBvvdbY9"}
-
- data: {"id":"chatcmpl-DUu7ksqxlXggVdt7buj81RDNH5dRr","object":"chat.completion.chunk","created":1776257952,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"60"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu7ksqxlXggVdt7buj81RDNH5dRr","object":"chat.completion.chunk","created":1776257952,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":",\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"IYKDSPIncTuhPBq"}
-
- data: {"id":"chatcmpl-DUu7ksqxlXggVdt7buj81RDNH5dRr","object":"chat.completion.chunk","created":1776257952,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"url"}}]},"finish_reason":null}],"usage":null,"obfuscation":"KHMAuMRNkn9jnEW"}
-
- data: {"id":"chatcmpl-DUu7ksqxlXggVdt7buj81RDNH5dRr","object":"chat.completion.chunk","created":1776257952,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"S5LPvy5gX3km8"}
-
- data: {"id":"chatcmpl-DUu7ksqxlXggVdt7buj81RDNH5dRr","object":"chat.completion.chunk","created":1776257952,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"https"}}]},"finish_reason":null}],"usage":null,"obfuscation":"3CKo6wKqGJkW4"}
+ data: {"id":"chatcmpl-f52b285094c2eddf20195bcc","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu7ksqxlXggVdt7buj81RDNH5dRr","object":"chat.completion.chunk","created":1776257952,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"://"}}]},"finish_reason":null}],"usage":null,"obfuscation":"9rISnQHi8RHYGwa"}
+ data: {"id":"chatcmpl-f52b285094c2eddf20195bcc","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_bd35","type":"function","index":0,"function":{"name":"fetch","arguments":""}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu7ksqxlXggVdt7buj81RDNH5dRr","object":"chat.completion.chunk","created":1776257952,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"example"}}]},"finish_reason":null}],"usage":null,"obfuscation":"tAIlsR8wQJA"}
+ data: {"id":"chatcmpl-f52b285094c2eddf20195bcc","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\"format\": \"text\", \"url\": \"https://example-files.online-convert.com/website/html/example.html\"}"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu7ksqxlXggVdt7buj81RDNH5dRr","object":"chat.completion.chunk","created":1776257952,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-files"}}]},"finish_reason":null}],"usage":null,"obfuscation":"Dy09TpffGAsZ"}
-
- data: {"id":"chatcmpl-DUu7ksqxlXggVdt7buj81RDNH5dRr","object":"chat.completion.chunk","created":1776257952,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".online"}}]},"finish_reason":null}],"usage":null,"obfuscation":"b44rT3IZg1K"}
-
- data: {"id":"chatcmpl-DUu7ksqxlXggVdt7buj81RDNH5dRr","object":"chat.completion.chunk","created":1776257952,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-con"}}]},"finish_reason":null}],"usage":null,"obfuscation":"AuBfBMMWkefMfX"}
-
- data: {"id":"chatcmpl-DUu7ksqxlXggVdt7buj81RDNH5dRr","object":"chat.completion.chunk","created":1776257952,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"vert"}}]},"finish_reason":null}],"usage":null,"obfuscation":"6ABwGiIF4Lsd1L"}
-
- data: {"id":"chatcmpl-DUu7ksqxlXggVdt7buj81RDNH5dRr","object":"chat.completion.chunk","created":1776257952,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".com"}}]},"finish_reason":null}],"usage":null,"obfuscation":"jGWMHn2e2GycKV"}
-
- data: {"id":"chatcmpl-DUu7ksqxlXggVdt7buj81RDNH5dRr","object":"chat.completion.chunk","created":1776257952,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/"}}]},"finish_reason":null}],"usage":null,"obfuscation":"E"}
-
- data: {"id":"chatcmpl-DUu7ksqxlXggVdt7buj81RDNH5dRr","object":"chat.completion.chunk","created":1776257952,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"website"}}]},"finish_reason":null}],"usage":null,"obfuscation":"36Qo22spoWx"}
-
- data: {"id":"chatcmpl-DUu7ksqxlXggVdt7buj81RDNH5dRr","object":"chat.completion.chunk","created":1776257952,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/html"}}]},"finish_reason":null}],"usage":null,"obfuscation":"RPXGtSj5lY5cO"}
-
- data: {"id":"chatcmpl-DUu7ksqxlXggVdt7buj81RDNH5dRr","object":"chat.completion.chunk","created":1776257952,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/example"}}]},"finish_reason":null}],"usage":null,"obfuscation":"Hk0XepbaYO"}
-
- data: {"id":"chatcmpl-DUu7ksqxlXggVdt7buj81RDNH5dRr","object":"chat.completion.chunk","created":1776257952,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".html"}}]},"finish_reason":null}],"usage":null,"obfuscation":"DwRLM0KUT787l"}
-
- data: {"id":"chatcmpl-DUu7ksqxlXggVdt7buj81RDNH5dRr","object":"chat.completion.chunk","created":1776257952,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"}"}}]},"finish_reason":null}],"usage":null,"obfuscation":"js5OALmPediUzcs"}
-
- data: {"id":"chatcmpl-DUu7ksqxlXggVdt7buj81RDNH5dRr","object":"chat.completion.chunk","created":1776257952,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu7ksqxlXggVdt7buj81RDNH5dRr","object":"chat.completion.chunk","created":1776257952,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6394,"completion_tokens":298,"total_tokens":6692,"prompt_tokens_details":{"cached_tokens":3456,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":256,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"9"}
+ data: {"id":"chatcmpl-f52b285094c2eddf20195bcc","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":{"prompt_tokens":145,"completion_tokens":49,"total_tokens":7138}}
data: [DONE]
headers:
Content-Type:
- - text/event-stream; charset=utf-8
+ - text/event-stream
status: 200 OK
code: 200
- duration: 5.281282209s
+ duration: 4.480027958s
- id: 2
request:
proto: HTTP/1.1
@@ -215,15 +161,15 @@ interactions:
- text/html; charset=UTF-8
status: 200 OK
code: 200
- duration: 407.700042ms
+ duration: 510.770375ms
- id: 3
request:
proto: HTTP/1.1
proto_major: 1
proto_minor: 1
- content_length: 33578
+ content_length: 33512
host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/fetch_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"fetch the content from https://example-files.online-convert.com/website/html/example.html and tell me if it contains the word 'John Doe'\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_bv9KrGPJ9DPWDQ8eWLccEJ32\",\"function\":{\"arguments\":\"{\\\"format\\\":\\\"text\\\",\\\"timeout\\\":60,\\\"url\\\":\\\"https://example-files.online-convert.com/website/html/example.html\\\"}\",\"name\":\"fetch\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"HTML test file Purpose: Provide example of this file type Document file type: HTML Version: 1.0 Remark: Example content: The names \\\"John Doe\\\" for males, \\\"Jane Doe\\\" or \\\"Jane Roe\\\" for females, or \\\"Jonnie Doe\\\" and \\\"Janie Doe\\\" for children, or just \\\"Doe\\\" non-gender-specifically are used as placeholder names for a party whose true identity is unknown or must be withheld in a legal action, case, or discussion. The names are also used to refer to acorpse or hospital patient whose identity is unknown. This practice is widely used in the United States and Canada, but is rarely used in other English-speaking countries including the United Kingdom itself, from where the use of \\\"John Doe\\\" in a legal context originates. The names Joe Bloggs or John Smith are used in the UK instead, as well as in Australia and New Zealand. John Doe is sometimes used to refer to a typical male in other contexts as well, in a similar manner to John Q. Public, known in Great Britain as Joe Public, John Smith or Joe Bloggs. For example, the first name listed on a form is often John Doe, along with a fictional address or other fictional information to provide an example of how to fill in the form. The name is also used frequently in popular culture, for example in the Frank Capra film Meet John Doe. John Doe was also the name of a 2002 American television series. Similarly, a child or baby whose identity is unknown may be referred to as Baby Doe. A notorious murder case in Kansas City, Missouri, referred to the baby victim as Precious Doe. Other unidentified female murder victims are Cali Doe and Princess Doe. Additional persons may be called James Doe, Judy Doe, etc. However, to avoid possible confusion, if two anonymous or unknown parties are cited in a specific case or action, the surnames Doe and Roe may be used simultaneously; for example, \\\"John Doe v. Jane Roe\\\". If several anonymous parties are referenced, they may simply be labelled John Doe #1, John Doe #2, etc. (the U.S. Operation Delego cited 21 (numbered) \\\"John Doe\\\"s) or labelled with other variants of Doe / Roe / Poe / etc. Other early alternatives such as John Stiles and Richard Miles are now rarely used, and Mary Major has been used in some American federal cases. File created by https://www.online-convert.com More example files: https://www.online-convert.com/file-type Text of \\\"Example content\\\": Wikipedia License: Attribution-ShareAlike 3.0 Unported Feel free to use and share the file according to license above.\",\"tool_call_id\":\"call_bv9KrGPJ9DPWDQ8eWLccEJ32\",\"role\":\"tool\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
+ body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/glm-5.1/fetch_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"fetch the content from https://example-files.online-convert.com/website/html/example.html and tell me if it contains the word 'John Doe'\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_bd35\",\"function\":{\"arguments\":\"{\\\"format\\\": \\\"text\\\", \\\"url\\\": \\\"https://example-files.online-convert.com/website/html/example.html\\\"}\",\"name\":\"fetch\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"HTML test file Purpose: Provide example of this file type Document file type: HTML Version: 1.0 Remark: Example content: The names \\\"John Doe\\\" for males, \\\"Jane Doe\\\" or \\\"Jane Roe\\\" for females, or \\\"Jonnie Doe\\\" and \\\"Janie Doe\\\" for children, or just \\\"Doe\\\" non-gender-specifically are used as placeholder names for a party whose true identity is unknown or must be withheld in a legal action, case, or discussion. The names are also used to refer to acorpse or hospital patient whose identity is unknown. This practice is widely used in the United States and Canada, but is rarely used in other English-speaking countries including the United Kingdom itself, from where the use of \\\"John Doe\\\" in a legal context originates. The names Joe Bloggs or John Smith are used in the UK instead, as well as in Australia and New Zealand. John Doe is sometimes used to refer to a typical male in other contexts as well, in a similar manner to John Q. Public, known in Great Britain as Joe Public, John Smith or Joe Bloggs. For example, the first name listed on a form is often John Doe, along with a fictional address or other fictional information to provide an example of how to fill in the form. The name is also used frequently in popular culture, for example in the Frank Capra film Meet John Doe. John Doe was also the name of a 2002 American television series. Similarly, a child or baby whose identity is unknown may be referred to as Baby Doe. A notorious murder case in Kansas City, Missouri, referred to the baby victim as Precious Doe. Other unidentified female murder victims are Cali Doe and Princess Doe. Additional persons may be called James Doe, Judy Doe, etc. However, to avoid possible confusion, if two anonymous or unknown parties are cited in a specific case or action, the surnames Doe and Roe may be used simultaneously; for example, \\\"John Doe v. Jane Roe\\\". If several anonymous parties are referenced, they may simply be labelled John Doe #1, John Doe #2, etc. (the U.S. Operation Delego cited 21 (numbered) \\\"John Doe\\\"s) or labelled with other variants of Doe / Roe / Poe / etc. Other early alternatives such as John Stiles and Richard Miles are now rarely used, and Mary Major has been used in some American federal cases. File created by https://www.online-convert.com More example files: https://www.online-convert.com/file-type Text of \\\"Example content\\\": Wikipedia License: Attribution-ShareAlike 3.0 Unported Feel free to use and share the file according to license above.\",\"tool_call_id\":\"call_bd35\",\"role\":\"tool\"}],\"model\":\"glm-5.1\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
headers:
Accept:
- application/json
@@ -231,27 +177,36 @@ interactions:
- application/json
User-Agent:
- Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
+ url: https://hyper.charm.land/v1/chat/completions
method: POST
response:
proto: HTTP/2.0
proto_major: 2
proto_minor: 0
content_length: -1
+ uncompressed: true
body: |+
- data: {"id":"chatcmpl-DUu7qJlUwKm7Kkdf3VKPQAcTkPtnI","object":"chat.completion.chunk","created":1776257958,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":"","refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"5SAUXxTFnS"}
+ data: {"id":"chatcmpl-d8d51f1ff93734d49f0279e9","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d8d51f1ff93734d49f0279e9","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"Yes, the page"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d8d51f1ff93734d49f0279e9","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" contains the word '"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d8d51f1ff93734d49f0279e9","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"John Doe' —"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d8d51f1ff93734d49f0279e9","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" it appears multiple times"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu7qJlUwKm7Kkdf3VKPQAcTkPtnI","object":"chat.completion.chunk","created":1776257958,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"Yes"},"finish_reason":null}],"usage":null,"obfuscation":"MkZxC01IC"}
+ data: {"id":"chatcmpl-d8d51f1ff93734d49f0279e9","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" throughout the content"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu7qJlUwKm7Kkdf3VKPQAcTkPtnI","object":"chat.completion.chunk","created":1776257958,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":null,"obfuscation":"fPF8yT"}
+ data: {"id":"chatcmpl-d8d51f1ff93734d49f0279e9","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"."},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu7qJlUwKm7Kkdf3VKPQAcTkPtnI","object":"chat.completion.chunk","created":1776257958,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6988,"completion_tokens":202,"total_tokens":7190,"prompt_tokens_details":{"cached_tokens":6272,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":192,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"U"}
+ data: {"id":"chatcmpl-d8d51f1ff93734d49f0279e9","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{"prompt_tokens":602,"completion_tokens":21,"total_tokens":7695}}
data: [DONE]
headers:
Content-Type:
- - text/event-stream; charset=utf-8
+ - text/event-stream
status: 200 OK
code: 200
- duration: 4.176065666s
+ duration: 912.446417ms
diff --git a/internal/agent/testdata/TestCoderAgent/glm-5.1/glob_tool.yaml b/internal/agent/testdata/TestCoderAgent/glm-5.1/glob_tool.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..27c2dc55f43a7dcabc4af3d7f3bba95d7568de98
--- /dev/null
+++ b/internal/agent/testdata/TestCoderAgent/glm-5.1/glob_tool.yaml
@@ -0,0 +1,143 @@
+---
+version: 2
+interactions:
+- id: 0
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 814
+ host: ""
+ body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse glob to find all .go files in the current directory\n \n\n","role":"user"}],"model":"gpt-oss-120b","max_completion_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
+ headers:
+ Accept:
+ - application/json
+ Content-Type:
+ - application/json
+ User-Agent:
+ - Charm-Crush/devel (https://charm.land/crush)
+ url: https://hyper.charm.land/v1/chat/completions
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |+
+ data: {"id":"chatcmpl-700285b94d03a05d9e3f02b8","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-700285b94d03a05d9e3f02b8","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":"Use"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-700285b94d03a05d9e3f02b8","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" glob"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-700285b94d03a05d9e3f02b8","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" to"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-700285b94d03a05d9e3f02b8","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" locate"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-700285b94d03a05d9e3f02b8","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" ."},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-700285b94d03a05d9e3f02b8","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":"go"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-700285b94d03a05d9e3f02b8","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" files"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-700285b94d03a05d9e3f02b8","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" in"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-700285b94d03a05d9e3f02b8","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" current"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-700285b94d03a05d9e3f02b8","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" directory"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-700285b94d03a05d9e3f02b8","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{"prompt_tokens":35,"completion_tokens":298,"total_tokens":509}}
+
+ data: [DONE]
+
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 3.3505585s
+- id: 1
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 30650
+ host: ""
+ body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/glm-5.1/glob_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use glob to find all .go files in the current directory\",\"role\":\"user\"}],\"model\":\"glm-5.1\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
+ headers:
+ Accept:
+ - application/json
+ Content-Type:
+ - application/json
+ User-Agent:
+ - Charm-Crush/devel (https://charm.land/crush)
+ url: https://hyper.charm.land/v1/chat/completions
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |+
+ data: {"id":"chatcmpl-2302edd01e4d115e64f8a74b","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-2302edd01e4d115e64f8a74b","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_ae51","type":"function","index":0,"function":{"name":"glob","arguments":""}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-2302edd01e4d115e64f8a74b","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\"pattern\": \"**/*.go\"}"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-2302edd01e4d115e64f8a74b","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":{"prompt_tokens":128,"completion_tokens":12,"total_tokens":7084}}
+
+ data: [DONE]
+
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 4.272323542s
+- id: 2
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 30903
+ host: ""
+ body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/glm-5.1/glob_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use glob to find all .go files in the current directory\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_ae51\",\"function\":{\"arguments\":\"{\\\"pattern\\\": \\\"**/*.go\\\"}\",\"name\":\"glob\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"/tmp/crush-test/TestCoderAgent/glm-5.1/glob_tool/main.go\",\"tool_call_id\":\"call_ae51\",\"role\":\"tool\"}],\"model\":\"glm-5.1\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
+ headers:
+ Accept:
+ - application/json
+ Content-Type:
+ - application/json
+ User-Agent:
+ - Charm-Crush/devel (https://charm.land/crush)
+ url: https://hyper.charm.land/v1/chat/completions
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |+
+ data: {"id":"chatcmpl-4eef785e9e5e3968e73723e6","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-4eef785e9e5e3968e73723e6","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"Found"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-4eef785e9e5e3968e73723e6","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" 1 `.go"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-4eef785e9e5e3968e73723e6","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"` file: `"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-4eef785e9e5e3968e73723e6","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"main.go`"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-4eef785e9e5e3968e73723e6","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{"prompt_tokens":35,"completion_tokens":13,"total_tokens":7120}}
+
+ data: [DONE]
+
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 897.754875ms
diff --git a/internal/agent/testdata/TestCoderAgent/glm-5.1/grep_tool.yaml b/internal/agent/testdata/TestCoderAgent/glm-5.1/grep_tool.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..881eca1912b99168481c3e3588b437678257e447
--- /dev/null
+++ b/internal/agent/testdata/TestCoderAgent/glm-5.1/grep_tool.yaml
@@ -0,0 +1,143 @@
+---
+version: 2
+interactions:
+- id: 0
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 812
+ host: ""
+ body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse grep to search for the word ''package'' in go files\n \n\n","role":"user"}],"model":"gpt-oss-120b","max_completion_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
+ headers:
+ Accept:
+ - application/json
+ Content-Type:
+ - application/json
+ User-Agent:
+ - Charm-Crush/devel (https://charm.land/crush)
+ url: https://hyper.charm.land/v1/chat/completions
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |+
+ data: {"id":"chatcmpl-d83c26df9c6591d438b30a94","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d83c26df9c6591d438b30a94","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":"G"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d83c26df9c6591d438b30a94","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":"rep"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d83c26df9c6591d438b30a94","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" for"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d83c26df9c6591d438b30a94","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" package"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d83c26df9c6591d438b30a94","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" keyword"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d83c26df9c6591d438b30a94","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" in"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d83c26df9c6591d438b30a94","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" Go"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d83c26df9c6591d438b30a94","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" source"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d83c26df9c6591d438b30a94","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" files"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d83c26df9c6591d438b30a94","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{"prompt_tokens":36,"completion_tokens":197,"total_tokens":409}}
+
+ data: [DONE]
+
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 2.981118041s
+- id: 1
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 30648
+ host: ""
+ body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/glm-5.1/grep_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use grep to search for the word 'package' in go files\",\"role\":\"user\"}],\"model\":\"glm-5.1\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
+ headers:
+ Accept:
+ - application/json
+ Content-Type:
+ - application/json
+ User-Agent:
+ - Charm-Crush/devel (https://charm.land/crush)
+ url: https://hyper.charm.land/v1/chat/completions
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |+
+ data: {"id":"chatcmpl-42f08541adc9ffe21182bf11","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-42f08541adc9ffe21182bf11","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_ebae","type":"function","index":0,"function":{"name":"grep","arguments":""}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-42f08541adc9ffe21182bf11","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\"include\": \"*.go\", \"path\": \"/tmp/crush-test/TestCoderAgent/glm-5.1/grep_tool\", \"pattern\": \"package\"}"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-42f08541adc9ffe21182bf11","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":{"prompt_tokens":129,"completion_tokens":54,"total_tokens":7127}}
+
+ data: [DONE]
+
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 4.310790167s
+- id: 2
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 31040
+ host: ""
+ body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/glm-5.1/grep_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use grep to search for the word 'package' in go files\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_ebae\",\"function\":{\"arguments\":\"{\\\"include\\\": \\\"*.go\\\", \\\"path\\\": \\\"/tmp/crush-test/TestCoderAgent/glm-5.1/grep_tool\\\", \\\"pattern\\\": \\\"package\\\"}\",\"name\":\"grep\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"Found 1 matches\\n/tmp/crush-test/TestCoderAgent/glm-5.1/grep_tool/main.go:\\n Line 1, Char 1: package main\\n\",\"tool_call_id\":\"call_ebae\",\"role\":\"tool\"}],\"model\":\"glm-5.1\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
+ headers:
+ Accept:
+ - application/json
+ Content-Type:
+ - application/json
+ User-Agent:
+ - Charm-Crush/devel (https://charm.land/crush)
+ url: https://hyper.charm.land/v1/chat/completions
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |+
+ data: {"id":"chatcmpl-2bb3cae38a9c3157de8b6208","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-2bb3cae38a9c3157de8b6208","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"Found 1 match"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-2bb3cae38a9c3157de8b6208","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":": `main"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-2bb3cae38a9c3157de8b6208","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":".go:1`"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-2bb3cae38a9c3157de8b6208","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" — `package main"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-2bb3cae38a9c3157de8b6208","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"`"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-2bb3cae38a9c3157de8b6208","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{"prompt_tokens":80,"completion_tokens":17,"total_tokens":7169}}
+
+ data: [DONE]
+
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 1.096788458s
diff --git a/internal/agent/testdata/TestCoderAgent/glm-5.1/ls_tool.yaml b/internal/agent/testdata/TestCoderAgent/glm-5.1/ls_tool.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..9b89d9b62abb8bd6ffa05362240b1bce930cfccd
--- /dev/null
+++ b/internal/agent/testdata/TestCoderAgent/glm-5.1/ls_tool.yaml
@@ -0,0 +1,137 @@
+---
+version: 2
+interactions:
+- id: 0
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 808
+ host: ""
+ body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse ls to list the files in the current directory\n \n\n","role":"user"}],"model":"gpt-oss-120b","max_completion_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
+ headers:
+ Accept:
+ - application/json
+ Content-Type:
+ - application/json
+ User-Agent:
+ - Charm-Crush/devel (https://charm.land/crush)
+ url: https://hyper.charm.land/v1/chat/completions
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |+
+ data: {"id":"chatcmpl-d58720ab3cc6652c5a236a50","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d58720ab3cc6652c5a236a50","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":"Using"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d58720ab3cc6652c5a236a50","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" ls"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d58720ab3cc6652c5a236a50","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" to"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d58720ab3cc6652c5a236a50","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" list"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d58720ab3cc6652c5a236a50","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" current"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d58720ab3cc6652c5a236a50","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" directory"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d58720ab3cc6652c5a236a50","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" files"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d58720ab3cc6652c5a236a50","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{"prompt_tokens":33,"completion_tokens":145,"total_tokens":354}}
+
+ data: [DONE]
+
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 1.908952917s
+- id: 1
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 30642
+ host: ""
+ body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/glm-5.1/ls_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use ls to list the files in the current directory\",\"role\":\"user\"}],\"model\":\"glm-5.1\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
+ headers:
+ Accept:
+ - application/json
+ Content-Type:
+ - application/json
+ User-Agent:
+ - Charm-Crush/devel (https://charm.land/crush)
+ url: https://hyper.charm.land/v1/chat/completions
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |+
+ data: {"id":"chatcmpl-efa0d1177fbf30d0d80a6215","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-efa0d1177fbf30d0d80a6215","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_2496","type":"function","index":0,"function":{"name":"ls","arguments":""}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-efa0d1177fbf30d0d80a6215","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{}"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-efa0d1177fbf30d0d80a6215","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":{"prompt_tokens":126,"completion_tokens":4,"total_tokens":7074}}
+
+ data: [DONE]
+
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 3.25056s
+- id: 2
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 30891
+ host: ""
+ body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/glm-5.1/ls_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use ls to list the files in the current directory\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_2496\",\"function\":{\"arguments\":\"{}\",\"name\":\"ls\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n- /tmp/crush-test/TestCoderAgent/glm-5.1/ls_tool/\\n - go.mod\\n - main.go\\n\",\"tool_call_id\":\"call_2496\",\"role\":\"tool\"}],\"model\":\"glm-5.1\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
+ headers:
+ Accept:
+ - application/json
+ Content-Type:
+ - application/json
+ User-Agent:
+ - Charm-Crush/devel (https://charm.land/crush)
+ url: https://hyper.charm.land/v1/chat/completions
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |+
+ data: {"id":"chatcmpl-ff453f3d64106d48e7af0211","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-ff453f3d64106d48e7af0211","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"-"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-ff453f3d64106d48e7af0211","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" `go.mod`\n"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-ff453f3d64106d48e7af0211","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"- `main.go"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-ff453f3d64106d48e7af0211","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"`"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-ff453f3d64106d48e7af0211","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{"prompt_tokens":37,"completion_tokens":11,"total_tokens":7120}}
+
+ data: [DONE]
+
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 1.573505208s
diff --git a/internal/agent/testdata/TestCoderAgent/glm-5.1/multiedit_tool.yaml b/internal/agent/testdata/TestCoderAgent/glm-5.1/multiedit_tool.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..c4ec356ae5027c986dc446ed8847a1e91b05a222
--- /dev/null
+++ b/internal/agent/testdata/TestCoderAgent/glm-5.1/multiedit_tool.yaml
@@ -0,0 +1,233 @@
+---
+version: 2
+interactions:
+- id: 0
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 887
+ host: ""
+ body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse multiedit to change ''Hello, World!'' to ''Hello, Crush!'' and add a comment ''// Greeting'' above the fmt.Println line in main.go\n \n\n","role":"user"}],"model":"gpt-oss-120b","max_completion_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
+ headers:
+ Accept:
+ - application/json
+ Content-Type:
+ - application/json
+ User-Agent:
+ - Charm-Crush/devel (https://charm.land/crush)
+ url: https://hyper.charm.land/v1/chat/completions
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |+
+ data: {"id":"chatcmpl-d5ad9fe3c47867293b4749b1","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d5ad9fe3c47867293b4749b1","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":"Modify"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d5ad9fe3c47867293b4749b1","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" Hello"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d5ad9fe3c47867293b4749b1","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" World"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d5ad9fe3c47867293b4749b1","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" to"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d5ad9fe3c47867293b4749b1","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" Hello"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d5ad9fe3c47867293b4749b1","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" Crush"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d5ad9fe3c47867293b4749b1","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" with"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d5ad9fe3c47867293b4749b1","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" comment"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-d5ad9fe3c47867293b4749b1","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{"prompt_tokens":55,"completion_tokens":155,"total_tokens":386}}
+
+ data: [DONE]
+
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 1.903782417s
+- id: 1
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 30728
+ host: ""
+ body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/glm-5.1/multiedit_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use multiedit to change 'Hello, World!' to 'Hello, Crush!' and add a comment '// Greeting' above the fmt.Println line in main.go\",\"role\":\"user\"}],\"model\":\"glm-5.1\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
+ headers:
+ Accept:
+ - application/json
+ Content-Type:
+ - application/json
+ User-Agent:
+ - Charm-Crush/devel (https://charm.land/crush)
+ url: https://hyper.charm.land/v1/chat/completions
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |+
+ data: {"id":"chatcmpl-40971c2a0e4b872f33c84ffc","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-40971c2a0e4b872f33c84ffc","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_8d3c","type":"function","index":0,"function":{"name":"view","arguments":""}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-40971c2a0e4b872f33c84ffc","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\"file_path\": \"/tmp/crush-test/TestCoderAgent/glm-5.1/multiedit_tool/main.go\"}"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-40971c2a0e4b872f33c84ffc","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":{"prompt_tokens":151,"completion_tokens":44,"total_tokens":7139}}
+
+ data: [DONE]
+
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 3.5521825s
+- id: 2
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 31133
+ host: ""
+ body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/glm-5.1/multiedit_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use multiedit to change 'Hello, World!' to 'Hello, Crush!' and add a comment '// Greeting' above the fmt.Println line in main.go\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_8d3c\",\"function\":{\"arguments\":\"{\\\"file_path\\\": \\\"/tmp/crush-test/TestCoderAgent/glm-5.1/multiedit_tool/main.go\\\"}\",\"name\":\"view\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\tfmt.Println(\\\"Hello, World!\\\")\\n 7|}\\n\\n\",\"tool_call_id\":\"call_8d3c\",\"role\":\"tool\"}],\"model\":\"glm-5.1\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
+ headers:
+ Accept:
+ - application/json
+ Content-Type:
+ - application/json
+ User-Agent:
+ - Charm-Crush/devel (https://charm.land/crush)
+ url: https://hyper.charm.land/v1/chat/completions
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |+
+ data: {"id":"chatcmpl-440dd811605b41f84c1517a9","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-440dd811605b41f84c1517a9","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_eda6","type":"function","index":0,"function":{"name":"multiedit","arguments":""}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-440dd811605b41f84c1517a9","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\"edits\": [{\"new_string\": \"\\\"Hello, Crush!\\\"\", \"old_string\": \"\\\"Hello, World!\\\"\"}, {\"new_string\": \"\\t// Greeting\\n\\tfmt.Println(\\\"Hello, Crush!\\\")\", \"old_string\": \"\\tfmt.Println(\\\"Hello, Crush!\\\")\"}], \"file_path\": \"/tmp/crush-test/TestCoderAgent/glm-5.1/multiedit_tool/main.go\"}"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-440dd811605b41f84c1517a9","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":{"prompt_tokens":110,"completion_tokens":96,"total_tokens":7278}}
+
+ data: [DONE]
+
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 1.757463708s
+- id: 3
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 31738
+ host: ""
+ body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/glm-5.1/multiedit_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use multiedit to change 'Hello, World!' to 'Hello, Crush!' and add a comment '// Greeting' above the fmt.Println line in main.go\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_8d3c\",\"function\":{\"arguments\":\"{\\\"file_path\\\": \\\"/tmp/crush-test/TestCoderAgent/glm-5.1/multiedit_tool/main.go\\\"}\",\"name\":\"view\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\tfmt.Println(\\\"Hello, World!\\\")\\n 7|}\\n\\n\",\"tool_call_id\":\"call_8d3c\",\"role\":\"tool\"},{\"tool_calls\":[{\"id\":\"call_eda6\",\"function\":{\"arguments\":\"{\\\"edits\\\": [{\\\"new_string\\\": \\\"\\\\\\\"Hello, Crush!\\\\\\\"\\\", \\\"old_string\\\": \\\"\\\\\\\"Hello, World!\\\\\\\"\\\"}, {\\\"new_string\\\": \\\"\\\\t// Greeting\\\\n\\\\tfmt.Println(\\\\\\\"Hello, Crush!\\\\\\\")\\\", \\\"old_string\\\": \\\"\\\\tfmt.Println(\\\\\\\"Hello, Crush!\\\\\\\")\\\"}], \\\"file_path\\\": \\\"/tmp/crush-test/TestCoderAgent/glm-5.1/multiedit_tool/main.go\\\"}\",\"name\":\"multiedit\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\nApplied 2 edits to file: /tmp/crush-test/TestCoderAgent/glm-5.1/multiedit_tool/main.go\\n\\n\",\"tool_call_id\":\"call_eda6\",\"role\":\"tool\"}],\"model\":\"glm-5.1\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
+ headers:
+ Accept:
+ - application/json
+ Content-Type:
+ - application/json
+ User-Agent:
+ - Charm-Crush/devel (https://charm.land/crush)
+ url: https://hyper.charm.land/v1/chat/completions
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |+
+ data: {"id":"chatcmpl-0ea8bac1183b589903b5e9a2","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-0ea8bac1183b589903b5e9a2","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_56ee","type":"function","index":0,"function":{"name":"view","arguments":""}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-0ea8bac1183b589903b5e9a2","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\"file_path\": \"/tmp/crush-test/TestCoderAgent/glm-5.1/multiedit_tool/main.go\"}"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-0ea8bac1183b589903b5e9a2","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":{"prompt_tokens":86,"completion_tokens":30,"total_tokens":7348}}
+
+ data: [DONE]
+
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 1.172073916s
+- id: 4
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 32165
+ host: ""
+ body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/glm-5.1/multiedit_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use multiedit to change 'Hello, World!' to 'Hello, Crush!' and add a comment '// Greeting' above the fmt.Println line in main.go\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_8d3c\",\"function\":{\"arguments\":\"{\\\"file_path\\\": \\\"/tmp/crush-test/TestCoderAgent/glm-5.1/multiedit_tool/main.go\\\"}\",\"name\":\"view\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\tfmt.Println(\\\"Hello, World!\\\")\\n 7|}\\n\\n\",\"tool_call_id\":\"call_8d3c\",\"role\":\"tool\"},{\"tool_calls\":[{\"id\":\"call_eda6\",\"function\":{\"arguments\":\"{\\\"edits\\\": [{\\\"new_string\\\": \\\"\\\\\\\"Hello, Crush!\\\\\\\"\\\", \\\"old_string\\\": \\\"\\\\\\\"Hello, World!\\\\\\\"\\\"}, {\\\"new_string\\\": \\\"\\\\t// Greeting\\\\n\\\\tfmt.Println(\\\\\\\"Hello, Crush!\\\\\\\")\\\", \\\"old_string\\\": \\\"\\\\tfmt.Println(\\\\\\\"Hello, Crush!\\\\\\\")\\\"}], \\\"file_path\\\": \\\"/tmp/crush-test/TestCoderAgent/glm-5.1/multiedit_tool/main.go\\\"}\",\"name\":\"multiedit\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\nApplied 2 edits to file: /tmp/crush-test/TestCoderAgent/glm-5.1/multiedit_tool/main.go\\n\\n\",\"tool_call_id\":\"call_eda6\",\"role\":\"tool\"},{\"tool_calls\":[{\"id\":\"call_56ee\",\"function\":{\"arguments\":\"{\\\"file_path\\\": \\\"/tmp/crush-test/TestCoderAgent/glm-5.1/multiedit_tool/main.go\\\"}\",\"name\":\"view\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\t// Greeting\\n 7|\\tfmt.Println(\\\"Hello, Crush!\\\")\\n 8|}\\n\\n\",\"tool_call_id\":\"call_56ee\",\"role\":\"tool\"}],\"model\":\"glm-5.1\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
+ headers:
+ Accept:
+ - application/json
+ Content-Type:
+ - application/json
+ User-Agent:
+ - Charm-Crush/devel (https://charm.land/crush)
+ url: https://hyper.charm.land/v1/chat/completions
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |+
+ data: {"id":"chatcmpl-0fbaf0015d2f45e0605b3c58","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-0fbaf0015d2f45e0605b3c58","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"Done —"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-0fbaf0015d2f45e0605b3c58","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" both changes applied:"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-0fbaf0015d2f45e0605b3c58","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" `\""},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-0fbaf0015d2f45e0605b3c58","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"Hello, World!\""},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-0fbaf0015d2f45e0605b3c58","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"` → `\"Hello"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-0fbaf0015d2f45e0605b3c58","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":", Crush!\"`"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-0fbaf0015d2f45e0605b3c58","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" and `// G"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-0fbaf0015d2f45e0605b3c58","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"reeting` comment"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-0fbaf0015d2f45e0605b3c58","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" added above the `"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-0fbaf0015d2f45e0605b3c58","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"fmt.Println` line"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-0fbaf0015d2f45e0605b3c58","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"."},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-0fbaf0015d2f45e0605b3c58","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{"prompt_tokens":86,"completion_tokens":36,"total_tokens":7450}}
+
+ data: [DONE]
+
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 1.121115959s
diff --git a/internal/agent/testdata/TestCoderAgent/glm-5.1/parallel_tool_calls.yaml b/internal/agent/testdata/TestCoderAgent/glm-5.1/parallel_tool_calls.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..44bbddaa959696a28f8274ef66acce627fb0070c
--- /dev/null
+++ b/internal/agent/testdata/TestCoderAgent/glm-5.1/parallel_tool_calls.yaml
@@ -0,0 +1,151 @@
+---
+version: 2
+interactions:
+- id: 0
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 893
+ host: ""
+ body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse glob to find all .go files and use ls to list the current directory, it is very important that you run both tool calls in parallel\n \n\n","role":"user"}],"model":"gpt-oss-120b","max_completion_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
+ headers:
+ Accept:
+ - application/json
+ Content-Type:
+ - application/json
+ User-Agent:
+ - Charm-Crush/devel (https://charm.land/crush)
+ url: https://hyper.charm.land/v1/chat/completions
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |+
+ data: {"id":"chatcmpl-b284adff2f283f37275d565b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-b284adff2f283f37275d565b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":"Parallel"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-b284adff2f283f37275d565b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" glob"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-b284adff2f283f37275d565b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" for"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-b284adff2f283f37275d565b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" ."},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-b284adff2f283f37275d565b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":"go"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-b284adff2f283f37275d565b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" files"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-b284adff2f283f37275d565b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" and"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-b284adff2f283f37275d565b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" ls"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-b284adff2f283f37275d565b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" listing"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-b284adff2f283f37275d565b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{"prompt_tokens":36,"completion_tokens":137,"total_tokens":365}}
+
+ data: [DONE]
+
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 1.718419583s
+- id: 1
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 30739
+ host: ""
+ body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/glm-5.1/parallel_tool_calls\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use glob to find all .go files and use ls to list the current directory, it is very important that you run both tool calls in parallel\",\"role\":\"user\"}],\"model\":\"glm-5.1\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
+ headers:
+ Accept:
+ - application/json
+ Content-Type:
+ - application/json
+ User-Agent:
+ - Charm-Crush/devel (https://charm.land/crush)
+ url: https://hyper.charm.land/v1/chat/completions
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |+
+ data: {"id":"chatcmpl-7bf07380999966966d56309f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-7bf07380999966966d56309f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_42b2","type":"function","index":0,"function":{"name":"glob","arguments":""}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-7bf07380999966966d56309f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\"pattern\": \"**/*.go\"}"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-7bf07380999966966d56309f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_f20c","type":"function","index":1,"function":{"name":"ls","arguments":""}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-7bf07380999966966d56309f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":1,"function":{"arguments":"{}"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-7bf07380999966966d56309f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":{"prompt_tokens":146,"completion_tokens":15,"total_tokens":7105}}
+
+ data: [DONE]
+
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 2.60407775s
+- id: 2
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 31227
+ host: ""
+ body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/glm-5.1/parallel_tool_calls\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use glob to find all .go files and use ls to list the current directory, it is very important that you run both tool calls in parallel\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_42b2\",\"function\":{\"arguments\":\"{\\\"pattern\\\": \\\"**/*.go\\\"}\",\"name\":\"glob\"},\"type\":\"function\"},{\"id\":\"call_f20c\",\"function\":{\"arguments\":\"{}\",\"name\":\"ls\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"/tmp/crush-test/TestCoderAgent/glm-5.1/parallel_tool_calls/main.go\",\"tool_call_id\":\"call_42b2\",\"role\":\"tool\"},{\"content\":\"\\n- /tmp/crush-test/TestCoderAgent/glm-5.1/parallel_tool_calls/\\n - go.mod\\n - main.go\\n\",\"tool_call_id\":\"call_f20c\",\"role\":\"tool\"}],\"model\":\"glm-5.1\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
+ headers:
+ Accept:
+ - application/json
+ Content-Type:
+ - application/json
+ User-Agent:
+ - Charm-Crush/devel (https://charm.land/crush)
+ url: https://hyper.charm.land/v1/chat/completions
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |+
+ data: {"id":"chatcmpl-45340f8dbffe1649d58caf7e","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-45340f8dbffe1649d58caf7e","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"Found "},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-45340f8dbffe1649d58caf7e","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"1 `.go`"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-45340f8dbffe1649d58caf7e","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" file: `main"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-45340f8dbffe1649d58caf7e","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":".go`. Directory"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-45340f8dbffe1649d58caf7e","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" contains `go"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-45340f8dbffe1649d58caf7e","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":".mod` and `"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-45340f8dbffe1649d58caf7e","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"main.go`."},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-45340f8dbffe1649d58caf7e","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{"prompt_tokens":58,"completion_tokens":24,"total_tokens":7186}}
+
+ data: [DONE]
+
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 873.629708ms
diff --git a/internal/agent/testdata/TestCoderAgent/glm-5.1/read_a_file.yaml b/internal/agent/testdata/TestCoderAgent/glm-5.1/read_a_file.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..e4ef2c4b28d480b33145d1c04fa74c4f3ebea79e
--- /dev/null
+++ b/internal/agent/testdata/TestCoderAgent/glm-5.1/read_a_file.yaml
@@ -0,0 +1,257 @@
+---
+version: 2
+interactions:
+- id: 0
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 774
+ host: ""
+ body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nRead the go mod\n \n\n","role":"user"}],"model":"gpt-oss-120b","max_completion_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
+ headers:
+ Accept:
+ - application/json
+ Content-Type:
+ - application/json
+ User-Agent:
+ - Charm-Crush/devel (https://charm.land/crush)
+ url: https://hyper.charm.land/v1/chat/completions
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |+
+ data: {"id":"chatcmpl-20dcf6ef9ddc91fe00c616e2","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-20dcf6ef9ddc91fe00c616e2","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":"Reading"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-20dcf6ef9ddc91fe00c616e2","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" the"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-20dcf6ef9ddc91fe00c616e2","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" go"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-20dcf6ef9ddc91fe00c616e2","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":".mod"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-20dcf6ef9ddc91fe00c616e2","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" File"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-20dcf6ef9ddc91fe00c616e2","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{"prompt_tokens":27,"completion_tokens":77,"total_tokens":280}}
+
+ data: [DONE]
+
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 1.620399166s
+- id: 1
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 30612
+ host: ""
+ body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/glm-5.1/read_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"Read the go mod\",\"role\":\"user\"}],\"model\":\"glm-5.1\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
+ headers:
+ Accept:
+ - application/json
+ Content-Type:
+ - application/json
+ User-Agent:
+ - Charm-Crush/devel (https://charm.land/crush)
+ url: https://hyper.charm.land/v1/chat/completions
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |+
+ data: {"id":"chatcmpl-0114e9f0e420befe754de287","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-0114e9f0e420befe754de287","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_Enol63d8wGn6vUwS7wBqLVGN","type":"function","index":0,"function":{"name":"glob","arguments":""}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-0114e9f0e420befe754de287","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-0114e9f0e420befe754de287","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"pattern\": \""}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-0114e9f0e420befe754de287","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"**"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-0114e9f0e420befe754de287","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-0114e9f0e420befe754de287","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"go"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-0114e9f0e420befe754de287","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".mod"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-0114e9f0e420befe754de287","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"}"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-0114e9f0e420befe754de287","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":{"prompt_tokens":151,"completion_tokens":31,"total_tokens":7094}}
+
+ data: [DONE]
+
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 6.519424875s
+- id: 2
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 30908
+ host: ""
+ body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/glm-5.1/read_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"Read the go mod\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_Enol63d8wGn6vUwS7wBqLVGN\",\"function\":{\"arguments\":\"{\\\"pattern\\\": \\\"**/go.mod\\\"}\",\"name\":\"glob\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"/tmp/crush-test/TestCoderAgent/glm-5.1/read_a_file/go.mod\",\"tool_call_id\":\"call_Enol63d8wGn6vUwS7wBqLVGN\",\"role\":\"tool\"}],\"model\":\"glm-5.1\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
+ headers:
+ Accept:
+ - application/json
+ Content-Type:
+ - application/json
+ User-Agent:
+ - Charm-Crush/devel (https://charm.land/crush)
+ url: https://hyper.charm.land/v1/chat/completions
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |+
+ data: {"id":"chatcmpl-4c8438e22d9ad3951b88da2f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-4c8438e22d9ad3951b88da2f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_Sf1ByEr9L6MC5OLC2HuCu4AV","type":"function","index":0,"function":{"name":"view","arguments":""}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-4c8438e22d9ad3951b88da2f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-4c8438e22d9ad3951b88da2f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"file_path\": \""}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-4c8438e22d9ad3951b88da2f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/tmp"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-4c8438e22d9ad3951b88da2f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/c"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-4c8438e22d9ad3951b88da2f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"rush"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-4c8438e22d9ad3951b88da2f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-test"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-4c8438e22d9ad3951b88da2f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/Test"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-4c8438e22d9ad3951b88da2f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Coder"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-4c8438e22d9ad3951b88da2f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Agent"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-4c8438e22d9ad3951b88da2f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/g"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-4c8438e22d9ad3951b88da2f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"lm"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-4c8438e22d9ad3951b88da2f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-4c8438e22d9ad3951b88da2f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"5"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-4c8438e22d9ad3951b88da2f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"."}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-4c8438e22d9ad3951b88da2f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"1"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-4c8438e22d9ad3951b88da2f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/read"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-4c8438e22d9ad3951b88da2f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_a"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-4c8438e22d9ad3951b88da2f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_file"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-4c8438e22d9ad3951b88da2f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/go"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-4c8438e22d9ad3951b88da2f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".mod"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-4c8438e22d9ad3951b88da2f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"}"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-4c8438e22d9ad3951b88da2f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":{"prompt_tokens":59,"completion_tokens":29,"total_tokens":7128}}
+
+ data: [DONE]
+
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 818.888167ms
+- id: 3
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 31278
+ host: ""
+ body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/glm-5.1/read_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"Read the go mod\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_Enol63d8wGn6vUwS7wBqLVGN\",\"function\":{\"arguments\":\"{\\\"pattern\\\": \\\"**/go.mod\\\"}\",\"name\":\"glob\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"/tmp/crush-test/TestCoderAgent/glm-5.1/read_a_file/go.mod\",\"tool_call_id\":\"call_Enol63d8wGn6vUwS7wBqLVGN\",\"role\":\"tool\"},{\"tool_calls\":[{\"id\":\"call_Sf1ByEr9L6MC5OLC2HuCu4AV\",\"function\":{\"arguments\":\"{\\\"file_path\\\": \\\"/tmp/crush-test/TestCoderAgent/glm-5.1/read_a_file/go.mod\\\"}\",\"name\":\"view\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n 1|module example.com/testproject\\n 2|\\n 3|go 1.23\\n\\n\",\"tool_call_id\":\"call_Sf1ByEr9L6MC5OLC2HuCu4AV\",\"role\":\"tool\"}],\"model\":\"glm-5.1\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
+ headers:
+ Accept:
+ - application/json
+ Content-Type:
+ - application/json
+ User-Agent:
+ - Charm-Crush/devel (https://charm.land/crush)
+ url: https://hyper.charm.land/v1/chat/completions
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |+
+ data: {"id":"chatcmpl-e3a7dcc36329d7c4064f77b1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e3a7dcc36329d7c4064f77b1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"The"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e3a7dcc36329d7c4064f77b1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" `"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e3a7dcc36329d7c4064f77b1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"go"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e3a7dcc36329d7c4064f77b1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":".mod"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e3a7dcc36329d7c4064f77b1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"`"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e3a7dcc36329d7c4064f77b1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" defines"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e3a7dcc36329d7c4064f77b1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" module"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e3a7dcc36329d7c4064f77b1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" `"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e3a7dcc36329d7c4064f77b1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"example"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e3a7dcc36329d7c4064f77b1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":".com"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e3a7dcc36329d7c4064f77b1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"/test"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e3a7dcc36329d7c4064f77b1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"project"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e3a7dcc36329d7c4064f77b1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"`"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e3a7dcc36329d7c4064f77b1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" with"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e3a7dcc36329d7c4064f77b1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" Go"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e3a7dcc36329d7c4064f77b1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" "},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e3a7dcc36329d7c4064f77b1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"1"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e3a7dcc36329d7c4064f77b1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"."},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e3a7dcc36329d7c4064f77b1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"23"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e3a7dcc36329d7c4064f77b1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"."},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e3a7dcc36329d7c4064f77b1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{"prompt_tokens":58,"completion_tokens":22,"total_tokens":7184}}
+
+ data: [DONE]
+
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 928.872084ms
diff --git a/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/simple_test.yaml b/internal/agent/testdata/TestCoderAgent/glm-5.1/simple_test.yaml
similarity index 51%
rename from internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/simple_test.yaml
rename to internal/agent/testdata/TestCoderAgent/glm-5.1/simple_test.yaml
index 22edd6179dbafaa927893bf9405ce5605b05fda7..af13dcc3633dabd9832fb5ccab1fadfe6c7466ed 100644
--- a/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/simple_test.yaml
+++ b/internal/agent/testdata/TestCoderAgent/glm-5.1/simple_test.yaml
@@ -6,9 +6,9 @@ interactions:
proto: HTTP/1.1
proto_major: 1
proto_minor: 1
- content_length: 798
+ content_length: 764
host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nHello\n \n\n","role":"user"}],"model":"qwen/qwen3-next-80b-a3b-instruct","max_tokens":40,"stream_options":{"include_usage":true},"usage":{"include":true},"stream":true}'
+ body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nHello\n \n\n","role":"user"}],"model":"gpt-oss-120b","max_completion_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
headers:
Accept:
- application/json
@@ -16,23 +16,22 @@ interactions:
- application/json
User-Agent:
- Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
+ url: https://hyper.charm.land/v1/chat/completions
method: POST
response:
proto: HTTP/2.0
proto_major: 2
proto_minor: 0
content_length: -1
+ uncompressed: true
body: |+
- : OPENROUTER PROCESSING
+ data: {"id":"chatcmpl-f342b662fa22d62db967957b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
- : OPENROUTER PROCESSING
+ data: {"id":"chatcmpl-f342b662fa22d62db967957b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":"Simple"},"finish_reason":null}]}
- data: {"id":"gen-1776258045-M14anX94IZdtXfmbrZ6w","object":"chat.completion.chunk","created":1776258045,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Google","choices":[{"index":0,"delta":{"content":"Hello","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
+ data: {"id":"chatcmpl-f342b662fa22d62db967957b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" Greeting"},"finish_reason":null}]}
- data: {"id":"gen-1776258045-M14anX94IZdtXfmbrZ6w","object":"chat.completion.chunk","created":1776258045,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Google","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}]}
-
- data: {"id":"gen-1776258045-M14anX94IZdtXfmbrZ6w","object":"chat.completion.chunk","created":1776258045,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Google","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}],"usage":{"prompt_tokens":144,"completion_tokens":2,"total_tokens":146,"cost":0.000024,"is_byok":false,"prompt_tokens_details":{"cached_tokens":0,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.000024,"upstream_inference_prompt_cost":0.0000216,"upstream_inference_completions_cost":0.0000024},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
+ data: {"id":"chatcmpl-f342b662fa22d62db967957b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{"prompt_tokens":152,"completion_tokens":88,"total_tokens":288}}
data: [DONE]
@@ -41,15 +40,15 @@ interactions:
- text/event-stream
status: 200 OK
code: 200
- duration: 2.353106417s
+ duration: 1.80550475s
- id: 1
request:
proto: HTTP/1.1
proto_major: 1
proto_minor: 1
- content_length: 30765
+ content_length: 30602
host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/simple_test\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"Hello\",\"role\":\"user\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
+ body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/glm-5.1/simple_test\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"Hello\",\"role\":\"user\"}],\"model\":\"glm-5.1\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
headers:
Accept:
- application/json
@@ -57,33 +56,36 @@ interactions:
- application/json
User-Agent:
- Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
+ url: https://hyper.charm.land/v1/chat/completions
method: POST
response:
proto: HTTP/2.0
proto_major: 2
proto_minor: 0
content_length: -1
+ uncompressed: true
body: |+
- : OPENROUTER PROCESSING
+ data: {"id":"chatcmpl-ce8c02cd21955b76f9057498","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-ce8c02cd21955b76f9057498","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"Hi"},"finish_reason":null}]}
- : OPENROUTER PROCESSING
+ data: {"id":"chatcmpl-ce8c02cd21955b76f9057498","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"!"},"finish_reason":null}]}
- : OPENROUTER PROCESSING
+ data: {"id":"chatcmpl-ce8c02cd21955b76f9057498","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" How"},"finish_reason":null}]}
- : OPENROUTER PROCESSING
+ data: {"id":"chatcmpl-ce8c02cd21955b76f9057498","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" can"},"finish_reason":null}]}
- : OPENROUTER PROCESSING
+ data: {"id":"chatcmpl-ce8c02cd21955b76f9057498","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" I"},"finish_reason":null}]}
- : OPENROUTER PROCESSING
+ data: {"id":"chatcmpl-ce8c02cd21955b76f9057498","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" help"},"finish_reason":null}]}
- data: {"id":"gen-1776258045-wsJcL26GBcpIkWZx0CHY","object":"chat.completion.chunk","created":1776258045,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"Hello","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
+ data: {"id":"chatcmpl-ce8c02cd21955b76f9057498","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" you"},"finish_reason":null}]}
- data: {"id":"gen-1776258045-wsJcL26GBcpIkWZx0CHY","object":"chat.completion.chunk","created":1776258045,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":".","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
+ data: {"id":"chatcmpl-ce8c02cd21955b76f9057498","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" today"},"finish_reason":null}]}
- data: {"id":"gen-1776258045-wsJcL26GBcpIkWZx0CHY","object":"chat.completion.chunk","created":1776258045,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}]}
+ data: {"id":"chatcmpl-ce8c02cd21955b76f9057498","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"?"},"finish_reason":null}]}
- data: {"id":"gen-1776258045-wsJcL26GBcpIkWZx0CHY","object":"chat.completion.chunk","created":1776258045,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}],"usage":{"prompt_tokens":6483,"completion_tokens":3,"total_tokens":6486,"cost":0.0038973,"is_byok":false,"prompt_tokens_details":{"cached_tokens":0,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0038973,"upstream_inference_prompt_cost":0.0038898,"upstream_inference_completions_cost":0.0000075},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
+ data: {"id":"chatcmpl-ce8c02cd21955b76f9057498","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{"prompt_tokens":7058,"completion_tokens":24,"total_tokens":7082}}
data: [DONE]
@@ -92,4 +94,4 @@ interactions:
- text/event-stream
status: 200 OK
code: 200
- duration: 4.60443725s
+ duration: 2.985252875s
diff --git a/internal/agent/testdata/TestCoderAgent/openai-gpt-5/grep_tool.yaml b/internal/agent/testdata/TestCoderAgent/glm-5.1/sourcegraph_tool.yaml
similarity index 50%
rename from internal/agent/testdata/TestCoderAgent/openai-gpt-5/grep_tool.yaml
rename to internal/agent/testdata/TestCoderAgent/glm-5.1/sourcegraph_tool.yaml
index 60e48ade4228103370a20b7f04b9740523c5c024..a1bb01ad174b0bcb07227bd34f1e0543b27cd883 100644
--- a/internal/agent/testdata/TestCoderAgent/openai-gpt-5/grep_tool.yaml
+++ b/internal/agent/testdata/TestCoderAgent/glm-5.1/sourcegraph_tool.yaml
@@ -6,9 +6,9 @@ interactions:
proto: HTTP/1.1
proto_major: 1
proto_minor: 1
- content_length: 795
+ content_length: 819
host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse grep to search for the word ''package'' in go files\n \n\n","role":"user"}],"model":"gpt-4o","max_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
+ body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse sourcegraph to search for ''func main'' in Go repositories\n \n\n","role":"user"}],"model":"gpt-oss-120b","max_completion_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
headers:
Accept:
- application/json
@@ -16,48 +16,47 @@ interactions:
- application/json
User-Agent:
- Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
+ url: https://hyper.charm.land/v1/chat/completions
method: POST
response:
proto: HTTP/2.0
proto_major: 2
proto_minor: 0
content_length: -1
+ uncompressed: true
body: |+
- data: {"id":"chatcmpl-DUu848sQt4Ep7RV7uVnyx102d94YI","object":"chat.completion.chunk","created":1776257972,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"role":"assistant","content":"","refusal":null},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"uxNecKgAuOkR7p"}
+ data: {"id":"chatcmpl-620058f24b10a79ade1c095b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu848sQt4Ep7RV7uVnyx102d94YI","object":"chat.completion.chunk","created":1776257972,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":"Using"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"lGHYXKjoiKU"}
+ data: {"id":"chatcmpl-620058f24b10a79ade1c095b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":"Search"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu848sQt4Ep7RV7uVnyx102d94YI","object":"chat.completion.chunk","created":1776257972,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" grep"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"R1Pistrcm7c"}
+ data: {"id":"chatcmpl-620058f24b10a79ade1c095b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" for"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu848sQt4Ep7RV7uVnyx102d94YI","object":"chat.completion.chunk","created":1776257972,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" to"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"uTzT0W9oK8xlh"}
+ data: {"id":"chatcmpl-620058f24b10a79ade1c095b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" func"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu848sQt4Ep7RV7uVnyx102d94YI","object":"chat.completion.chunk","created":1776257972,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" find"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"55aUOVWEWre"}
+ data: {"id":"chatcmpl-620058f24b10a79ade1c095b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" main"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu848sQt4Ep7RV7uVnyx102d94YI","object":"chat.completion.chunk","created":1776257972,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" '"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"pS156bwGCUwr0i"}
+ data: {"id":"chatcmpl-620058f24b10a79ade1c095b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" in"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu848sQt4Ep7RV7uVnyx102d94YI","object":"chat.completion.chunk","created":1776257972,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":"package"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"7Dh6cgrtx"}
+ data: {"id":"chatcmpl-620058f24b10a79ade1c095b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" Go"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu848sQt4Ep7RV7uVnyx102d94YI","object":"chat.completion.chunk","created":1776257972,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":"'"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"FSDGrXERmN2EBLT"}
+ data: {"id":"chatcmpl-620058f24b10a79ade1c095b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" repos"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu848sQt4Ep7RV7uVnyx102d94YI","object":"chat.completion.chunk","created":1776257972,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" in"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"rJSDz5sFx6nrb"}
+ data: {"id":"chatcmpl-620058f24b10a79ade1c095b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" via"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu848sQt4Ep7RV7uVnyx102d94YI","object":"chat.completion.chunk","created":1776257972,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" Go"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"PQLYXTtYVqeIm"}
+ data: {"id":"chatcmpl-620058f24b10a79ade1c095b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" Source"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu848sQt4Ep7RV7uVnyx102d94YI","object":"chat.completion.chunk","created":1776257972,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" files"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"TgVKN33UYi"}
+ data: {"id":"chatcmpl-620058f24b10a79ade1c095b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":"graph"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu848sQt4Ep7RV7uVnyx102d94YI","object":"chat.completion.chunk","created":1776257972,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{},"logprobs":null,"finish_reason":"stop"}],"usage":null,"obfuscation":"6hGyE1De9N"}
-
- data: {"id":"chatcmpl-DUu848sQt4Ep7RV7uVnyx102d94YI","object":"chat.completion.chunk","created":1776257972,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[],"usage":{"prompt_tokens":154,"completion_tokens":10,"total_tokens":164,"prompt_tokens_details":{"cached_tokens":0,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":0,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"Ar0GwunAmv3vi"}
+ data: {"id":"chatcmpl-620058f24b10a79ade1c095b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{"prompt_tokens":36,"completion_tokens":323,"total_tokens":535}}
data: [DONE]
headers:
Content-Type:
- - text/event-stream; charset=utf-8
+ - text/event-stream
status: 200 OK
code: 200
- duration: 558.148708ms
+ duration: 3.6349505s
- id: 1
request:
proto: HTTP/1.1
@@ -65,7 +64,7 @@ interactions:
proto_minor: 1
content_length: 30662
host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/grep_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use grep to search for the word 'package' in go files\",\"role\":\"user\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
+ body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/glm-5.1/sourcegraph_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use sourcegraph to search for 'func main' in Go repositories\",\"role\":\"user\"}],\"model\":\"glm-5.1\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
headers:
Accept:
- application/json
@@ -73,182 +72,230 @@ interactions:
- application/json
User-Agent:
- Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
+ url: https://hyper.charm.land/v1/chat/completions
method: POST
response:
proto: HTTP/2.0
proto_major: 2
proto_minor: 0
content_length: -1
+ uncompressed: true
body: |+
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":null,"tool_calls":[{"index":0,"id":"call_qSoKEzV62FVDk2T32GPTxo95","type":"function","function":{"name":"grep","arguments":""}}],"refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"HPUThbxFXKeA"}
+ data: {"id":"chatcmpl-16dbeb5ad8ac324087308be1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"VFS3oLmd9CqUrtM"}
+ data: {"id":"chatcmpl-16dbeb5ad8ac324087308be1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_7f06","type":"function","index":0,"function":{"name":"sourcegraph","arguments":""}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"path"}}]},"finish_reason":null}],"usage":null,"obfuscation":"UKmKodpEli7bHJ"}
+ data: {"id":"chatcmpl-16dbeb5ad8ac324087308be1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\"count\": 20, \"query\": \"func main lang:go\"}"}}]},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"KZ8qE4KmaZsbLzL"}
+ data: {"id":"chatcmpl-16dbeb5ad8ac324087308be1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":{"prompt_tokens":129,"completion_tokens":41,"total_tokens":7114}}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"/"}}]},"finish_reason":null}],"usage":null,"obfuscation":"8zOBamDv10ygnEA"}
+ data: [DONE]
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"tmp"}}]},"finish_reason":null}],"usage":null,"obfuscation":"UiuwCZlS18haIvy"}
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 4.78124025s
+- id: 2
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 424
+ host: sourcegraph.com
+ body: '{"query":"query Search($query: String!) { search(query: $query, version: V2, patternType: keyword ) { results { matchCount, limitHit, resultCount, approximateResultCount, missing { name }, timedout { name }, indexUnavailable, results { __typename, ... on FileMatch { repository { name }, file { path, url, content }, lineMatches { preview, lineNumber, offsetAndLengths } } } } } }","variables":{"query":"func main lang:go"}}'
+ headers:
+ Content-Type:
+ - application/json
+ User-Agent:
+ - crush/1.0
+ url: https://sourcegraph.com/.api/graphql
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: '{"data":{"search":{"results":{"matchCount":30,"limitHit":true,"resultCount":30,"approximateResultCount":"30+","missing":[],"timedout":[],"indexUnavailable":false,"results":[{"__typename":"FileMatch","repository":{"name":"github.com/GoesToEleven/GolangTraining"},"file":{"path":"14_functions/10_func-expression/01_before-func-expression/main.go","url":"/r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/01_before-func-expression/main.go","content":"package main\n\nimport \"fmt\"\n\nfunc greeting() {\n\tfmt.Println(\"Hello world!\")\n}\n\nfunc main() {\n\tgreeting()\n}\n"},"lineMatches":[{"preview":"func main() {","lineNumber":8,"offsetAndLengths":[[0,9]]},{"preview":"package main","lineNumber":0,"offsetAndLengths":[[8,4]]},{"preview":"func greeting() {","lineNumber":4,"offsetAndLengths":[[0,4]]}]},{"__typename":"FileMatch","repository":{"name":"github.com/GoesToEleven/GolangTraining"},"file":{"path":"14_functions/10_func-expression/02_func-expression/main.go","url":"/r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/02_func-expression/main.go","content":"package main\n\nimport \"fmt\"\n\nfunc main() {\n\n\tgreeting := func() {\n\t\tfmt.Println(\"Hello world!\")\n\t}\n\n\tgreeting()\n}\n"},"lineMatches":[{"preview":"func main() {","lineNumber":4,"offsetAndLengths":[[0,9]]},{"preview":"\tgreeting := func() {","lineNumber":6,"offsetAndLengths":[[13,4]]},{"preview":"package main","lineNumber":0,"offsetAndLengths":[[8,4]]}]},{"__typename":"FileMatch","repository":{"name":"github.com/GoesToEleven/GolangTraining"},"file":{"path":"14_functions/10_func-expression/03_func-expression_shows-type/main.go","url":"/r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/03_func-expression_shows-type/main.go","content":"package main\n\nimport \"fmt\"\n\nfunc main() {\n\n\tgreeting := func() {\n\t\tfmt.Println(\"Hello world!\")\n\t}\n\n\tgreeting()\n\tfmt.Printf(\"%T\\n\", greeting)\n}\n"},"lineMatches":[{"preview":"func main() {","lineNumber":4,"offsetAndLengths":[[0,9]]},{"preview":"\tgreeting := func() {","lineNumber":6,"offsetAndLengths":[[13,4]]},{"preview":"package main","lineNumber":0,"offsetAndLengths":[[8,4]]}]},{"__typename":"FileMatch","repository":{"name":"github.com/GoesToEleven/GolangTraining"},"file":{"path":"14_functions/10_func-expression/04_another-way_func-expression/main.go","url":"/r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/04_another-way_func-expression/main.go","content":"package main\n\nimport \"fmt\"\n\nfunc makeGreeter() func() string {\n\treturn func() string {\n\t\treturn \"Hello world!\"\n\t}\n}\n\nfunc main() {\n\tgreet := makeGreeter()\n\tfmt.Println(greet())\n}\n"},"lineMatches":[{"preview":"func main() {","lineNumber":10,"offsetAndLengths":[[0,9]]},{"preview":"package main","lineNumber":0,"offsetAndLengths":[[8,4]]},{"preview":"func makeGreeter() func() string {","lineNumber":4,"offsetAndLengths":[[0,4],[19,4]]},{"preview":"\treturn func() string {","lineNumber":5,"offsetAndLengths":[[8,4]]}]},{"__typename":"FileMatch","repository":{"name":"github.com/jmbharathram/executeoncommand"},"file":{"path":"go/channel/main.go","url":"/r/github.com/jmbharathram/executeoncommand/-/blob/go/channel/main.go","content":"package main\n\nimport \"fmt\"\n\n// Example 0\nfunc dummy_func() {\n\tfmt.Println(\"dummy_func\")\n}\nfunc main() {\n\tgo dummy_func()\n fmt.Println(\"end of main\")\n}\n\n\n// Example 1\n\npackage main\n\nimport \"fmt\"\n\nfunc dummy_func(ch chan int) {\n // receive a value from channel\n val := \u003c-ch\n\tfmt.Println(val)\n}\nfunc main() {\n // create a channel\n\tch := make(chan int)\n // trigger a parallel go routine\n\tgo dummy_func(ch)\n // send a value into the channel\n\tch \u003c- 10\n}\n\n// Example 2\n\npackage main\n\nimport (\n \"fmt\"\n //\"flag\"\n)\n\nfunc dummy_func(ch chan int) {\n // receive a value from channel\n val := \u003c-ch\n\tfmt.Println(val)\n\n //send a value again\n ch \u003c- val*2\n}\nfunc main() {\n fmt.Println(\"Welcome to main function\")\n \n //flag statements for user input\n var input_value int\n input_value = 10\n //flag.IntVar(\u0026input_value, \"v\", 10, \"Input value to send/receive via channel.\" )\n //flag.Parse()\n\n // create a channel\n\tch := make(chan int)\n\n // trigger a parallel go routine\n\tgo dummy_func(ch)\n \n // send a value into the channel\n\tch \u003c- input_value\n\n // receive a value from channel\n response := \u003c- ch\n fmt.Println(response)\n\n}\n\n// example 3\n\n\npackage main\n\nimport(\n\"fmt\"\n\"time\"\n\"math/rand\"\n)\n\nfunc routine1(ch chan int, min int, max int) {\n\n rand.Seed(time.Now().UnixNano())\n t := rand.Intn(max - min) + min\n fmt.Println(\"routine 1 is waiting: \", t, \"secs\")\n time.Sleep(time.Duration(t)*time.Second)\n\n ch \u003c- t\n}\n\nfunc routine2(ch chan int, min int, max int) {\n\n rand.Seed(time.Now().UnixNano())\n t := rand.Intn(max - min) + min\n fmt.Println(\"routine 2 is waiting: \", t, \"secs\")\n time.Sleep(time.Duration(t)*time.Second)\n\n ch \u003c- t\n}\n\nfunc main(){\n\nmin := 1\nmax := 10\n\nch1:= make(chan int)\nch2:= make(chan int)\n\ngo routine1(ch1, min, max)\ngo routine2(ch2, min, max)\n\nselect{\n\ncase \u003c- ch1:\n fmt.Println(\"routine 1 finished\")\n\ncase \u003c- ch2:\n fmt.Println(\"routine 2 finished\")\n}\n\nfmt.Println(\"end of main function\")\n\n}\n"},"lineMatches":[{"preview":"func main(){","lineNumber":106,"offsetAndLengths":[[0,9]]},{"preview":"func dummy_func() {","lineNumber":5,"offsetAndLengths":[[0,4],[11,4]]},{"preview":"\tfmt.Println(\"dummy_func\")","lineNumber":6,"offsetAndLengths":[[20,4]]},{"preview":"func main() {","lineNumber":8,"offsetAndLengths":[[0,9]]},{"preview":"\tgo dummy_func()","lineNumber":9,"offsetAndLengths":[[10,4]]},{"preview":" fmt.Println(\"end of main\")","lineNumber":10,"offsetAndLengths":[[24,4]]},{"preview":"func main() {","lineNumber":25,"offsetAndLengths":[[0,9]]},{"preview":"func main() {","lineNumber":51,"offsetAndLengths":[[0,9]]},{"preview":" fmt.Println(\"Welcome to main function\")","lineNumber":52,"offsetAndLengths":[[28,4],[33,4]]},{"preview":"fmt.Println(\"end of main function\")","lineNumber":126,"offsetAndLengths":[[20,4],[25,4]]},{"preview":"package main","lineNumber":78,"offsetAndLengths":[[8,4]]},{"preview":"package main","lineNumber":16,"offsetAndLengths":[[8,4]]},{"preview":"package main","lineNumber":0,"offsetAndLengths":[[8,4]]}]}]}}}}'
+ headers:
+ Content-Type:
+ - application/json
+ status: 200 OK
+ code: 200
+ duration: 1.519403333s
+- id: 3
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 39935
+ host: ""
+ body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/glm-5.1/sourcegraph_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use sourcegraph to search for 'func main' in Go repositories\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_7f06\",\"function\":{\"arguments\":\"{\\\"count\\\": 20, \\\"query\\\": \\\"func main lang:go\\\"}\",\"name\":\"sourcegraph\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"# Sourcegraph Search Results\\n\\nFound 30 matches across 30 results\\n(Result limit reached, try a more specific query)\\n\\n## Result 1: github.com/GoesToEleven/GolangTraining/14_functions/10_func-expression/01_before-func-expression/main.go\\n\\nURL: /r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/01_before-func-expression/main.go\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func greeting() {\\n6| \\tfmt.Println(\\\"Hello world!\\\")\\n7| }\\n8| func main() {\\n9| func main() {\\n10| \\tgreeting()\\n11| }\\n12| \\n```\\n\\n```\\n0| package main\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func greeting() {\\n6| \\tfmt.Println(\\\"Hello world!\\\")\\n7| }\\n8| \\n9| func main() {\\n10| \\tgreeting()\\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| func greeting() {\\n5| func greeting() {\\n6| \\tfmt.Println(\\\"Hello world!\\\")\\n7| }\\n8| \\n9| func main() {\\n10| \\tgreeting()\\n11| }\\n12| \\n```\\n\\n## Result 2: github.com/GoesToEleven/GolangTraining/14_functions/10_func-expression/02_func-expression/main.go\\n\\nURL: /r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/02_func-expression/main.go\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| func main() {\\n5| func main() {\\n6| \\n7| \\tgreeting := func() {\\n8| \\t\\tfmt.Println(\\\"Hello world!\\\")\\n9| \\t}\\n10| \\n11| \\tgreeting()\\n12| }\\n13| \\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func main() {\\n6| \\tgreeting := func() {\\n7| \\tgreeting := func() {\\n8| \\t\\tfmt.Println(\\\"Hello world!\\\")\\n9| \\t}\\n10| \\n11| \\tgreeting()\\n12| }\\n13| \\n```\\n\\n```\\n0| package main\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func main() {\\n6| \\n7| \\tgreeting := func() {\\n8| \\t\\tfmt.Println(\\\"Hello world!\\\")\\n9| \\t}\\n10| \\n```\\n\\n## Result 3: github.com/GoesToEleven/GolangTraining/14_functions/10_func-expression/03_func-expression_shows-type/main.go\\n\\nURL: /r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/03_func-expression_shows-type/main.go\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| func main() {\\n5| func main() {\\n6| \\n7| \\tgreeting := func() {\\n8| \\t\\tfmt.Println(\\\"Hello world!\\\")\\n9| \\t}\\n10| \\n11| \\tgreeting()\\n12| \\tfmt.Printf(\\\"%T\\\\n\\\", greeting)\\n13| }\\n14| \\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func main() {\\n6| \\tgreeting := func() {\\n7| \\tgreeting := func() {\\n8| \\t\\tfmt.Println(\\\"Hello world!\\\")\\n9| \\t}\\n10| \\n11| \\tgreeting()\\n12| \\tfmt.Printf(\\\"%T\\\\n\\\", greeting)\\n13| }\\n14| \\n```\\n\\n```\\n0| package main\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func main() {\\n6| \\n7| \\tgreeting := func() {\\n8| \\t\\tfmt.Println(\\\"Hello world!\\\")\\n9| \\t}\\n10| \\n```\\n\\n## Result 4: github.com/GoesToEleven/GolangTraining/14_functions/10_func-expression/04_another-way_func-expression/main.go\\n\\nURL: /r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/04_another-way_func-expression/main.go\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func makeGreeter() func() string {\\n6| \\treturn func() string {\\n7| \\t\\treturn \\\"Hello world!\\\"\\n8| \\t}\\n9| }\\n10| func main() {\\n11| func main() {\\n12| \\tgreet := makeGreeter()\\n13| \\tfmt.Println(greet())\\n14| }\\n15| \\n```\\n\\n```\\n0| package main\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func makeGreeter() func() string {\\n6| \\treturn func() string {\\n7| \\t\\treturn \\\"Hello world!\\\"\\n8| \\t}\\n9| }\\n10| \\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| func makeGreeter() func() string {\\n5| func makeGreeter() func() string {\\n6| \\treturn func() string {\\n7| \\t\\treturn \\\"Hello world!\\\"\\n8| \\t}\\n9| }\\n10| \\n11| func main() {\\n12| \\tgreet := makeGreeter()\\n13| \\tfmt.Println(greet())\\n14| }\\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| \\treturn func() string {\\n6| \\treturn func() string {\\n7| \\t\\treturn \\\"Hello world!\\\"\\n8| \\t}\\n9| }\\n10| \\n11| func main() {\\n12| \\tgreet := makeGreeter()\\n13| \\tfmt.Println(greet())\\n14| }\\n15| \\n```\\n\\n## Result 5: github.com/jmbharathram/executeoncommand/go/channel/main.go\\n\\nURL: /r/github.com/jmbharathram/executeoncommand/-/blob/go/channel/main.go\\n\\n```\\n96| \\n97| func routine2(ch chan int, min int, max int) {\\n98| \\n99| rand.Seed(time.Now().UnixNano())\\n100| t := rand.Intn(max - min) + min\\n101| fmt.Println(\\\"routine 2 is waiting: \\\", t, \\\"secs\\\")\\n102| time.Sleep(time.Duration(t)*time.Second)\\n103| \\n104| ch <- t\\n105| }\\n106| func main(){\\n107| func main(){\\n108| \\n109| min := 1\\n110| max := 10\\n111| \\n112| ch1:= make(chan int)\\n113| ch2:= make(chan int)\\n114| \\n115| go routine1(ch1, min, max)\\n116| go routine2(ch2, min, max)\\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func dummy_func() {\\n6| func dummy_func() {\\n7| \\tfmt.Println(\\\"dummy_func\\\")\\n8| }\\n9| func main() {\\n10| \\tgo dummy_func()\\n11| fmt.Println(\\\"end of main\\\")\\n12| }\\n13| \\n14| \\n15| // Example 1\\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| // Example 0\\n6| \\tfmt.Println(\\\"dummy_func\\\")\\n7| \\tfmt.Println(\\\"dummy_func\\\")\\n8| }\\n9| func main() {\\n10| \\tgo dummy_func()\\n11| fmt.Println(\\\"end of main\\\")\\n12| }\\n13| \\n14| \\n15| // Example 1\\n16| \\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| // Example 0\\n6| func dummy_func() {\\n7| \\tfmt.Println(\\\"dummy_func\\\")\\n8| func main() {\\n9| func main() {\\n10| \\tgo dummy_func()\\n11| fmt.Println(\\\"end of main\\\")\\n12| }\\n13| \\n14| \\n15| // Example 1\\n16| \\n17| package main\\n18| \\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| // Example 0\\n6| func dummy_func() {\\n7| \\tfmt.Println(\\\"dummy_func\\\")\\n8| }\\n9| \\tgo dummy_func()\\n10| \\tgo dummy_func()\\n11| fmt.Println(\\\"end of main\\\")\\n12| }\\n13| \\n14| \\n15| // Example 1\\n16| \\n17| package main\\n18| \\n19| import \\\"fmt\\\"\\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| // Example 0\\n6| func dummy_func() {\\n7| \\tfmt.Println(\\\"dummy_func\\\")\\n8| }\\n9| func main() {\\n10| fmt.Println(\\\"end of main\\\")\\n11| fmt.Println(\\\"end of main\\\")\\n12| }\\n13| \\n14| \\n15| // Example 1\\n16| \\n17| package main\\n18| \\n19| import \\\"fmt\\\"\\n20| \\n```\\n\\n```\\n15| // Example 1\\n16| \\n17| package main\\n18| \\n19| import \\\"fmt\\\"\\n20| \\n21| func dummy_func(ch chan int) {\\n22| // receive a value from channel\\n23| val := <-ch\\n24| \\tfmt.Println(val)\\n25| func main() {\\n26| func main() {\\n27| // create a channel\\n28| \\tch := make(chan int)\\n29| // trigger a parallel go routine\\n30| \\tgo dummy_func(ch)\\n31| // send a value into the channel\\n32| \\tch <- 10\\n33| }\\n34| \\n35| // Example 2\\n```\\n\\n```\\n41| //\\\"flag\\\"\\n42| )\\n43| \\n44| func dummy_func(ch chan int) {\\n45| // receive a value from channel\\n46| val := <-ch\\n47| \\tfmt.Println(val)\\n48| \\n49| //send a value again\\n50| ch <- val*2\\n51| func main() {\\n52| func main() {\\n53| fmt.Println(\\\"Welcome to main function\\\")\\n54| \\n55| //flag statements for user input\\n56| var input_value int\\n57| input_value = 10\\n58| //flag.IntVar(&input_value, \\\"v\\\", 10, \\\"Input value to send/receive via channel.\\\" )\\n59| //flag.Parse()\\n60| \\n61| // create a channel\\n```\\n\\n```\\n42| )\\n43| \\n44| func dummy_func(ch chan int) {\\n45| // receive a value from channel\\n46| val := <-ch\\n47| \\tfmt.Println(val)\\n48| \\n49| //send a value again\\n50| ch <- val*2\\n51| }\\n52| fmt.Println(\\\"Welcome to main function\\\")\\n53| fmt.Println(\\\"Welcome to main function\\\")\\n54| \\n55| //flag statements for user input\\n56| var input_value int\\n57| input_value = 10\\n58| //flag.IntVar(&input_value, \\\"v\\\", 10, \\\"Input value to send/receive via channel.\\\" )\\n59| //flag.Parse()\\n60| \\n61| // create a channel\\n62| \\tch := make(chan int)\\n```\\n\\n```\\n116| go routine2(ch2, min, max)\\n117| \\n118| select{\\n119| \\n120| case <- ch1:\\n121| fmt.Println(\\\"routine 1 finished\\\")\\n122| \\n123| case <- ch2:\\n124| fmt.Println(\\\"routine 2 finished\\\")\\n125| }\\n126| fmt.Println(\\\"end of main function\\\")\\n127| fmt.Println(\\\"end of main function\\\")\\n128| \\n129| }\\n130| \\n```\\n\\n```\\n68| \\tch <- input_value\\n69| \\n70| // receive a value from channel\\n71| response := <- ch\\n72| fmt.Println(response)\\n73| \\n74| }\\n75| \\n76| // example 3\\n77| \\n78| package main\\n79| package main\\n80| \\n81| import(\\n82| \\\"fmt\\\"\\n83| \\\"time\\\"\\n84| \\\"math/rand\\\"\\n85| )\\n86| \\n87| func routine1(ch chan int, min int, max int) {\\n88| \\n```\\n\\n```\\n6| func dummy_func() {\\n7| \\tfmt.Println(\\\"dummy_func\\\")\\n8| }\\n9| func main() {\\n10| \\tgo dummy_func()\\n11| fmt.Println(\\\"end of main\\\")\\n12| }\\n13| \\n14| \\n15| // Example 1\\n16| package main\\n17| package main\\n18| \\n19| import \\\"fmt\\\"\\n20| \\n21| func dummy_func(ch chan int) {\\n22| // receive a value from channel\\n23| val := <-ch\\n24| \\tfmt.Println(val)\\n25| }\\n26| func main() {\\n```\\n\\n```\\n0| package main\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| // Example 0\\n6| func dummy_func() {\\n7| \\tfmt.Println(\\\"dummy_func\\\")\\n8| }\\n9| func main() {\\n10| \\tgo dummy_func()\\n```\\n\\n\",\"tool_call_id\":\"call_7f06\",\"role\":\"tool\"}],\"model\":\"glm-5.1\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
+ headers:
+ Accept:
+ - application/json
+ Content-Type:
+ - application/json
+ User-Agent:
+ - Charm-Crush/devel (https://charm.land/crush)
+ url: https://hyper.charm.land/v1/chat/completions
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |+
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/cr"}}]},"finish_reason":null}],"usage":null,"obfuscation":"euVNTINMdqHJqMN"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"Found **"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ush"}}]},"finish_reason":null}],"usage":null,"obfuscation":"lYaT6fQ0R2wmlwA"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"30 matches** across"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"Cb1P64DS0sG2M"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" 30"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/Test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"Qar8OseluYpxn"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" Go repositories."},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Coder"}}]},"finish_reason":null}],"usage":null,"obfuscation":"EXNcpSrQedUrf"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" Here are some highlights"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Agent"}}]},"finish_reason":null}],"usage":null,"obfuscation":"WCxSE0ADzYBIt"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":":\n\n| #"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/open"}}]},"finish_reason":null}],"usage":null,"obfuscation":"OAEjS9QhzpKlD"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" | Repository"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ai"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" | File |\n"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-g"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"|---|-----------"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"pt"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"|------"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-"}}]},"finish_reason":null}],"usage":null,"obfuscation":"c"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"|\n| "},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"5"}}]},"finish_reason":null}],"usage":null,"obfuscation":"6"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"1 | `G"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/"}}]},"finish_reason":null}],"usage":null,"obfuscation":"q"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"oesToEleven"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"grep"}}]},"finish_reason":null}],"usage":null,"obfuscation":"6E5oRNSP8e7ugF"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"/GolangTraining`"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_tool"}}]},"finish_reason":null}],"usage":null,"obfuscation":"a38PAU2H5STf7"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" | `14_functions"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\",\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"y3DarH4nNPpwG"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"/10_func-expression"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"pattern"}}]},"finish_reason":null}],"usage":null,"obfuscation":"AuOVQBaMryd"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"/01_before-f"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"6pkMRxDhmQhJW"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"unc-expression/main.go"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"package"}}]},"finish_reason":null}],"usage":null,"obfuscation":"7RcuwebuIvO"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"` |\n| "},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\",\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"wXKywwgW4OYww"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"2 | `G"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"include"}}]},"finish_reason":null}],"usage":null,"obfuscation":"vCJMRuXYJyx"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"oesToEleven"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"MqM7kmErFEjUW"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"/GolangTraining`"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"*."}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" | `14_functions"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"go"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"/10_func-expression"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\",\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"hgxglhHwcrkpJ"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"/02_func-expression"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"literal"}}]},"finish_reason":null}],"usage":null,"obfuscation":"3QmLNouMWKG"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"/main.go` |\n"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_text"}}]},"finish_reason":null}],"usage":null,"obfuscation":"6cQOB9D3FHs0n"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"| 3 |"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"A02yueH3107yPz1"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" `GoesTo"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"true"}}]},"finish_reason":null}],"usage":null,"obfuscation":"1H3aQI3rIX46mA"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"Eleven/Golang"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"}"}}]},"finish_reason":null}],"usage":null,"obfuscation":"H"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"Training` | `"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":null,"obfuscation":""}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"14_functions/10"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu84nO266uam89WksGodDIfF8dlz","object":"chat.completion.chunk","created":1776257972,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6378,"completion_tokens":309,"total_tokens":6687,"prompt_tokens_details":{"cached_tokens":3456,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":256,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"O"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"_func-expression/03"},"finish_reason":null}]}
- data: [DONE]
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"_func-expression_shows"},"finish_reason":null}]}
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 5.718010542s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31121
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/grep_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use grep to search for the word 'package' in go files\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_qSoKEzV62FVDk2T32GPTxo95\",\"function\":{\"arguments\":\"{\\\"path\\\":\\\"/tmp/crush-test/TestCoderAgent/openai-gpt-5/grep_tool\\\",\\\"pattern\\\":\\\"package\\\",\\\"include\\\":\\\"*.go\\\",\\\"literal_text\\\":true}\",\"name\":\"grep\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"Found 1 matches\\n/tmp/crush-test/TestCoderAgent/openai-gpt-5/grep_tool/main.go:\\n Line 1, Char 1: package main\\n\",\"tool_call_id\":\"call_qSoKEzV62FVDk2T32GPTxo95\",\"role\":\"tool\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu8AvZGmiqsr5cTYfc8X4l2m2dcq","object":"chat.completion.chunk","created":1776257978,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":"","refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"E4X9Gx23xu"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"-type/main.go`"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" |\n| 4"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" | `Goes"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"ToEleven/G"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"olangTraining` |"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" `14_functions/"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"10_func-expression/"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"04_another-way"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"_func-expression/main.go"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"` |\n| "},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu8AvZGmiqsr5cTYfc8X4l2m2dcq","object":"chat.completion.chunk","created":1776257978,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/"},"finish_reason":null}],"usage":null,"obfuscation":"dq9hfX0ZsUb"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"5 | `j"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu8AvZGmiqsr5cTYfc8X4l2m2dcq","object":"chat.completion.chunk","created":1776257978,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"tmp"},"finish_reason":null}],"usage":null,"obfuscation":"TofmZ0wqE"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"mbharathram"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu8AvZGmiqsr5cTYfc8X4l2m2dcq","object":"chat.completion.chunk","created":1776257978,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/cr"},"finish_reason":null}],"usage":null,"obfuscation":"fZAphUdyf"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"/executeoncommand"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu8AvZGmiqsr5cTYfc8X4l2m2dcq","object":"chat.completion.chunk","created":1776257978,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"ush"},"finish_reason":null}],"usage":null,"obfuscation":"J5jjjKPHt"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"` | `go"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu8AvZGmiqsr5cTYfc8X4l2m2dcq","object":"chat.completion.chunk","created":1776257978,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"-test"},"finish_reason":null}],"usage":null,"obfuscation":"fwbT6JE"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"/channel/main.go`"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu8AvZGmiqsr5cTYfc8X4l2m2dcq","object":"chat.completion.chunk","created":1776257978,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/Test"},"finish_reason":null}],"usage":null,"obfuscation":"73dOXvi"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" |\n\nMost results are"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu8AvZGmiqsr5cTYfc8X4l2m2dcq","object":"chat.completion.chunk","created":1776257978,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"Coder"},"finish_reason":null}],"usage":null,"obfuscation":"KPUdHWb"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" from Go training/tutorial"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu8AvZGmiqsr5cTYfc8X4l2m2dcq","object":"chat.completion.chunk","created":1776257978,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"Agent"},"finish_reason":null}],"usage":null,"obfuscation":"qOa4eTo"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" repos"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu8AvZGmiqsr5cTYfc8X4l2m2dcq","object":"chat.completion.chunk","created":1776257978,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/open"},"finish_reason":null}],"usage":null,"obfuscation":"59z7lDk"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" containing"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu8AvZGmiqsr5cTYfc8X4l2m2dcq","object":"chat.completion.chunk","created":1776257978,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"ai"},"finish_reason":null}],"usage":null,"obfuscation":"5LJnTggQYs"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" simple"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu8AvZGmiqsr5cTYfc8X4l2m2dcq","object":"chat.completion.chunk","created":1776257978,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"-g"},"finish_reason":null}],"usage":null,"obfuscation":"sLz9Q67VKF"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" `func main()`"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu8AvZGmiqsr5cTYfc8X4l2m2dcq","object":"chat.completion.chunk","created":1776257978,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"pt"},"finish_reason":null}],"usage":null,"obfuscation":"NEnkRAVSPh"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" entry points —"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu8AvZGmiqsr5cTYfc8X4l2m2dcq","object":"chat.completion.chunk","created":1776257978,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"-"},"finish_reason":null}],"usage":null,"obfuscation":"ofNU73tKqAH"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" function expressions, gor"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu8AvZGmiqsr5cTYfc8X4l2m2dcq","object":"chat.completion.chunk","created":1776257978,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"5"},"finish_reason":null}],"usage":null,"obfuscation":"ZB0q4EglFxn"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"outines, and"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu8AvZGmiqsr5cTYfc8X4l2m2dcq","object":"chat.completion.chunk","created":1776257978,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/"},"finish_reason":null}],"usage":null,"obfuscation":"k0wmXXdtRkt"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" channel examples. The"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu8AvZGmiqsr5cTYfc8X4l2m2dcq","object":"chat.completion.chunk","created":1776257978,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"grep"},"finish_reason":null}],"usage":null,"obfuscation":"TiKooM6i"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" query"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu8AvZGmiqsr5cTYfc8X4l2m2dcq","object":"chat.completion.chunk","created":1776257978,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"_tool"},"finish_reason":null}],"usage":null,"obfuscation":"skys9IR"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" matched"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu8AvZGmiqsr5cTYfc8X4l2m2dcq","object":"chat.completion.chunk","created":1776257978,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/main"},"finish_reason":null}],"usage":null,"obfuscation":"myFtf2C"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" broadly"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu8AvZGmiqsr5cTYfc8X4l2m2dcq","object":"chat.completion.chunk","created":1776257978,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":".go"},"finish_reason":null}],"usage":null,"obfuscation":"BNDJrHvrY"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":";"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu8AvZGmiqsr5cTYfc8X4l2m2dcq","object":"chat.completion.chunk","created":1776257978,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":":"},"finish_reason":null}],"usage":null,"obfuscation":"ihaV8upg5Jr"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" you"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu8AvZGmiqsr5cTYfc8X4l2m2dcq","object":"chat.completion.chunk","created":1776257978,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"1"},"finish_reason":null}],"usage":null,"obfuscation":"usQVk6Yaczp"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" can narrow it"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu8AvZGmiqsr5cTYfc8X4l2m2dcq","object":"chat.completion.chunk","created":1776257978,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":":"},"finish_reason":null}],"usage":null,"obfuscation":"GC2B1eGmg4h"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" with additional"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu8AvZGmiqsr5cTYfc8X4l2m2dcq","object":"chat.completion.chunk","created":1776257978,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"1"},"finish_reason":null}],"usage":null,"obfuscation":"jRwX3c6slaX"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" filters like `repo"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu8AvZGmiqsr5cTYfc8X4l2m2dcq","object":"chat.completion.chunk","created":1776257978,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":" package"},"finish_reason":null}],"usage":null,"obfuscation":"rB9O"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":":` or specific"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu8AvZGmiqsr5cTYfc8X4l2m2dcq","object":"chat.completion.chunk","created":1776257978,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":" main"},"finish_reason":null}],"usage":null,"obfuscation":"k0MDyxv"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" content"},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu8AvZGmiqsr5cTYfc8X4l2m2dcq","object":"chat.completion.chunk","created":1776257978,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":null,"obfuscation":"JdTGyq"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" patterns."},"finish_reason":null}]}
- data: {"id":"chatcmpl-DUu8AvZGmiqsr5cTYfc8X4l2m2dcq","object":"chat.completion.chunk","created":1776257978,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6471,"completion_tokens":28,"total_tokens":6499,"prompt_tokens_details":{"cached_tokens":6272,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":0,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"Cc09"}
+ data: {"id":"chatcmpl-cc747e4f3ce9b98ccca2f87f","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{"prompt_tokens":2886,"completion_tokens":230,"total_tokens":10188}}
data: [DONE]
headers:
Content-Type:
- - text/event-stream; charset=utf-8
+ - text/event-stream
status: 200 OK
code: 200
- duration: 806.88925ms
+ duration: 1.138374583s
diff --git a/internal/agent/testdata/TestCoderAgent/glm-5.1/update_a_file.yaml b/internal/agent/testdata/TestCoderAgent/glm-5.1/update_a_file.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..759b3d46c64d2d0f070a60fcd898f11404c87a12
--- /dev/null
+++ b/internal/agent/testdata/TestCoderAgent/glm-5.1/update_a_file.yaml
@@ -0,0 +1,317 @@
+---
+version: 2
+interactions:
+- id: 0
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 828
+ host: ""
+ body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nupdate the main.go file by changing the print to say hello from crush\n \n\n","role":"user"}],"model":"gpt-oss-120b","max_completion_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
+ headers:
+ Accept:
+ - application/json
+ Content-Type:
+ - application/json
+ User-Agent:
+ - Charm-Crush/devel (https://charm.land/crush)
+ url: https://hyper.charm.land/v1/chat/completions
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |+
+ data: {"id":"chatcmpl-1d2779ed381e7de8462ddb1b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-1d2779ed381e7de8462ddb1b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":"Update"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-1d2779ed381e7de8462ddb1b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" main"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-1d2779ed381e7de8462ddb1b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":".go"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-1d2779ed381e7de8462ddb1b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" to"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-1d2779ed381e7de8462ddb1b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" print"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-1d2779ed381e7de8462ddb1b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" hello"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-1d2779ed381e7de8462ddb1b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" from"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-1d2779ed381e7de8462ddb1b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" crush"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-1d2779ed381e7de8462ddb1b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{"prompt_tokens":37,"completion_tokens":174,"total_tokens":387}}
+
+ data: [DONE]
+
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 2.104933458s
+- id: 1
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 30668
+ host: ""
+ body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/glm-5.1/update_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"update the main.go file by changing the print to say hello from crush\",\"role\":\"user\"}],\"model\":\"glm-5.1\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
+ headers:
+ Accept:
+ - application/json
+ Content-Type:
+ - application/json
+ User-Agent:
+ - Charm-Crush/devel (https://charm.land/crush)
+ url: https://hyper.charm.land/v1/chat/completions
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |+
+ data: {"id":"chatcmpl-62f494c9d4daac20d6160a32","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-62f494c9d4daac20d6160a32","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_LejjtJRemmR0IkblRpWeKJcU","type":"function","index":0,"function":{"name":"view","arguments":""}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-62f494c9d4daac20d6160a32","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-62f494c9d4daac20d6160a32","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"file_path\": \""}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-62f494c9d4daac20d6160a32","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/tmp"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-62f494c9d4daac20d6160a32","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/c"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-62f494c9d4daac20d6160a32","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"rush"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-62f494c9d4daac20d6160a32","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-test"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-62f494c9d4daac20d6160a32","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/Test"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-62f494c9d4daac20d6160a32","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Coder"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-62f494c9d4daac20d6160a32","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Agent"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-62f494c9d4daac20d6160a32","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/g"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-62f494c9d4daac20d6160a32","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"lm"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-62f494c9d4daac20d6160a32","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-62f494c9d4daac20d6160a32","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"5"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-62f494c9d4daac20d6160a32","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"."}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-62f494c9d4daac20d6160a32","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"1"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-62f494c9d4daac20d6160a32","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/update"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-62f494c9d4daac20d6160a32","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_a"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-62f494c9d4daac20d6160a32","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_file"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-62f494c9d4daac20d6160a32","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/main"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-62f494c9d4daac20d6160a32","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".go"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-62f494c9d4daac20d6160a32","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"}"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-62f494c9d4daac20d6160a32","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":{"prompt_tokens":161,"completion_tokens":43,"total_tokens":7116}}
+
+ data: [DONE]
+
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 3.436352834s
+- id: 2
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 31112
+ host: ""
+ body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/glm-5.1/update_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"update the main.go file by changing the print to say hello from crush\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_LejjtJRemmR0IkblRpWeKJcU\",\"function\":{\"arguments\":\"{\\\"file_path\\\": \\\"/tmp/crush-test/TestCoderAgent/glm-5.1/update_a_file/main.go\\\"}\",\"name\":\"view\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\tfmt.Println(\\\"Hello, World!\\\")\\n 7|}\\n\\n\",\"tool_call_id\":\"call_LejjtJRemmR0IkblRpWeKJcU\",\"role\":\"tool\"}],\"model\":\"glm-5.1\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
+ headers:
+ Accept:
+ - application/json
+ Content-Type:
+ - application/json
+ User-Agent:
+ - Charm-Crush/devel (https://charm.land/crush)
+ url: https://hyper.charm.land/v1/chat/completions
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_kIs4jcLNFzh6CN4tvT3OnYVq","type":"function","index":0,"function":{"name":"edit","arguments":""}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"file_path\": \""}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/tmp"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/c"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"rush"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-test"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/Test"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Coder"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Agent"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/g"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"lm"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"5"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"."}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"1"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/update"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_a"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_file"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/main"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".go"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\", \"new_string\": \""}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"fmt"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".Println"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"(\\\""}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"hello"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" from"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" crush"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\\\")"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\", \"old_string\": \""}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"fmt"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".Println"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"(\\\""}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Hello"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":","}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" World"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"!\\\")"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"}"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e2f1220c3d02e84f0704bdc1","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":{"prompt_tokens":118,"completion_tokens":55,"total_tokens":7213}}
+
+ data: [DONE]
+
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 966.519125ms
+- id: 3
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 31628
+ host: ""
+ body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/glm-5.1/update_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"update the main.go file by changing the print to say hello from crush\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_LejjtJRemmR0IkblRpWeKJcU\",\"function\":{\"arguments\":\"{\\\"file_path\\\": \\\"/tmp/crush-test/TestCoderAgent/glm-5.1/update_a_file/main.go\\\"}\",\"name\":\"view\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\tfmt.Println(\\\"Hello, World!\\\")\\n 7|}\\n\\n\",\"tool_call_id\":\"call_LejjtJRemmR0IkblRpWeKJcU\",\"role\":\"tool\"},{\"tool_calls\":[{\"id\":\"call_kIs4jcLNFzh6CN4tvT3OnYVq\",\"function\":{\"arguments\":\"{\\\"file_path\\\": \\\"/tmp/crush-test/TestCoderAgent/glm-5.1/update_a_file/main.go\\\", \\\"new_string\\\": \\\"fmt.Println(\\\\\\\"hello from crush\\\\\\\")\\\", \\\"old_string\\\": \\\"fmt.Println(\\\\\\\"Hello, World!\\\\\\\")\\\"}\",\"name\":\"edit\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\nContent replaced in file: /tmp/crush-test/TestCoderAgent/glm-5.1/update_a_file/main.go\\n\\n\",\"tool_call_id\":\"call_kIs4jcLNFzh6CN4tvT3OnYVq\",\"role\":\"tool\"}],\"model\":\"glm-5.1\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
+ headers:
+ Accept:
+ - application/json
+ Content-Type:
+ - application/json
+ User-Agent:
+ - Charm-Crush/devel (https://charm.land/crush)
+ url: https://hyper.charm.land/v1/chat/completions
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |+
+ data: {"id":"chatcmpl-e71f6fb409f160fa5736ed0b","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e71f6fb409f160fa5736ed0b","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"Done"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e71f6fb409f160fa5736ed0b","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"."},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e71f6fb409f160fa5736ed0b","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" Updated"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e71f6fb409f160fa5736ed0b","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" `"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e71f6fb409f160fa5736ed0b","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"main"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e71f6fb409f160fa5736ed0b","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":".go"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e71f6fb409f160fa5736ed0b","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":":"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e71f6fb409f160fa5736ed0b","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"6"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e71f6fb409f160fa5736ed0b","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"`"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e71f6fb409f160fa5736ed0b","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" to"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e71f6fb409f160fa5736ed0b","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" print"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e71f6fb409f160fa5736ed0b","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" `\""},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e71f6fb409f160fa5736ed0b","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"hello"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e71f6fb409f160fa5736ed0b","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" from"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e71f6fb409f160fa5736ed0b","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":" crush"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e71f6fb409f160fa5736ed0b","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"\""},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e71f6fb409f160fa5736ed0b","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"`."},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-e71f6fb409f160fa5736ed0b","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{"prompt_tokens":80,"completion_tokens":19,"total_tokens":7267}}
+
+ data: [DONE]
+
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 932.001416ms
diff --git a/internal/agent/testdata/TestCoderAgent/glm-5.1/write_tool.yaml b/internal/agent/testdata/TestCoderAgent/glm-5.1/write_tool.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..688e233b0d2e5d84e1b8ab1f19e96195748a18b0
--- /dev/null
+++ b/internal/agent/testdata/TestCoderAgent/glm-5.1/write_tool.yaml
@@ -0,0 +1,133 @@
+---
+version: 2
+interactions:
+- id: 0
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 30705
+ host: ""
+ body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/glm-5.1/write_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use write to create a new file called config.json with content '{\\\"name\\\": \\\"test\\\", \\\"version\\\": \\\"1.0.0\\\"}'\",\"role\":\"user\"}],\"model\":\"glm-5.1\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
+ headers:
+ Accept:
+ - application/json
+ Content-Type:
+ - application/json
+ User-Agent:
+ - Charm-Crush/devel (https://charm.land/crush)
+ url: https://hyper.charm.land/v1/chat/completions
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |+
+ data: {"id":"chatcmpl-847e5ee14927528c7c205dbc","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-847e5ee14927528c7c205dbc","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_88a5","type":"function","index":0,"function":{"name":"write","arguments":""}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-847e5ee14927528c7c205dbc","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\"content\": \"{\\\"name\\\": \\\"test\\\", \\\"version\\\": \\\"1.0.0\\\"}\", \"file_path\": \"/tmp/crush-test/TestCoderAgent/glm-5.1/write_tool/config.json\"}"}}]},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-847e5ee14927528c7c205dbc","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":{"prompt_tokens":16,"completion_tokens":59,"total_tokens":7147}}
+
+ data: [DONE]
+
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 2.147664208s
+- id: 1
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 868
+ host: ""
+ body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse write to create a new file called config.json with content ''{\"name\": \"test\", \"version\": \"1.0.0\"}''\n \n\n","role":"user"}],"model":"gpt-oss-120b","max_completion_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
+ headers:
+ Accept:
+ - application/json
+ Content-Type:
+ - application/json
+ User-Agent:
+ - Charm-Crush/devel (https://charm.land/crush)
+ url: https://hyper.charm.land/v1/chat/completions
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |+
+ data: {"id":"chatcmpl-03f58ca81ce6554f3cf3714b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-03f58ca81ce6554f3cf3714b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":"Create"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-03f58ca81ce6554f3cf3714b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" config"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-03f58ca81ce6554f3cf3714b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":".json"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-03f58ca81ce6554f3cf3714b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" with"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-03f58ca81ce6554f3cf3714b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" name"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-03f58ca81ce6554f3cf3714b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" and"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-03f58ca81ce6554f3cf3714b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{"content":" version"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-03f58ca81ce6554f3cf3714b","object":"chat.completion.chunk","created":0,"model":"gpt-oss-120b","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{"prompt_tokens":3,"completion_tokens":128,"total_tokens":355}}
+
+ data: [DONE]
+
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 3.764015042s
+- id: 2
+ request:
+ proto: HTTP/1.1
+ proto_major: 1
+ proto_minor: 1
+ content_length: 31147
+ host: ""
+ body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/glm-5.1/write_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use write to create a new file called config.json with content '{\\\"name\\\": \\\"test\\\", \\\"version\\\": \\\"1.0.0\\\"}'\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_88a5\",\"function\":{\"arguments\":\"{\\\"content\\\": \\\"{\\\\\\\"name\\\\\\\": \\\\\\\"test\\\\\\\", \\\\\\\"version\\\\\\\": \\\\\\\"1.0.0\\\\\\\"}\\\", \\\"file_path\\\": \\\"/tmp/crush-test/TestCoderAgent/glm-5.1/write_tool/config.json\\\"}\",\"name\":\"write\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\nFile successfully written: /tmp/crush-test/TestCoderAgent/glm-5.1/write_tool/config.json\\n\",\"tool_call_id\":\"call_88a5\",\"role\":\"tool\"}],\"model\":\"glm-5.1\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
+ headers:
+ Accept:
+ - application/json
+ Content-Type:
+ - application/json
+ User-Agent:
+ - Charm-Crush/devel (https://charm.land/crush)
+ url: https://hyper.charm.land/v1/chat/completions
+ method: POST
+ response:
+ proto: HTTP/2.0
+ proto_major: 2
+ proto_minor: 0
+ content_length: -1
+ uncompressed: true
+ body: |+
+ data: {"id":"chatcmpl-f377bf2fa672782039d88341","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-f377bf2fa672782039d88341","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"Created `config.json"},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-f377bf2fa672782039d88341","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{"content":"`."},"finish_reason":null}]}
+
+ data: {"id":"chatcmpl-f377bf2fa672782039d88341","object":"chat.completion.chunk","created":0,"model":"glm-5.1","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{"prompt_tokens":7138,"completion_tokens":6,"total_tokens":7176}}
+
+ data: [DONE]
+
+ headers:
+ Content-Type:
+ - text/event-stream
+ status: 200 OK
+ code: 200
+ duration: 2.336614709s
diff --git a/internal/agent/testdata/TestCoderAgent/openai-gpt-5/bash_tool.yaml b/internal/agent/testdata/TestCoderAgent/openai-gpt-5/bash_tool.yaml
deleted file mode 100644
index f91b29817bdfec3e80f87c97dd307ab6766211c1..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/openai-gpt-5/bash_tool.yaml
+++ /dev/null
@@ -1,170 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 836
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse bash to create a file named test.txt with content ''hello bash''. do not print its timestamp\n \n\n","role":"user"}],"model":"gpt-4o","max_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu766HHRJZGfCbzBIsqnXXbWbRg6","object":"chat.completion.chunk","created":1776257912,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"role":"assistant","content":"","refusal":null},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"btFJOm8415mI2m"}
-
- data: {"id":"chatcmpl-DUu766HHRJZGfCbzBIsqnXXbWbRg6","object":"chat.completion.chunk","created":1776257912,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":"Creating"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"HuT3Bpv5"}
-
- data: {"id":"chatcmpl-DUu766HHRJZGfCbzBIsqnXXbWbRg6","object":"chat.completion.chunk","created":1776257912,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":" a"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"b4r8VMOYrYZAuM"}
-
- data: {"id":"chatcmpl-DUu766HHRJZGfCbzBIsqnXXbWbRg6","object":"chat.completion.chunk","created":1776257912,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":" File"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"EQgGDtlho8Q"}
-
- data: {"id":"chatcmpl-DUu766HHRJZGfCbzBIsqnXXbWbRg6","object":"chat.completion.chunk","created":1776257912,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":" with"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"1aMQ6gMv4rM"}
-
- data: {"id":"chatcmpl-DUu766HHRJZGfCbzBIsqnXXbWbRg6","object":"chat.completion.chunk","created":1776257912,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":" Bash"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"FzGsXtCzL7e"}
-
- data: {"id":"chatcmpl-DUu766HHRJZGfCbzBIsqnXXbWbRg6","object":"chat.completion.chunk","created":1776257912,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":" Without"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"LtwOp6S2"}
-
- data: {"id":"chatcmpl-DUu766HHRJZGfCbzBIsqnXXbWbRg6","object":"chat.completion.chunk","created":1776257912,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":" Timestamp"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"9CGJGr"}
-
- data: {"id":"chatcmpl-DUu766HHRJZGfCbzBIsqnXXbWbRg6","object":"chat.completion.chunk","created":1776257912,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{},"logprobs":null,"finish_reason":"stop"}],"usage":null,"obfuscation":"O7d6fVtofG"}
-
- data: {"id":"chatcmpl-DUu766HHRJZGfCbzBIsqnXXbWbRg6","object":"chat.completion.chunk","created":1776257912,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[],"usage":{"prompt_tokens":161,"completion_tokens":7,"total_tokens":168,"prompt_tokens_details":{"cached_tokens":0,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":0,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"PDcwbMlWzQteaR"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 476.599417ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30703
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/bash_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use bash to create a file named test.txt with content 'hello bash'. do not print its timestamp\",\"role\":\"user\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu77HB1K1jw6OrWCgmUl6d2MUot8","object":"chat.completion.chunk","created":1776257913,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":null,"tool_calls":[{"index":0,"id":"call_5ErEJQ9boiXg4O7SoB2c4Atc","type":"function","function":{"name":"bash","arguments":""}}],"refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"Ap8Etdpwgu8e"}
-
- data: {"id":"chatcmpl-DUu77HB1K1jw6OrWCgmUl6d2MUot8","object":"chat.completion.chunk","created":1776257913,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"qG7oH2MmmLuI8gy"}
-
- data: {"id":"chatcmpl-DUu77HB1K1jw6OrWCgmUl6d2MUot8","object":"chat.completion.chunk","created":1776257913,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"command"}}]},"finish_reason":null}],"usage":null,"obfuscation":"LS9bVfzxQ0q"}
-
- data: {"id":"chatcmpl-DUu77HB1K1jw6OrWCgmUl6d2MUot8","object":"chat.completion.chunk","created":1776257913,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"OubrMABDsj0hc"}
-
- data: {"id":"chatcmpl-DUu77HB1K1jw6OrWCgmUl6d2MUot8","object":"chat.completion.chunk","created":1776257913,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"printf"}}]},"finish_reason":null}],"usage":null,"obfuscation":"vmMCM9gR04Ll"}
-
- data: {"id":"chatcmpl-DUu77HB1K1jw6OrWCgmUl6d2MUot8","object":"chat.completion.chunk","created":1776257913,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" '"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu77HB1K1jw6OrWCgmUl6d2MUot8","object":"chat.completion.chunk","created":1776257913,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"hello"}}]},"finish_reason":null}],"usage":null,"obfuscation":"okdnCwSz8KWEf"}
-
- data: {"id":"chatcmpl-DUu77HB1K1jw6OrWCgmUl6d2MUot8","object":"chat.completion.chunk","created":1776257913,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" bash"}}]},"finish_reason":null}],"usage":null,"obfuscation":"FbZthw6boKHC5"}
-
- data: {"id":"chatcmpl-DUu77HB1K1jw6OrWCgmUl6d2MUot8","object":"chat.completion.chunk","created":1776257913,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"'"}}]},"finish_reason":null}],"usage":null,"obfuscation":"k"}
-
- data: {"id":"chatcmpl-DUu77HB1K1jw6OrWCgmUl6d2MUot8","object":"chat.completion.chunk","created":1776257913,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" >"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu77HB1K1jw6OrWCgmUl6d2MUot8","object":"chat.completion.chunk","created":1776257913,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"mYdFkLllmX7hR"}
-
- data: {"id":"chatcmpl-DUu77HB1K1jw6OrWCgmUl6d2MUot8","object":"chat.completion.chunk","created":1776257913,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".txt"}}]},"finish_reason":null}],"usage":null,"obfuscation":"uClZaUL8c16OFX"}
-
- data: {"id":"chatcmpl-DUu77HB1K1jw6OrWCgmUl6d2MUot8","object":"chat.completion.chunk","created":1776257913,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\",\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"Zsa83MZO0KMhj"}
-
- data: {"id":"chatcmpl-DUu77HB1K1jw6OrWCgmUl6d2MUot8","object":"chat.completion.chunk","created":1776257913,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"description"}}]},"finish_reason":null}],"usage":null,"obfuscation":"tOX014I"}
-
- data: {"id":"chatcmpl-DUu77HB1K1jw6OrWCgmUl6d2MUot8","object":"chat.completion.chunk","created":1776257913,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"M71muS1h2YwuY"}
-
- data: {"id":"chatcmpl-DUu77HB1K1jw6OrWCgmUl6d2MUot8","object":"chat.completion.chunk","created":1776257913,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Create"}}]},"finish_reason":null}],"usage":null,"obfuscation":"WnLZDl9baGuf"}
-
- data: {"id":"chatcmpl-DUu77HB1K1jw6OrWCgmUl6d2MUot8","object":"chat.completion.chunk","created":1776257913,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"4eNakoWEQqqPm"}
-
- data: {"id":"chatcmpl-DUu77HB1K1jw6OrWCgmUl6d2MUot8","object":"chat.completion.chunk","created":1776257913,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".txt"}}]},"finish_reason":null}],"usage":null,"obfuscation":"L2J8hyK2y0pBDt"}
-
- data: {"id":"chatcmpl-DUu77HB1K1jw6OrWCgmUl6d2MUot8","object":"chat.completion.chunk","created":1776257913,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" with"}}]},"finish_reason":null}],"usage":null,"obfuscation":"tKXcd0kZZGjjg"}
-
- data: {"id":"chatcmpl-DUu77HB1K1jw6OrWCgmUl6d2MUot8","object":"chat.completion.chunk","created":1776257913,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" content"}}]},"finish_reason":null}],"usage":null,"obfuscation":"NXreWV1sMC"}
-
- data: {"id":"chatcmpl-DUu77HB1K1jw6OrWCgmUl6d2MUot8","object":"chat.completion.chunk","created":1776257913,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"}"}}]},"finish_reason":null}],"usage":null,"obfuscation":"XdiO0S2n4KjpAW0"}
-
- data: {"id":"chatcmpl-DUu77HB1K1jw6OrWCgmUl6d2MUot8","object":"chat.completion.chunk","created":1776257913,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu77HB1K1jw6OrWCgmUl6d2MUot8","object":"chat.completion.chunk","created":1776257913,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6384,"completion_tokens":870,"total_tokens":7254,"prompt_tokens_details":{"cached_tokens":3456,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":832,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"V"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 17.858426542s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31020
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/bash_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use bash to create a file named test.txt with content 'hello bash'. do not print its timestamp\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_5ErEJQ9boiXg4O7SoB2c4Atc\",\"function\":{\"arguments\":\"{\\\"command\\\":\\\"printf 'hello bash' > test.txt\\\",\\\"description\\\":\\\"Create test.txt with content\\\"}\",\"name\":\"bash\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"no output\",\"tool_call_id\":\"call_5ErEJQ9boiXg4O7SoB2c4Atc\",\"role\":\"tool\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu7OiHHOqlp9N8BolJlgAIuPyIcv","object":"chat.completion.chunk","created":1776257930,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":"","refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"2qqsAMTO6O"}
-
- data: {"id":"chatcmpl-DUu7OiHHOqlp9N8BolJlgAIuPyIcv","object":"chat.completion.chunk","created":1776257930,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"Done"},"finish_reason":null}],"usage":null,"obfuscation":"3QylD3Db"}
-
- data: {"id":"chatcmpl-DUu7OiHHOqlp9N8BolJlgAIuPyIcv","object":"chat.completion.chunk","created":1776257930,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":null,"obfuscation":"c3sLoW"}
-
- data: {"id":"chatcmpl-DUu7OiHHOqlp9N8BolJlgAIuPyIcv","object":"chat.completion.chunk","created":1776257930,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6428,"completion_tokens":4,"total_tokens":6432,"prompt_tokens_details":{"cached_tokens":6272,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":0,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"ezOhc"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 697.361375ms
diff --git a/internal/agent/testdata/TestCoderAgent/openai-gpt-5/download_tool.yaml b/internal/agent/testdata/TestCoderAgent/openai-gpt-5/download_tool.yaml
deleted file mode 100644
index 0e0895d6f7f863e0686fd989ca5eeb6fc291af5a..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/openai-gpt-5/download_tool.yaml
+++ /dev/null
@@ -1,270 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 857
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\ndownload the file from https://example-files.online-convert.com/document/txt/example.txt and save it as example.txt\n \n\n","role":"user"}],"model":"gpt-4o","max_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu7P8bArE2m7oXyibZyJZnXnPJGS","object":"chat.completion.chunk","created":1776257931,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"role":"assistant","content":"","refusal":null},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"u2Zl5nMG0dL9CP"}
-
- data: {"id":"chatcmpl-DUu7P8bArE2m7oXyibZyJZnXnPJGS","object":"chat.completion.chunk","created":1776257931,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":"Downloading"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"Ra9ua"}
-
- data: {"id":"chatcmpl-DUu7P8bArE2m7oXyibZyJZnXnPJGS","object":"chat.completion.chunk","created":1776257931,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" and"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"iobtzjlXdwSz"}
-
- data: {"id":"chatcmpl-DUu7P8bArE2m7oXyibZyJZnXnPJGS","object":"chat.completion.chunk","created":1776257931,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" Saving"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"xeaxb2qW1"}
-
- data: {"id":"chatcmpl-DUu7P8bArE2m7oXyibZyJZnXnPJGS","object":"chat.completion.chunk","created":1776257931,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" a"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"iUZ9jtz5dNraEH"}
-
- data: {"id":"chatcmpl-DUu7P8bArE2m7oXyibZyJZnXnPJGS","object":"chat.completion.chunk","created":1776257931,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" Text"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"JNLJDvRUXy2"}
-
- data: {"id":"chatcmpl-DUu7P8bArE2m7oXyibZyJZnXnPJGS","object":"chat.completion.chunk","created":1776257931,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" File"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"g7qIehCB3JO"}
-
- data: {"id":"chatcmpl-DUu7P8bArE2m7oXyibZyJZnXnPJGS","object":"chat.completion.chunk","created":1776257931,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" from"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"aZ7zn892ySL"}
-
- data: {"id":"chatcmpl-DUu7P8bArE2m7oXyibZyJZnXnPJGS","object":"chat.completion.chunk","created":1776257931,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" URL"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"154tqRwBY6vR"}
-
- data: {"id":"chatcmpl-DUu7P8bArE2m7oXyibZyJZnXnPJGS","object":"chat.completion.chunk","created":1776257931,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{},"logprobs":null,"finish_reason":"stop"}],"usage":null,"obfuscation":"C35gRaSm0j"}
-
- data: {"id":"chatcmpl-DUu7P8bArE2m7oXyibZyJZnXnPJGS","object":"chat.completion.chunk","created":1776257931,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[],"usage":{"prompt_tokens":164,"completion_tokens":8,"total_tokens":172,"prompt_tokens_details":{"cached_tokens":0,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":0,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"MtWLF0P0cBrYuV"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 383.557166ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30728
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/download_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"download the file from https://example-files.online-convert.com/document/txt/example.txt and save it as example.txt\",\"role\":\"user\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":null,"tool_calls":[{"index":0,"id":"call_LEGkO6tOwIDrt6IOmfhkJ1Yd","type":"function","function":{"name":"download","arguments":""}}],"refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"U4pnOwEu"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"7mA1IHeLkSnHJOk"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"url"}}]},"finish_reason":null}],"usage":null,"obfuscation":"xKf8fzHO55qVDjT"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"R359hWOA8RhhW"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"https"}}]},"finish_reason":null}],"usage":null,"obfuscation":"q84yHHaEXfql2"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"://"}}]},"finish_reason":null}],"usage":null,"obfuscation":"qnOViUI2dI9U1vM"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"example"}}]},"finish_reason":null}],"usage":null,"obfuscation":"wsj0vyBAIfy"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-files"}}]},"finish_reason":null}],"usage":null,"obfuscation":"hS5XQVFJSdVZ"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".online"}}]},"finish_reason":null}],"usage":null,"obfuscation":"qnZGfu9lzdt"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-con"}}]},"finish_reason":null}],"usage":null,"obfuscation":"OGvW7q2RT8waiN"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"vert"}}]},"finish_reason":null}],"usage":null,"obfuscation":"DoYuivHZ5ILcHF"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".com"}}]},"finish_reason":null}],"usage":null,"obfuscation":"ds99c1XewVFqKK"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/document"}}]},"finish_reason":null}],"usage":null,"obfuscation":"kVBiqNvy4"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/t"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"xt"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/example"}}]},"finish_reason":null}],"usage":null,"obfuscation":"TKe79z9Nl3"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".txt"}}]},"finish_reason":null}],"usage":null,"obfuscation":"ss1ZdFPMWT02zD"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\",\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"gNgncQrBkiLJ7"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"file"}}]},"finish_reason":null}],"usage":null,"obfuscation":"fiaYb2t1flZiEQ"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_path"}}]},"finish_reason":null}],"usage":null,"obfuscation":"XpUIULSRPg5UP"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"0xbp0rKOHvcsPdv"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"/"}}]},"finish_reason":null}],"usage":null,"obfuscation":"6HJbREVgDaXay6G"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"tmp"}}]},"finish_reason":null}],"usage":null,"obfuscation":"ovuEEZcsVKSMxCF"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/cr"}}]},"finish_reason":null}],"usage":null,"obfuscation":"3WXULr2UiVMvCdt"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ush"}}]},"finish_reason":null}],"usage":null,"obfuscation":"zSxEveXmDch2uH0"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"LHH9P3BWJOWfj"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/Test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"SXD5a7ewDdszZ"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Coder"}}]},"finish_reason":null}],"usage":null,"obfuscation":"SKyuvdRmBZQ4m"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Agent"}}]},"finish_reason":null}],"usage":null,"obfuscation":"l8aX2HsQjEckm"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/open"}}]},"finish_reason":null}],"usage":null,"obfuscation":"v1d5pDG0gHJJ8"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ai"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-g"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"pt"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-"}}]},"finish_reason":null}],"usage":null,"obfuscation":"0"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"5"}}]},"finish_reason":null}],"usage":null,"obfuscation":"w"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/download"}}]},"finish_reason":null}],"usage":null,"obfuscation":"ra2JzKlCM"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_tool"}}]},"finish_reason":null}],"usage":null,"obfuscation":"PvJfql2C1nsUz"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/example"}}]},"finish_reason":null}],"usage":null,"obfuscation":"MW498P0rBE"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".txt"}}]},"finish_reason":null}],"usage":null,"obfuscation":"wm7sfTL2qXaFmm"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\",\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"08pOUxtexwmg8"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"timeout"}}]},"finish_reason":null}],"usage":null,"obfuscation":"TG0m17b7b5p"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"yLsBA2E1pD0OjX2"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"120"}}]},"finish_reason":null}],"usage":null,"obfuscation":"9iPsohiyJ2syf5e"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"}"}}]},"finish_reason":null}],"usage":null,"obfuscation":"y"}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu7PES1QeEgCFYMPKfTqdOBBkUjl","object":"chat.completion.chunk","created":1776257931,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6387,"completion_tokens":636,"total_tokens":7023,"prompt_tokens_details":{"cached_tokens":3456,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":576,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"l"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 13.55090525s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 0
- host: example-files.online-convert.com
- headers:
- User-Agent:
- - crush/1.0
- url: https://example-files.online-convert.com/document/txt/example.txt
- method: GET
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |-
- TXT test file
- Purpose: Provide example of this file type
- Document file type: TXT
- Version: 1.0
- Remark:
-
- Example content:
- The names "John Doe" for males, "Jane Doe" or "Jane Roe" for females, or "Jonnie Doe" and "Janie Doe" for children, or just "Doe" non-gender-specifically are used as placeholder names for a party whose true identity is unknown or must be withheld in a legal action, case, or discussion. The names are also used to refer to acorpse or hospital patient whose identity is unknown. This practice is widely used in the United States and Canada, but is rarely used in other English-speaking countries including the United Kingdom itself, from where the use of "John Doe" in a legal context originates. The names Joe Bloggs or John Smith are used in the UK instead, as well as in Australia and New Zealand.
-
- John Doe is sometimes used to refer to a typical male in other contexts as well, in a similar manner to John Q. Public, known in Great Britain as Joe Public, John Smith or Joe Bloggs. For example, the first name listed on a form is often John Doe, along with a fictional address or other fictional information to provide an example of how to fill in the form. The name is also used frequently in popular culture, for example in the Frank Capra film Meet John Doe. John Doe was also the name of a 2002 American television series.
-
- Similarly, a child or baby whose identity is unknown may be referred to as Baby Doe. A notorious murder case in Kansas City, Missouri, referred to the baby victim as Precious Doe. Other unidentified female murder victims are Cali Doe and Princess Doe. Additional persons may be called James Doe, Judy Doe, etc. However, to avoid possible confusion, if two anonymous or unknown parties are cited in a specific case or action, the surnames Doe and Roe may be used simultaneously; for example, "John Doe v. Jane Roe". If several anonymous parties are referenced, they may simply be labelled John Doe #1, John Doe #2, etc. (the U.S. Operation Delego cited 21 (numbered) "John Doe"s) or labelled with other variants of Doe / Roe / Poe / etc. Other early alternatives such as John Stiles and Richard Miles are now rarely used, and Mary Major has been used in some American federal cases.
-
-
-
- File created by https://www.online-convert.com
- More example files: https://www.online-convert.com/file-type
- Text of Example content: Wikipedia (https://en.wikipedia.org/wiki/John_Doe)
- License: Attribution-ShareAlike 4.0 (https://creativecommons.org/licenses/by-sa/4.0/)
-
- Feel free to use and share the file according to the license above.
- headers:
- Content-Type:
- - text/plain; charset=UTF-8
- status: 200 OK
- code: 200
- duration: 728.2115ms
-- id: 3
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31217
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/download_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"download the file from https://example-files.online-convert.com/document/txt/example.txt and save it as example.txt\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_LEGkO6tOwIDrt6IOmfhkJ1Yd\",\"function\":{\"arguments\":\"{\\\"url\\\":\\\"https://example-files.online-convert.com/document/txt/example.txt\\\",\\\"file_path\\\":\\\"/tmp/crush-test/TestCoderAgent/openai-gpt-5/download_tool/example.txt\\\",\\\"timeout\\\":120}\",\"name\":\"download\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"Successfully downloaded 2574 bytes to example.txt (Content-Type: text/plain; charset=UTF-8)\",\"tool_call_id\":\"call_LEGkO6tOwIDrt6IOmfhkJ1Yd\",\"role\":\"tool\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu7eqTx6JlSEZGmbx0lv0d2Btout","object":"chat.completion.chunk","created":1776257946,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":"","refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"1AwiqbzpV2"}
-
- data: {"id":"chatcmpl-DUu7eqTx6JlSEZGmbx0lv0d2Btout","object":"chat.completion.chunk","created":1776257946,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"Saved"},"finish_reason":null}],"usage":null,"obfuscation":"yn2VibI"}
-
- data: {"id":"chatcmpl-DUu7eqTx6JlSEZGmbx0lv0d2Btout","object":"chat.completion.chunk","created":1776257946,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":" to"},"finish_reason":null}],"usage":null,"obfuscation":"90aN8YbLN"}
-
- data: {"id":"chatcmpl-DUu7eqTx6JlSEZGmbx0lv0d2Btout","object":"chat.completion.chunk","created":1776257946,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":" example"},"finish_reason":null}],"usage":null,"obfuscation":"UwFx"}
-
- data: {"id":"chatcmpl-DUu7eqTx6JlSEZGmbx0lv0d2Btout","object":"chat.completion.chunk","created":1776257946,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":".txt"},"finish_reason":null}],"usage":null,"obfuscation":"7yTPdAnl"}
-
- data: {"id":"chatcmpl-DUu7eqTx6JlSEZGmbx0lv0d2Btout","object":"chat.completion.chunk","created":1776257946,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":null,"obfuscation":"DqfgKZ"}
-
- data: {"id":"chatcmpl-DUu7eqTx6JlSEZGmbx0lv0d2Btout","object":"chat.completion.chunk","created":1776257946,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6472,"completion_tokens":205,"total_tokens":6677,"prompt_tokens_details":{"cached_tokens":6272,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":192,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"1"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 5.580927958s
diff --git a/internal/agent/testdata/TestCoderAgent/openai-gpt-5/glob_tool.yaml b/internal/agent/testdata/TestCoderAgent/openai-gpt-5/glob_tool.yaml
deleted file mode 100644
index 19c0bef7ed2d4b09309f4b0afc0bca6b8cb0a1bc..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/openai-gpt-5/glob_tool.yaml
+++ /dev/null
@@ -1,222 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 797
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse glob to find all .go files in the current directory\n \n\n","role":"user"}],"model":"gpt-4o","max_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu7vwRcZeq1bGZKjOFL4UnjWIszV","object":"chat.completion.chunk","created":1776257963,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"role":"assistant","content":"","refusal":null},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"a2x71ppZ5Z1qs5"}
-
- data: {"id":"chatcmpl-DUu7vwRcZeq1bGZKjOFL4UnjWIszV","object":"chat.completion.chunk","created":1776257963,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":"Finding"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"Tn2CYUEym"}
-
- data: {"id":"chatcmpl-DUu7vwRcZeq1bGZKjOFL4UnjWIszV","object":"chat.completion.chunk","created":1776257963,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" ."},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"xVsGTBkumRMEVe"}
-
- data: {"id":"chatcmpl-DUu7vwRcZeq1bGZKjOFL4UnjWIszV","object":"chat.completion.chunk","created":1776257963,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":"go"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"QFn1JT6GE1J1tl"}
-
- data: {"id":"chatcmpl-DUu7vwRcZeq1bGZKjOFL4UnjWIszV","object":"chat.completion.chunk","created":1776257963,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" Files"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"RcPpPAinfY"}
-
- data: {"id":"chatcmpl-DUu7vwRcZeq1bGZKjOFL4UnjWIszV","object":"chat.completion.chunk","created":1776257963,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" with"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"qOdC5qqw3lD"}
-
- data: {"id":"chatcmpl-DUu7vwRcZeq1bGZKjOFL4UnjWIszV","object":"chat.completion.chunk","created":1776257963,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" Glob"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"o7pZGxVzN4M"}
-
- data: {"id":"chatcmpl-DUu7vwRcZeq1bGZKjOFL4UnjWIszV","object":"chat.completion.chunk","created":1776257963,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" in"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"60gBO8RrsjiQq"}
-
- data: {"id":"chatcmpl-DUu7vwRcZeq1bGZKjOFL4UnjWIszV","object":"chat.completion.chunk","created":1776257963,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" Current"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"M2evSfdW"}
-
- data: {"id":"chatcmpl-DUu7vwRcZeq1bGZKjOFL4UnjWIszV","object":"chat.completion.chunk","created":1776257963,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" Directory"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"RCtmze"}
-
- data: {"id":"chatcmpl-DUu7vwRcZeq1bGZKjOFL4UnjWIszV","object":"chat.completion.chunk","created":1776257963,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{},"logprobs":null,"finish_reason":"stop"}],"usage":null,"obfuscation":"St0cm2OqCX"}
-
- data: {"id":"chatcmpl-DUu7vwRcZeq1bGZKjOFL4UnjWIszV","object":"chat.completion.chunk","created":1776257963,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[],"usage":{"prompt_tokens":153,"completion_tokens":9,"total_tokens":162,"prompt_tokens_details":{"cached_tokens":0,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":0,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"i4S0SUY7py1rLP"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 462.902458ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30664
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/glob_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use glob to find all .go files in the current directory\",\"role\":\"user\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu7v78WXqADOJrLY8zfcGcgGiRej","object":"chat.completion.chunk","created":1776257963,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":null,"tool_calls":[{"index":0,"id":"call_GzUdF2vbdg8WBcIjXYCaFZ6Y","type":"function","function":{"name":"glob","arguments":""}}],"refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"73Z8R257cEJM"}
-
- data: {"id":"chatcmpl-DUu7v78WXqADOJrLY8zfcGcgGiRej","object":"chat.completion.chunk","created":1776257963,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"KUdpmJaVrU3o1Uw"}
-
- data: {"id":"chatcmpl-DUu7v78WXqADOJrLY8zfcGcgGiRej","object":"chat.completion.chunk","created":1776257963,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"path"}}]},"finish_reason":null}],"usage":null,"obfuscation":"kQDS4xN2EDzDFc"}
-
- data: {"id":"chatcmpl-DUu7v78WXqADOJrLY8zfcGcgGiRej","object":"chat.completion.chunk","created":1776257963,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"EAPVKfiyB9yhI2z"}
-
- data: {"id":"chatcmpl-DUu7v78WXqADOJrLY8zfcGcgGiRej","object":"chat.completion.chunk","created":1776257963,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"/"}}]},"finish_reason":null}],"usage":null,"obfuscation":"BKfqmu69NrMdjFt"}
-
- data: {"id":"chatcmpl-DUu7v78WXqADOJrLY8zfcGcgGiRej","object":"chat.completion.chunk","created":1776257963,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"tmp"}}]},"finish_reason":null}],"usage":null,"obfuscation":"ePtvkDa1g0Pjw95"}
-
- data: {"id":"chatcmpl-DUu7v78WXqADOJrLY8zfcGcgGiRej","object":"chat.completion.chunk","created":1776257963,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/cr"}}]},"finish_reason":null}],"usage":null,"obfuscation":"WP3YsFWOVzYJByU"}
-
- data: {"id":"chatcmpl-DUu7v78WXqADOJrLY8zfcGcgGiRej","object":"chat.completion.chunk","created":1776257963,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ush"}}]},"finish_reason":null}],"usage":null,"obfuscation":"ZFZq0xHpV0ILASs"}
-
- data: {"id":"chatcmpl-DUu7v78WXqADOJrLY8zfcGcgGiRej","object":"chat.completion.chunk","created":1776257963,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"esPRDuvzD8hVS"}
-
- data: {"id":"chatcmpl-DUu7v78WXqADOJrLY8zfcGcgGiRej","object":"chat.completion.chunk","created":1776257963,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/Test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"GfJH3BKR2RyMX"}
-
- data: {"id":"chatcmpl-DUu7v78WXqADOJrLY8zfcGcgGiRej","object":"chat.completion.chunk","created":1776257963,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Coder"}}]},"finish_reason":null}],"usage":null,"obfuscation":"NroS24sY98I5M"}
-
- data: {"id":"chatcmpl-DUu7v78WXqADOJrLY8zfcGcgGiRej","object":"chat.completion.chunk","created":1776257963,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Agent"}}]},"finish_reason":null}],"usage":null,"obfuscation":"6aFw6eX7DNvlf"}
-
- data: {"id":"chatcmpl-DUu7v78WXqADOJrLY8zfcGcgGiRej","object":"chat.completion.chunk","created":1776257963,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/open"}}]},"finish_reason":null}],"usage":null,"obfuscation":"uJ2Y4BfsAEB6U"}
-
- data: {"id":"chatcmpl-DUu7v78WXqADOJrLY8zfcGcgGiRej","object":"chat.completion.chunk","created":1776257963,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ai"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu7v78WXqADOJrLY8zfcGcgGiRej","object":"chat.completion.chunk","created":1776257963,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-g"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu7v78WXqADOJrLY8zfcGcgGiRej","object":"chat.completion.chunk","created":1776257963,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"pt"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu7v78WXqADOJrLY8zfcGcgGiRej","object":"chat.completion.chunk","created":1776257963,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-"}}]},"finish_reason":null}],"usage":null,"obfuscation":"n"}
-
- data: {"id":"chatcmpl-DUu7v78WXqADOJrLY8zfcGcgGiRej","object":"chat.completion.chunk","created":1776257963,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"5"}}]},"finish_reason":null}],"usage":null,"obfuscation":"w"}
-
- data: {"id":"chatcmpl-DUu7v78WXqADOJrLY8zfcGcgGiRej","object":"chat.completion.chunk","created":1776257963,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/gl"}}]},"finish_reason":null}],"usage":null,"obfuscation":"ApsbCk5RZ0mSDhr"}
-
- data: {"id":"chatcmpl-DUu7v78WXqADOJrLY8zfcGcgGiRej","object":"chat.completion.chunk","created":1776257963,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ob"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu7v78WXqADOJrLY8zfcGcgGiRej","object":"chat.completion.chunk","created":1776257963,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_tool"}}]},"finish_reason":null}],"usage":null,"obfuscation":"2Icq8PbY7MItz"}
-
- data: {"id":"chatcmpl-DUu7v78WXqADOJrLY8zfcGcgGiRej","object":"chat.completion.chunk","created":1776257963,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\",\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"WT4vNosuN4WCg"}
-
- data: {"id":"chatcmpl-DUu7v78WXqADOJrLY8zfcGcgGiRej","object":"chat.completion.chunk","created":1776257963,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"pattern"}}]},"finish_reason":null}],"usage":null,"obfuscation":"Bn864sNk6X2"}
-
- data: {"id":"chatcmpl-DUu7v78WXqADOJrLY8zfcGcgGiRej","object":"chat.completion.chunk","created":1776257963,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"uxD8E77SF1BrX"}
-
- data: {"id":"chatcmpl-DUu7v78WXqADOJrLY8zfcGcgGiRej","object":"chat.completion.chunk","created":1776257963,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"*."}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu7v78WXqADOJrLY8zfcGcgGiRej","object":"chat.completion.chunk","created":1776257963,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"go"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu7v78WXqADOJrLY8zfcGcgGiRej","object":"chat.completion.chunk","created":1776257963,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"}"}}]},"finish_reason":null}],"usage":null,"obfuscation":"suKU6jui17H0Bcl"}
-
- data: {"id":"chatcmpl-DUu7v78WXqADOJrLY8zfcGcgGiRej","object":"chat.completion.chunk","created":1776257963,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu7v78WXqADOJrLY8zfcGcgGiRej","object":"chat.completion.chunk","created":1776257963,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6377,"completion_tokens":363,"total_tokens":6740,"prompt_tokens_details":{"cached_tokens":3456,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":320,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"J"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 7.630558916s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31025
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/glob_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use glob to find all .go files in the current directory\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_GzUdF2vbdg8WBcIjXYCaFZ6Y\",\"function\":{\"arguments\":\"{\\\"path\\\":\\\"/tmp/crush-test/TestCoderAgent/openai-gpt-5/glob_tool\\\",\\\"pattern\\\":\\\"*.go\\\"}\",\"name\":\"glob\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"/tmp/crush-test/TestCoderAgent/openai-gpt-5/glob_tool/main.go\",\"tool_call_id\":\"call_GzUdF2vbdg8WBcIjXYCaFZ6Y\",\"role\":\"tool\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu83kkwh0pIcbfWtSN0p3xUEZsut","object":"chat.completion.chunk","created":1776257971,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":"","refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"ggyjPsMdgW"}
-
- data: {"id":"chatcmpl-DUu83kkwh0pIcbfWtSN0p3xUEZsut","object":"chat.completion.chunk","created":1776257971,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/"},"finish_reason":null}],"usage":null,"obfuscation":"uQzHFDtcIJy"}
-
- data: {"id":"chatcmpl-DUu83kkwh0pIcbfWtSN0p3xUEZsut","object":"chat.completion.chunk","created":1776257971,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"tmp"},"finish_reason":null}],"usage":null,"obfuscation":"69dn5IKQd"}
-
- data: {"id":"chatcmpl-DUu83kkwh0pIcbfWtSN0p3xUEZsut","object":"chat.completion.chunk","created":1776257971,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/cr"},"finish_reason":null}],"usage":null,"obfuscation":"9R5HmVjCi"}
-
- data: {"id":"chatcmpl-DUu83kkwh0pIcbfWtSN0p3xUEZsut","object":"chat.completion.chunk","created":1776257971,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"ush"},"finish_reason":null}],"usage":null,"obfuscation":"z7Ajo465K"}
-
- data: {"id":"chatcmpl-DUu83kkwh0pIcbfWtSN0p3xUEZsut","object":"chat.completion.chunk","created":1776257971,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"-test"},"finish_reason":null}],"usage":null,"obfuscation":"HXUEYZX"}
-
- data: {"id":"chatcmpl-DUu83kkwh0pIcbfWtSN0p3xUEZsut","object":"chat.completion.chunk","created":1776257971,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/Test"},"finish_reason":null}],"usage":null,"obfuscation":"0Ki7wyB"}
-
- data: {"id":"chatcmpl-DUu83kkwh0pIcbfWtSN0p3xUEZsut","object":"chat.completion.chunk","created":1776257971,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"Coder"},"finish_reason":null}],"usage":null,"obfuscation":"2YB3Ijz"}
-
- data: {"id":"chatcmpl-DUu83kkwh0pIcbfWtSN0p3xUEZsut","object":"chat.completion.chunk","created":1776257971,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"Agent"},"finish_reason":null}],"usage":null,"obfuscation":"FJ2xhZB"}
-
- data: {"id":"chatcmpl-DUu83kkwh0pIcbfWtSN0p3xUEZsut","object":"chat.completion.chunk","created":1776257971,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/open"},"finish_reason":null}],"usage":null,"obfuscation":"uOA1wNX"}
-
- data: {"id":"chatcmpl-DUu83kkwh0pIcbfWtSN0p3xUEZsut","object":"chat.completion.chunk","created":1776257971,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"ai"},"finish_reason":null}],"usage":null,"obfuscation":"qo1JDaxIM1"}
-
- data: {"id":"chatcmpl-DUu83kkwh0pIcbfWtSN0p3xUEZsut","object":"chat.completion.chunk","created":1776257971,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"-g"},"finish_reason":null}],"usage":null,"obfuscation":"nPzveAE4Qy"}
-
- data: {"id":"chatcmpl-DUu83kkwh0pIcbfWtSN0p3xUEZsut","object":"chat.completion.chunk","created":1776257971,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"pt"},"finish_reason":null}],"usage":null,"obfuscation":"j6piAxvx2Y"}
-
- data: {"id":"chatcmpl-DUu83kkwh0pIcbfWtSN0p3xUEZsut","object":"chat.completion.chunk","created":1776257971,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"-"},"finish_reason":null}],"usage":null,"obfuscation":"qGQVUtFrJuW"}
-
- data: {"id":"chatcmpl-DUu83kkwh0pIcbfWtSN0p3xUEZsut","object":"chat.completion.chunk","created":1776257971,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"5"},"finish_reason":null}],"usage":null,"obfuscation":"pSVqh287L0p"}
-
- data: {"id":"chatcmpl-DUu83kkwh0pIcbfWtSN0p3xUEZsut","object":"chat.completion.chunk","created":1776257971,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/gl"},"finish_reason":null}],"usage":null,"obfuscation":"Ag29cmEVg"}
-
- data: {"id":"chatcmpl-DUu83kkwh0pIcbfWtSN0p3xUEZsut","object":"chat.completion.chunk","created":1776257971,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"ob"},"finish_reason":null}],"usage":null,"obfuscation":"a1rs0UQo8x"}
-
- data: {"id":"chatcmpl-DUu83kkwh0pIcbfWtSN0p3xUEZsut","object":"chat.completion.chunk","created":1776257971,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"_tool"},"finish_reason":null}],"usage":null,"obfuscation":"BOtLaMC"}
-
- data: {"id":"chatcmpl-DUu83kkwh0pIcbfWtSN0p3xUEZsut","object":"chat.completion.chunk","created":1776257971,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/main"},"finish_reason":null}],"usage":null,"obfuscation":"zmuz7oU"}
-
- data: {"id":"chatcmpl-DUu83kkwh0pIcbfWtSN0p3xUEZsut","object":"chat.completion.chunk","created":1776257971,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":".go"},"finish_reason":null}],"usage":null,"obfuscation":"m7VfkSuSj"}
-
- data: {"id":"chatcmpl-DUu83kkwh0pIcbfWtSN0p3xUEZsut","object":"chat.completion.chunk","created":1776257971,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":null,"obfuscation":"ZXCc4F"}
-
- data: {"id":"chatcmpl-DUu83kkwh0pIcbfWtSN0p3xUEZsut","object":"chat.completion.chunk","created":1776257971,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6442,"completion_tokens":22,"total_tokens":6464,"prompt_tokens_details":{"cached_tokens":6272,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":0,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"JbOo"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 662.510417ms
diff --git a/internal/agent/testdata/TestCoderAgent/openai-gpt-5/ls_tool.yaml b/internal/agent/testdata/TestCoderAgent/openai-gpt-5/ls_tool.yaml
deleted file mode 100644
index cded06cbec3b06bb750b692b9a3cb240b1424f1f..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/openai-gpt-5/ls_tool.yaml
+++ /dev/null
@@ -1,188 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 791
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse ls to list the files in the current directory\n \n\n","role":"user"}],"model":"gpt-4o","max_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu8By9ZhTZA2ibLQodzTs4CJg9HK","object":"chat.completion.chunk","created":1776257979,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"role":"assistant","content":"","refusal":null},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"KdWvH71jPQncsC"}
-
- data: {"id":"chatcmpl-DUu8By9ZhTZA2ibLQodzTs4CJg9HK","object":"chat.completion.chunk","created":1776257979,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":"Listing"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"QTgvyWYJF"}
-
- data: {"id":"chatcmpl-DUu8By9ZhTZA2ibLQodzTs4CJg9HK","object":"chat.completion.chunk","created":1776257979,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" Files"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"95S2InrgnJ"}
-
- data: {"id":"chatcmpl-DUu8By9ZhTZA2ibLQodzTs4CJg9HK","object":"chat.completion.chunk","created":1776257979,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" with"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"BXlMfiG1WWO"}
-
- data: {"id":"chatcmpl-DUu8By9ZhTZA2ibLQodzTs4CJg9HK","object":"chat.completion.chunk","created":1776257979,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" LS"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"6Rx07ZTyhSMbA"}
-
- data: {"id":"chatcmpl-DUu8By9ZhTZA2ibLQodzTs4CJg9HK","object":"chat.completion.chunk","created":1776257979,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" Command"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"SPSL3vJ4"}
-
- data: {"id":"chatcmpl-DUu8By9ZhTZA2ibLQodzTs4CJg9HK","object":"chat.completion.chunk","created":1776257979,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{},"logprobs":null,"finish_reason":"stop"}],"usage":null,"obfuscation":"09O2pUvzlg"}
-
- data: {"id":"chatcmpl-DUu8By9ZhTZA2ibLQodzTs4CJg9HK","object":"chat.completion.chunk","created":1776257979,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[],"usage":{"prompt_tokens":151,"completion_tokens":5,"total_tokens":156,"prompt_tokens_details":{"cached_tokens":0,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":0,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"AWangyIfGvW1fn"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 495.352166ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30656
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/ls_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use ls to list the files in the current directory\",\"role\":\"user\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu8BRYZmGiN6COk7XzDNUZD9gb7G","object":"chat.completion.chunk","created":1776257979,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":null,"tool_calls":[{"index":0,"id":"call_TVPf6rox6w0Zw0bkccPWVD9I","type":"function","function":{"name":"ls","arguments":""}}],"refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"5xjLrtSXvQOPds"}
-
- data: {"id":"chatcmpl-DUu8BRYZmGiN6COk7XzDNUZD9gb7G","object":"chat.completion.chunk","created":1776257979,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"fVi6aSqkWKUZwMt"}
-
- data: {"id":"chatcmpl-DUu8BRYZmGiN6COk7XzDNUZD9gb7G","object":"chat.completion.chunk","created":1776257979,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"path"}}]},"finish_reason":null}],"usage":null,"obfuscation":"JVXYVaZnebroKI"}
-
- data: {"id":"chatcmpl-DUu8BRYZmGiN6COk7XzDNUZD9gb7G","object":"chat.completion.chunk","created":1776257979,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"244pNCgF6auLquo"}
-
- data: {"id":"chatcmpl-DUu8BRYZmGiN6COk7XzDNUZD9gb7G","object":"chat.completion.chunk","created":1776257979,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" \"/"}}]},"finish_reason":null}],"usage":null,"obfuscation":"NkngRtdkLQnqRn"}
-
- data: {"id":"chatcmpl-DUu8BRYZmGiN6COk7XzDNUZD9gb7G","object":"chat.completion.chunk","created":1776257979,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"tmp"}}]},"finish_reason":null}],"usage":null,"obfuscation":"2Og8HMg9kq3V0sW"}
-
- data: {"id":"chatcmpl-DUu8BRYZmGiN6COk7XzDNUZD9gb7G","object":"chat.completion.chunk","created":1776257979,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/cr"}}]},"finish_reason":null}],"usage":null,"obfuscation":"fmYKZrgRy2S5hKJ"}
-
- data: {"id":"chatcmpl-DUu8BRYZmGiN6COk7XzDNUZD9gb7G","object":"chat.completion.chunk","created":1776257979,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ush"}}]},"finish_reason":null}],"usage":null,"obfuscation":"JErvZatIPcrxlsz"}
-
- data: {"id":"chatcmpl-DUu8BRYZmGiN6COk7XzDNUZD9gb7G","object":"chat.completion.chunk","created":1776257979,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"yOphPVNh2AwQB"}
-
- data: {"id":"chatcmpl-DUu8BRYZmGiN6COk7XzDNUZD9gb7G","object":"chat.completion.chunk","created":1776257979,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/Test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"E7Q4BTM5N0Qrm"}
-
- data: {"id":"chatcmpl-DUu8BRYZmGiN6COk7XzDNUZD9gb7G","object":"chat.completion.chunk","created":1776257979,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Coder"}}]},"finish_reason":null}],"usage":null,"obfuscation":"icfp1AiSZzwp6"}
-
- data: {"id":"chatcmpl-DUu8BRYZmGiN6COk7XzDNUZD9gb7G","object":"chat.completion.chunk","created":1776257979,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Agent"}}]},"finish_reason":null}],"usage":null,"obfuscation":"0oL4w968Hgb7I"}
-
- data: {"id":"chatcmpl-DUu8BRYZmGiN6COk7XzDNUZD9gb7G","object":"chat.completion.chunk","created":1776257979,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/open"}}]},"finish_reason":null}],"usage":null,"obfuscation":"nnUFAO8Q4JCvL"}
-
- data: {"id":"chatcmpl-DUu8BRYZmGiN6COk7XzDNUZD9gb7G","object":"chat.completion.chunk","created":1776257979,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ai"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu8BRYZmGiN6COk7XzDNUZD9gb7G","object":"chat.completion.chunk","created":1776257979,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-g"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu8BRYZmGiN6COk7XzDNUZD9gb7G","object":"chat.completion.chunk","created":1776257979,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"pt"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu8BRYZmGiN6COk7XzDNUZD9gb7G","object":"chat.completion.chunk","created":1776257979,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-"}}]},"finish_reason":null}],"usage":null,"obfuscation":"m"}
-
- data: {"id":"chatcmpl-DUu8BRYZmGiN6COk7XzDNUZD9gb7G","object":"chat.completion.chunk","created":1776257979,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"5"}}]},"finish_reason":null}],"usage":null,"obfuscation":"W"}
-
- data: {"id":"chatcmpl-DUu8BRYZmGiN6COk7XzDNUZD9gb7G","object":"chat.completion.chunk","created":1776257979,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/"}}]},"finish_reason":null}],"usage":null,"obfuscation":"1"}
-
- data: {"id":"chatcmpl-DUu8BRYZmGiN6COk7XzDNUZD9gb7G","object":"chat.completion.chunk","created":1776257979,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ls"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu8BRYZmGiN6COk7XzDNUZD9gb7G","object":"chat.completion.chunk","created":1776257979,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_tool"}}]},"finish_reason":null}],"usage":null,"obfuscation":"PxDVPjJUfT4qU"}
-
- data: {"id":"chatcmpl-DUu8BRYZmGiN6COk7XzDNUZD9gb7G","object":"chat.completion.chunk","created":1776257979,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\","}}]},"finish_reason":null}],"usage":null,"obfuscation":"nKnV1ADwPkDz2z6"}
-
- data: {"id":"chatcmpl-DUu8BRYZmGiN6COk7XzDNUZD9gb7G","object":"chat.completion.chunk","created":1776257979,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" \""}}]},"finish_reason":null}],"usage":null,"obfuscation":"uTGeyZBIjPJjD9c"}
-
- data: {"id":"chatcmpl-DUu8BRYZmGiN6COk7XzDNUZD9gb7G","object":"chat.completion.chunk","created":1776257979,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"depth"}}]},"finish_reason":null}],"usage":null,"obfuscation":"TBwmmNIksIetK"}
-
- data: {"id":"chatcmpl-DUu8BRYZmGiN6COk7XzDNUZD9gb7G","object":"chat.completion.chunk","created":1776257979,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"J5DENVvE83oVySX"}
-
- data: {"id":"chatcmpl-DUu8BRYZmGiN6COk7XzDNUZD9gb7G","object":"chat.completion.chunk","created":1776257979,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" "}}]},"finish_reason":null}],"usage":null,"obfuscation":"T"}
-
- data: {"id":"chatcmpl-DUu8BRYZmGiN6COk7XzDNUZD9gb7G","object":"chat.completion.chunk","created":1776257979,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"1"}}]},"finish_reason":null}],"usage":null,"obfuscation":"H"}
-
- data: {"id":"chatcmpl-DUu8BRYZmGiN6COk7XzDNUZD9gb7G","object":"chat.completion.chunk","created":1776257979,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"}"}}]},"finish_reason":null}],"usage":null,"obfuscation":"M"}
-
- data: {"id":"chatcmpl-DUu8BRYZmGiN6COk7XzDNUZD9gb7G","object":"chat.completion.chunk","created":1776257979,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu8BRYZmGiN6COk7XzDNUZD9gb7G","object":"chat.completion.chunk","created":1776257979,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6375,"completion_tokens":237,"total_tokens":6612,"prompt_tokens_details":{"cached_tokens":3456,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":192,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"W"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 5.000794875s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31029
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/ls_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use ls to list the files in the current directory\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_TVPf6rox6w0Zw0bkccPWVD9I\",\"function\":{\"arguments\":\"{\\\"path\\\": \\\"/tmp/crush-test/TestCoderAgent/openai-gpt-5/ls_tool\\\", \\\"depth\\\": 1}\",\"name\":\"ls\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n- /tmp/crush-test/TestCoderAgent/openai-gpt-5/ls_tool/\\n - go.mod\\n - main.go\\n\",\"tool_call_id\":\"call_TVPf6rox6w0Zw0bkccPWVD9I\",\"role\":\"tool\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu8G4sTLAnQlW0NhJPeFW5fDfOJ2","object":"chat.completion.chunk","created":1776257984,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":"","refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"fpclxhV6YS"}
-
- data: {"id":"chatcmpl-DUu8G4sTLAnQlW0NhJPeFW5fDfOJ2","object":"chat.completion.chunk","created":1776257984,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"go"},"finish_reason":null}],"usage":null,"obfuscation":"h95DxsH2g7"}
-
- data: {"id":"chatcmpl-DUu8G4sTLAnQlW0NhJPeFW5fDfOJ2","object":"chat.completion.chunk","created":1776257984,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":".mod"},"finish_reason":null}],"usage":null,"obfuscation":"alTNAjXW"}
-
- data: {"id":"chatcmpl-DUu8G4sTLAnQlW0NhJPeFW5fDfOJ2","object":"chat.completion.chunk","created":1776257984,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":","},"finish_reason":null}],"usage":null,"obfuscation":"whXA8YLdI5m"}
-
- data: {"id":"chatcmpl-DUu8G4sTLAnQlW0NhJPeFW5fDfOJ2","object":"chat.completion.chunk","created":1776257984,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":" main"},"finish_reason":null}],"usage":null,"obfuscation":"iHUMblQ"}
-
- data: {"id":"chatcmpl-DUu8G4sTLAnQlW0NhJPeFW5fDfOJ2","object":"chat.completion.chunk","created":1776257984,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":".go"},"finish_reason":null}],"usage":null,"obfuscation":"fnyduyYKl"}
-
- data: {"id":"chatcmpl-DUu8G4sTLAnQlW0NhJPeFW5fDfOJ2","object":"chat.completion.chunk","created":1776257984,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":null,"obfuscation":"Sl0UCl"}
-
- data: {"id":"chatcmpl-DUu8G4sTLAnQlW0NhJPeFW5fDfOJ2","object":"chat.completion.chunk","created":1776257984,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6452,"completion_tokens":8,"total_tokens":6460,"prompt_tokens_details":{"cached_tokens":6272,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":0,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"b2fBN"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 658.968875ms
diff --git a/internal/agent/testdata/TestCoderAgent/openai-gpt-5/multiedit_tool.yaml b/internal/agent/testdata/TestCoderAgent/openai-gpt-5/multiedit_tool.yaml
deleted file mode 100644
index 9735b27521fedf0383100281a06fdb44153e15c3..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/openai-gpt-5/multiedit_tool.yaml
+++ /dev/null
@@ -1,607 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 870
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse multiedit to change ''Hello, World!'' to ''Hello, Crush!'' and add a comment ''// Greeting'' above the fmt.Println line in main.go\n \n\n","role":"user"}],"model":"gpt-4o","max_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUuHv4cCwsfQOFgckGzR68rs71VKw","object":"chat.completion.chunk","created":1776258583,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"role":"assistant","content":"","refusal":null},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"pdXDH27RbLoHxt"}
-
- data: {"id":"chatcmpl-DUuHv4cCwsfQOFgckGzR68rs71VKw","object":"chat.completion.chunk","created":1776258583,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":"Updating"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"7ZzORgyJ"}
-
- data: {"id":"chatcmpl-DUuHv4cCwsfQOFgckGzR68rs71VKw","object":"chat.completion.chunk","created":1776258583,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":" Greeting"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"pWByjnu"}
-
- data: {"id":"chatcmpl-DUuHv4cCwsfQOFgckGzR68rs71VKw","object":"chat.completion.chunk","created":1776258583,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":" Message"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"6IOUbSDj"}
-
- data: {"id":"chatcmpl-DUuHv4cCwsfQOFgckGzR68rs71VKw","object":"chat.completion.chunk","created":1776258583,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":" in"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"qs8mFwfRiROzU"}
-
- data: {"id":"chatcmpl-DUuHv4cCwsfQOFgckGzR68rs71VKw","object":"chat.completion.chunk","created":1776258583,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":" main"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"AOKru5lV1pk"}
-
- data: {"id":"chatcmpl-DUuHv4cCwsfQOFgckGzR68rs71VKw","object":"chat.completion.chunk","created":1776258583,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":".go"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"r4L6nUPGX7SM4"}
-
- data: {"id":"chatcmpl-DUuHv4cCwsfQOFgckGzR68rs71VKw","object":"chat.completion.chunk","created":1776258583,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":" with"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"98sGxlUf9T9"}
-
- data: {"id":"chatcmpl-DUuHv4cCwsfQOFgckGzR68rs71VKw","object":"chat.completion.chunk","created":1776258583,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":" Comment"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"ILxbKeYR"}
-
- data: {"id":"chatcmpl-DUuHv4cCwsfQOFgckGzR68rs71VKw","object":"chat.completion.chunk","created":1776258583,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{},"logprobs":null,"finish_reason":"stop"}],"usage":null,"obfuscation":"HV7P1UFb9I"}
-
- data: {"id":"chatcmpl-DUuHv4cCwsfQOFgckGzR68rs71VKw","object":"chat.completion.chunk","created":1776258583,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[],"usage":{"prompt_tokens":173,"completion_tokens":8,"total_tokens":181,"prompt_tokens_details":{"cached_tokens":0,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":0,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"HVTixg6ANeOJbV"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 662.977083ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30742
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/multiedit_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use multiedit to change 'Hello, World!' to 'Hello, Crush!' and add a comment '// Greeting' above the fmt.Println line in main.go\",\"role\":\"user\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":null,"tool_calls":[{"index":0,"id":"call_1EaQeVFsEBO0jON0b6tMK6ZT","type":"function","function":{"name":"ls","arguments":""}}],"refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"CIlOocx6By1Iy8"}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"CcRKOmfRruIlL0E"}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"path"}}]},"finish_reason":null}],"usage":null,"obfuscation":"whew0yAPopZPLC"}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"tI4voixQzOLhlMg"}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" \"/"}}]},"finish_reason":null}],"usage":null,"obfuscation":"oM24Q4VGCieoql"}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"tmp"}}]},"finish_reason":null}],"usage":null,"obfuscation":"5dKwjDOXtw5yLt0"}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/cr"}}]},"finish_reason":null}],"usage":null,"obfuscation":"Vl2UwLz3YyJxIyQ"}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ush"}}]},"finish_reason":null}],"usage":null,"obfuscation":"GJ5Xc6NrHe0a0gs"}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"TgN2NCKMbPkVF"}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/Test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"cauUULTy804tI"}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Coder"}}]},"finish_reason":null}],"usage":null,"obfuscation":"AHQxOGZBDfGkL"}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Agent"}}]},"finish_reason":null}],"usage":null,"obfuscation":"mtX2QVGqKSxh8"}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/open"}}]},"finish_reason":null}],"usage":null,"obfuscation":"qWgk1B9ndaVDQ"}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ai"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-g"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"pt"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-"}}]},"finish_reason":null}],"usage":null,"obfuscation":"T"}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"5"}}]},"finish_reason":null}],"usage":null,"obfuscation":"b"}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/m"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ult"}}]},"finish_reason":null}],"usage":null,"obfuscation":"zliNoGvZfH5AxcH"}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ied"}}]},"finish_reason":null}],"usage":null,"obfuscation":"eGmVzkHK5hUn2Yd"}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"it"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_tool"}}]},"finish_reason":null}],"usage":null,"obfuscation":"AeEgPfFoydKaC"}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\","}}]},"finish_reason":null}],"usage":null,"obfuscation":"bLC8nRAQfURhe02"}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" \""}}]},"finish_reason":null}],"usage":null,"obfuscation":"ZLCbSMj7k9xptuA"}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"depth"}}]},"finish_reason":null}],"usage":null,"obfuscation":"wtpWiAmbzARpW"}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"fxA3htF8CP5nHJi"}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" "}}]},"finish_reason":null}],"usage":null,"obfuscation":"V"}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"2"}}]},"finish_reason":null}],"usage":null,"obfuscation":"0"}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"}"}}]},"finish_reason":null}],"usage":null,"obfuscation":"F"}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHv2uC5AkO1R282VUsUc7ypT7Yg","object":"chat.completion.chunk","created":1776258583,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6399,"completion_tokens":175,"total_tokens":6574,"prompt_tokens_details":{"cached_tokens":6272,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":128,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"v"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 2.264347834s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31129
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/multiedit_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use multiedit to change 'Hello, World!' to 'Hello, Crush!' and add a comment '// Greeting' above the fmt.Println line in main.go\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_1EaQeVFsEBO0jON0b6tMK6ZT\",\"function\":{\"arguments\":\"{\\\"path\\\": \\\"/tmp/crush-test/TestCoderAgent/openai-gpt-5/multiedit_tool\\\", \\\"depth\\\": 2}\",\"name\":\"ls\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n- /tmp/crush-test/TestCoderAgent/openai-gpt-5/multiedit_tool/\\n - go.mod\\n - main.go\\n\",\"tool_call_id\":\"call_1EaQeVFsEBO0jON0b6tMK6ZT\",\"role\":\"tool\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":null,"tool_calls":[{"index":0,"id":"call_Jx0qaG3KLJe5lS83V4Ke8s1s","type":"function","function":{"name":"view","arguments":""}}],"refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"60Ff6y039l5o"}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"z97PtBBjWkt5rtX"}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"file"}}]},"finish_reason":null}],"usage":null,"obfuscation":"08O0FWFSyKZpAQ"}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_path"}}]},"finish_reason":null}],"usage":null,"obfuscation":"6nPrYZKuEr6Bj"}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"1BS3pvUKwiVAKPj"}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"/"}}]},"finish_reason":null}],"usage":null,"obfuscation":"59cNESyOpulFsVL"}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"tmp"}}]},"finish_reason":null}],"usage":null,"obfuscation":"Pds1asGvjg0atPB"}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/cr"}}]},"finish_reason":null}],"usage":null,"obfuscation":"z6FoOW1VyMDnAfY"}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ush"}}]},"finish_reason":null}],"usage":null,"obfuscation":"itSRPYRYYvc6JiN"}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"TSw7m5u5B32gA"}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/Test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"eB2LO2sn2BWo1"}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Coder"}}]},"finish_reason":null}],"usage":null,"obfuscation":"WsbNX8KiAx7ci"}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Agent"}}]},"finish_reason":null}],"usage":null,"obfuscation":"KlTjucjI0lQDl"}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/open"}}]},"finish_reason":null}],"usage":null,"obfuscation":"9FURnJWPgvdvs"}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ai"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-g"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"pt"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-"}}]},"finish_reason":null}],"usage":null,"obfuscation":"Y"}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"5"}}]},"finish_reason":null}],"usage":null,"obfuscation":"3"}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/m"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ult"}}]},"finish_reason":null}],"usage":null,"obfuscation":"qtvEYIuqTlcYmOP"}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ied"}}]},"finish_reason":null}],"usage":null,"obfuscation":"7YjBSvPr1iz59xa"}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"it"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_tool"}}]},"finish_reason":null}],"usage":null,"obfuscation":"aW65NJ8untd6F"}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/main"}}]},"finish_reason":null}],"usage":null,"obfuscation":"aywNJfFZQ3eOz"}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".go"}}]},"finish_reason":null}],"usage":null,"obfuscation":"vDebugwTBXfsOOT"}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\",\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"aRKq5LF5PELqD"}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"limit"}}]},"finish_reason":null}],"usage":null,"obfuscation":"1SMkqAHP2VLW6"}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"7RFxwQVRGNzBLyb"}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"200"}}]},"finish_reason":null}],"usage":null,"obfuscation":"F5gDpwBfPQzmVuM"}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"0"}}]},"finish_reason":null}],"usage":null,"obfuscation":"1"}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"}"}}]},"finish_reason":null}],"usage":null,"obfuscation":"j"}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHx8c3G3CHHU6JoQmNLSFDrntMY","object":"chat.completion.chunk","created":1776258585,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6480,"completion_tokens":42,"total_tokens":6522,"prompt_tokens_details":{"cached_tokens":6400,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":0,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"M786"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 878.976792ms
-- id: 3
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31593
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/multiedit_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use multiedit to change 'Hello, World!' to 'Hello, Crush!' and add a comment '// Greeting' above the fmt.Println line in main.go\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_1EaQeVFsEBO0jON0b6tMK6ZT\",\"function\":{\"arguments\":\"{\\\"path\\\": \\\"/tmp/crush-test/TestCoderAgent/openai-gpt-5/multiedit_tool\\\", \\\"depth\\\": 2}\",\"name\":\"ls\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n- /tmp/crush-test/TestCoderAgent/openai-gpt-5/multiedit_tool/\\n - go.mod\\n - main.go\\n\",\"tool_call_id\":\"call_1EaQeVFsEBO0jON0b6tMK6ZT\",\"role\":\"tool\"},{\"tool_calls\":[{\"id\":\"call_Jx0qaG3KLJe5lS83V4Ke8s1s\",\"function\":{\"arguments\":\"{\\\"file_path\\\":\\\"/tmp/crush-test/TestCoderAgent/openai-gpt-5/multiedit_tool/main.go\\\",\\\"limit\\\":2000}\",\"name\":\"view\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\tfmt.Println(\\\"Hello, World!\\\")\\n 7|}\\n\\n\",\"tool_call_id\":\"call_Jx0qaG3KLJe5lS83V4Ke8s1s\",\"role\":\"tool\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":null,"tool_calls":[{"index":0,"id":"call_2hzHz91aslzg48tDaSK7ttbD","type":"function","function":{"name":"multiedit","arguments":""}}],"refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"KNlve4Q"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"IT8KXf41ih3oChi"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"file"}}]},"finish_reason":null}],"usage":null,"obfuscation":"EeHoqMhnG6ycMX"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_path"}}]},"finish_reason":null}],"usage":null,"obfuscation":"9VaQlqAL0733K"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"X8fPJ7F7fySxfcX"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"/"}}]},"finish_reason":null}],"usage":null,"obfuscation":"aK2f74mp2wM2sRp"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"tmp"}}]},"finish_reason":null}],"usage":null,"obfuscation":"oX7slWAWnHLIux3"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/cr"}}]},"finish_reason":null}],"usage":null,"obfuscation":"ZxNcdN9UuQZJnvz"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ush"}}]},"finish_reason":null}],"usage":null,"obfuscation":"F1ytiLT4ScIcfJ7"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"PGGohRCNriN1e"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/Test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"QSx40O0j61otG"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Coder"}}]},"finish_reason":null}],"usage":null,"obfuscation":"Ym8vcGe9jg3Yz"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Agent"}}]},"finish_reason":null}],"usage":null,"obfuscation":"qrvolLILQtDF4"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/open"}}]},"finish_reason":null}],"usage":null,"obfuscation":"0Kvu4BcJ0Wcow"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ai"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-g"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"pt"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-"}}]},"finish_reason":null}],"usage":null,"obfuscation":"D"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"5"}}]},"finish_reason":null}],"usage":null,"obfuscation":"d"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/m"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ult"}}]},"finish_reason":null}],"usage":null,"obfuscation":"Xyut1lPf2o3rtn8"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ied"}}]},"finish_reason":null}],"usage":null,"obfuscation":"nbl0M7BwiqjaheI"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"it"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_tool"}}]},"finish_reason":null}],"usage":null,"obfuscation":"yaUYgmlGQ2zuY"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/main"}}]},"finish_reason":null}],"usage":null,"obfuscation":"4N6OB88TiMRfd"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".go"}}]},"finish_reason":null}],"usage":null,"obfuscation":"CVUVMkIFN8RtSD5"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\",\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"mcwTKy0j9jaNQ"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ed"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"its"}}]},"finish_reason":null}],"usage":null,"obfuscation":"8Iu91COZLxsNdOF"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":["}}]},"finish_reason":null}],"usage":null,"obfuscation":"naJ0b33zQMcsD6"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"qO9g1TF37oYom1L"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"old"}}]},"finish_reason":null}],"usage":null,"obfuscation":"1GlNpZpV1joOlQ8"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_string"}}]},"finish_reason":null}],"usage":null,"obfuscation":"69lv9viYBNt"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"yZac6ciCYA8Au"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\\"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"tf"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"mt"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".Println"}}]},"finish_reason":null}],"usage":null,"obfuscation":"D45fag7olb"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"(\\\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"0YvD1z3x3aU5B"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Hello"}}]},"finish_reason":null}],"usage":null,"obfuscation":"BqeuV3LGh9ZCE"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":","}}]},"finish_reason":null}],"usage":null,"obfuscation":"Q"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" World"}}]},"finish_reason":null}],"usage":null,"obfuscation":"CIpikPLKmFBG"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"!\\"}}]},"finish_reason":null}],"usage":null,"obfuscation":"wFUtP5rBXhqh8mQ"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\")"}}]},"finish_reason":null}],"usage":null,"obfuscation":"FPg8X0LBvNtDG18"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\",\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"IAsvrIjIy5BSm"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"new"}}]},"finish_reason":null}],"usage":null,"obfuscation":"6deeZaT1YLHBTUa"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_string"}}]},"finish_reason":null}],"usage":null,"obfuscation":"E0RuwBGMArQ"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"yI38vn5tLT4Jz"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\\"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"tf"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"mt"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".Println"}}]},"finish_reason":null}],"usage":null,"obfuscation":"2aHJdp6ejI"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"(\\\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"jHcBbuB7Sd8e4"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Hello"}}]},"finish_reason":null}],"usage":null,"obfuscation":"H6O3IzJS1Zz1S"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":","}}]},"finish_reason":null}],"usage":null,"obfuscation":"U"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" Crush"}}]},"finish_reason":null}],"usage":null,"obfuscation":"MZHACarcT9rF"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"!\\"}}]},"finish_reason":null}],"usage":null,"obfuscation":"kUYWcLIlVRBJvrw"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\")"}}]},"finish_reason":null}],"usage":null,"obfuscation":"MiXyKSqFPFkgQTI"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"},{\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"946UM26RvHA"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"old"}}]},"finish_reason":null}],"usage":null,"obfuscation":"F6liAxZksmCVl6v"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_string"}}]},"finish_reason":null}],"usage":null,"obfuscation":"aPin7evU21a"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"RrRx62KTPFFoP"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\\"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"tf"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"mt"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".Println"}}]},"finish_reason":null}],"usage":null,"obfuscation":"p7ihowQSXv"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"(\\\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"MAZ91GSdE4NB7"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Hello"}}]},"finish_reason":null}],"usage":null,"obfuscation":"rSrhYDlju0ZUF"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":","}}]},"finish_reason":null}],"usage":null,"obfuscation":"L"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" Crush"}}]},"finish_reason":null}],"usage":null,"obfuscation":"5USN1NGkzKBL"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"!\\"}}]},"finish_reason":null}],"usage":null,"obfuscation":"FTQ1Rol6ZRzlJKy"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\")"}}]},"finish_reason":null}],"usage":null,"obfuscation":"lipmAa5WiSle2zz"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\",\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"gcun3Y0NOoBDF"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"new"}}]},"finish_reason":null}],"usage":null,"obfuscation":"P10rlprDSWIDAXP"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_string"}}]},"finish_reason":null}],"usage":null,"obfuscation":"wtVc9Ljodge"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"tZt4WE7j7sYJ4"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\\"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"t"}}]},"finish_reason":null}],"usage":null,"obfuscation":"I"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"//"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" Greeting"}}]},"finish_reason":null}],"usage":null,"obfuscation":"pQag3Khmg"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\\n"}}]},"finish_reason":null}],"usage":null,"obfuscation":"td7PEuIZxSzLj1d"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\\t"}}]},"finish_reason":null}],"usage":null,"obfuscation":"GLRdi3FrtUPlzdD"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"fmt"}}]},"finish_reason":null}],"usage":null,"obfuscation":"XTvOm1QIzdaTskN"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".Println"}}]},"finish_reason":null}],"usage":null,"obfuscation":"lscyWgSYlB"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"(\\\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"b52w7WJ4faPGV"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Hello"}}]},"finish_reason":null}],"usage":null,"obfuscation":"derHfTikhC25q"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":","}}]},"finish_reason":null}],"usage":null,"obfuscation":"v"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" Crush"}}]},"finish_reason":null}],"usage":null,"obfuscation":"YJCoL3yPCHLR"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"!\\"}}]},"finish_reason":null}],"usage":null,"obfuscation":"tiO9LTAC7cw0jdh"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\")"}}]},"finish_reason":null}],"usage":null,"obfuscation":"FLfgAsJggoaxgPV"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"}"}}]},"finish_reason":null}],"usage":null,"obfuscation":"CTGfRkCpvpyv0OS"}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"]}"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuHy7Mtk3xpDpzn9m6awpCn3ScGg","object":"chat.completion.chunk","created":1776258586,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6584,"completion_tokens":2030,"total_tokens":8614,"prompt_tokens_details":{"cached_tokens":6400,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":1920,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"fDgNkrTkw9EpwIN"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 26.705454083s
-- id: 4
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 32270
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/multiedit_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use multiedit to change 'Hello, World!' to 'Hello, Crush!' and add a comment '// Greeting' above the fmt.Println line in main.go\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_1EaQeVFsEBO0jON0b6tMK6ZT\",\"function\":{\"arguments\":\"{\\\"path\\\": \\\"/tmp/crush-test/TestCoderAgent/openai-gpt-5/multiedit_tool\\\", \\\"depth\\\": 2}\",\"name\":\"ls\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n- /tmp/crush-test/TestCoderAgent/openai-gpt-5/multiedit_tool/\\n - go.mod\\n - main.go\\n\",\"tool_call_id\":\"call_1EaQeVFsEBO0jON0b6tMK6ZT\",\"role\":\"tool\"},{\"tool_calls\":[{\"id\":\"call_Jx0qaG3KLJe5lS83V4Ke8s1s\",\"function\":{\"arguments\":\"{\\\"file_path\\\":\\\"/tmp/crush-test/TestCoderAgent/openai-gpt-5/multiedit_tool/main.go\\\",\\\"limit\\\":2000}\",\"name\":\"view\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\tfmt.Println(\\\"Hello, World!\\\")\\n 7|}\\n\\n\",\"tool_call_id\":\"call_Jx0qaG3KLJe5lS83V4Ke8s1s\",\"role\":\"tool\"},{\"tool_calls\":[{\"id\":\"call_2hzHz91aslzg48tDaSK7ttbD\",\"function\":{\"arguments\":\"{\\\"file_path\\\":\\\"/tmp/crush-test/TestCoderAgent/openai-gpt-5/multiedit_tool/main.go\\\",\\\"edits\\\":[{\\\"old_string\\\":\\\"\\\\tfmt.Println(\\\\\\\"Hello, World!\\\\\\\")\\\",\\\"new_string\\\":\\\"\\\\tfmt.Println(\\\\\\\"Hello, Crush!\\\\\\\")\\\"},{\\\"old_string\\\":\\\"\\\\tfmt.Println(\\\\\\\"Hello, Crush!\\\\\\\")\\\",\\\"new_string\\\":\\\"\\\\t// Greeting\\\\n\\\\tfmt.Println(\\\\\\\"Hello, Crush!\\\\\\\")\\\"}]}\",\"name\":\"multiedit\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\nApplied 2 edits to file: /tmp/crush-test/TestCoderAgent/openai-gpt-5/multiedit_tool/main.go\\n\\n\",\"tool_call_id\":\"call_2hzHz91aslzg48tDaSK7ttbD\",\"role\":\"tool\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":null,"tool_calls":[{"index":0,"id":"call_o2H2M83Fmvv7NupbzGGdAOdT","type":"function","function":{"name":"view","arguments":""}}],"refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"ogBClHPkE2gR"}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"LLsPuvVnIizrlyT"}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"file"}}]},"finish_reason":null}],"usage":null,"obfuscation":"cnKBC8yZWSYScL"}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_path"}}]},"finish_reason":null}],"usage":null,"obfuscation":"dLrJC6BzG21KP"}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"f7Z4f3Cah5PqmCB"}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"/"}}]},"finish_reason":null}],"usage":null,"obfuscation":"lTT0X2T3n2xrzAk"}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"tmp"}}]},"finish_reason":null}],"usage":null,"obfuscation":"W4809tvLRgSjBx7"}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/cr"}}]},"finish_reason":null}],"usage":null,"obfuscation":"f0UJxZewQe7PHyb"}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ush"}}]},"finish_reason":null}],"usage":null,"obfuscation":"cefvjCk9o2V5isI"}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"5fvWhWjvK6dTy"}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/Test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"lhdWOG5rM8r0d"}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Coder"}}]},"finish_reason":null}],"usage":null,"obfuscation":"0hnM3RqQcWFJK"}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Agent"}}]},"finish_reason":null}],"usage":null,"obfuscation":"6mDIadQGb9sIl"}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/open"}}]},"finish_reason":null}],"usage":null,"obfuscation":"2JjcJqyMqYHJT"}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ai"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-g"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"pt"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-"}}]},"finish_reason":null}],"usage":null,"obfuscation":"X"}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"5"}}]},"finish_reason":null}],"usage":null,"obfuscation":"3"}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/m"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ult"}}]},"finish_reason":null}],"usage":null,"obfuscation":"ulhw97oYtCeG3dN"}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ied"}}]},"finish_reason":null}],"usage":null,"obfuscation":"Eq8AwXyZgyWCAay"}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"it"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_tool"}}]},"finish_reason":null}],"usage":null,"obfuscation":"VH0XyHYohddGS"}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/main"}}]},"finish_reason":null}],"usage":null,"obfuscation":"sjnUFxqQQ0SZx"}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".go"}}]},"finish_reason":null}],"usage":null,"obfuscation":"ques4BNk8ODtOye"}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\",\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"0AoVupG6G8ygm"}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"limit"}}]},"finish_reason":null}],"usage":null,"obfuscation":"h3auj0l9hhg8L"}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"PRXqQxJqEa6rtxX"}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"200"}}]},"finish_reason":null}],"usage":null,"obfuscation":"GkKd4lsjzVbg4nI"}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"0"}}]},"finish_reason":null}],"usage":null,"obfuscation":"c"}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"}"}}]},"finish_reason":null}],"usage":null,"obfuscation":"i"}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUuIQLHqPulQpfigACidOJUgtm50c","object":"chat.completion.chunk","created":1776258614,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6735,"completion_tokens":42,"total_tokens":6777,"prompt_tokens_details":{"cached_tokens":6528,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":0,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"xSkE"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 723.4035ms
-- id: 5
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 32756
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/multiedit_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use multiedit to change 'Hello, World!' to 'Hello, Crush!' and add a comment '// Greeting' above the fmt.Println line in main.go\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_1EaQeVFsEBO0jON0b6tMK6ZT\",\"function\":{\"arguments\":\"{\\\"path\\\": \\\"/tmp/crush-test/TestCoderAgent/openai-gpt-5/multiedit_tool\\\", \\\"depth\\\": 2}\",\"name\":\"ls\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n- /tmp/crush-test/TestCoderAgent/openai-gpt-5/multiedit_tool/\\n - go.mod\\n - main.go\\n\",\"tool_call_id\":\"call_1EaQeVFsEBO0jON0b6tMK6ZT\",\"role\":\"tool\"},{\"tool_calls\":[{\"id\":\"call_Jx0qaG3KLJe5lS83V4Ke8s1s\",\"function\":{\"arguments\":\"{\\\"file_path\\\":\\\"/tmp/crush-test/TestCoderAgent/openai-gpt-5/multiedit_tool/main.go\\\",\\\"limit\\\":2000}\",\"name\":\"view\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\tfmt.Println(\\\"Hello, World!\\\")\\n 7|}\\n\\n\",\"tool_call_id\":\"call_Jx0qaG3KLJe5lS83V4Ke8s1s\",\"role\":\"tool\"},{\"tool_calls\":[{\"id\":\"call_2hzHz91aslzg48tDaSK7ttbD\",\"function\":{\"arguments\":\"{\\\"file_path\\\":\\\"/tmp/crush-test/TestCoderAgent/openai-gpt-5/multiedit_tool/main.go\\\",\\\"edits\\\":[{\\\"old_string\\\":\\\"\\\\tfmt.Println(\\\\\\\"Hello, World!\\\\\\\")\\\",\\\"new_string\\\":\\\"\\\\tfmt.Println(\\\\\\\"Hello, Crush!\\\\\\\")\\\"},{\\\"old_string\\\":\\\"\\\\tfmt.Println(\\\\\\\"Hello, Crush!\\\\\\\")\\\",\\\"new_string\\\":\\\"\\\\t// Greeting\\\\n\\\\tfmt.Println(\\\\\\\"Hello, Crush!\\\\\\\")\\\"}]}\",\"name\":\"multiedit\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\nApplied 2 edits to file: /tmp/crush-test/TestCoderAgent/openai-gpt-5/multiedit_tool/main.go\\n\\n\",\"tool_call_id\":\"call_2hzHz91aslzg48tDaSK7ttbD\",\"role\":\"tool\"},{\"tool_calls\":[{\"id\":\"call_o2H2M83Fmvv7NupbzGGdAOdT\",\"function\":{\"arguments\":\"{\\\"file_path\\\":\\\"/tmp/crush-test/TestCoderAgent/openai-gpt-5/multiedit_tool/main.go\\\",\\\"limit\\\":2000}\",\"name\":\"view\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\t// Greeting\\n 7|\\tfmt.Println(\\\"Hello, Crush!\\\")\\n 8|}\\n\\n\",\"tool_call_id\":\"call_o2H2M83Fmvv7NupbzGGdAOdT\",\"role\":\"tool\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUuIR6uPE9nOcpsbeIuRrIUHNUZVo","object":"chat.completion.chunk","created":1776258615,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":"","refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"MKXmyB5rnz"}
-
- data: {"id":"chatcmpl-DUuIR6uPE9nOcpsbeIuRrIUHNUZVo","object":"chat.completion.chunk","created":1776258615,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"Done"},"finish_reason":null}],"usage":null,"obfuscation":"MNqTvJzC"}
-
- data: {"id":"chatcmpl-DUuIR6uPE9nOcpsbeIuRrIUHNUZVo","object":"chat.completion.chunk","created":1776258615,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":null,"obfuscation":"9biSOR"}
-
- data: {"id":"chatcmpl-DUuIR6uPE9nOcpsbeIuRrIUHNUZVo","object":"chat.completion.chunk","created":1776258615,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6847,"completion_tokens":4,"total_tokens":6851,"prompt_tokens_details":{"cached_tokens":6656,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":0,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"Qg5Ef"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 736.791291ms
diff --git a/internal/agent/testdata/TestCoderAgent/openai-gpt-5/read_a_file.yaml b/internal/agent/testdata/TestCoderAgent/openai-gpt-5/read_a_file.yaml
deleted file mode 100644
index 72d45b9ce97e53436fbf244a342fa06e1829618d..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/openai-gpt-5/read_a_file.yaml
+++ /dev/null
@@ -1,293 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 757
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nRead the go mod\n \n\n","role":"user"}],"model":"gpt-4o","max_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu6dqBgbhU1UPRnCxHakFCyc9onb","object":"chat.completion.chunk","created":1776257883,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"role":"assistant","content":"","refusal":null},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"EwmOllIzTgofTi"}
-
- data: {"id":"chatcmpl-DUu6dqBgbhU1UPRnCxHakFCyc9onb","object":"chat.completion.chunk","created":1776257883,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":"Reading"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"cjOicLixi"}
-
- data: {"id":"chatcmpl-DUu6dqBgbhU1UPRnCxHakFCyc9onb","object":"chat.completion.chunk","created":1776257883,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" Go"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"QtnAAPDcRXyMB"}
-
- data: {"id":"chatcmpl-DUu6dqBgbhU1UPRnCxHakFCyc9onb","object":"chat.completion.chunk","created":1776257883,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" Module"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"ZWwjoOccl"}
-
- data: {"id":"chatcmpl-DUu6dqBgbhU1UPRnCxHakFCyc9onb","object":"chat.completion.chunk","created":1776257883,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" Information"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"rxaq"}
-
- data: {"id":"chatcmpl-DUu6dqBgbhU1UPRnCxHakFCyc9onb","object":"chat.completion.chunk","created":1776257883,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{},"logprobs":null,"finish_reason":"stop"}],"usage":null,"obfuscation":"VeSW8VDmIo"}
-
- data: {"id":"chatcmpl-DUu6dqBgbhU1UPRnCxHakFCyc9onb","object":"chat.completion.chunk","created":1776257883,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[],"usage":{"prompt_tokens":145,"completion_tokens":4,"total_tokens":149,"prompt_tokens_details":{"cached_tokens":0,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":0,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"r9JMj8ytipcFIU"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 1.372738958s
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30626
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/read_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"Read the go mod\",\"role\":\"user\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu6dbcDDCxFP4pPfEuGDi8hPOPj5","object":"chat.completion.chunk","created":1776257883,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":null,"tool_calls":[{"index":0,"id":"call_CIO1zmREyjXAUexomrJaocFg","type":"function","function":{"name":"ls","arguments":""}}],"refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"yGujjKPe9YFQG2"}
-
- data: {"id":"chatcmpl-DUu6dbcDDCxFP4pPfEuGDi8hPOPj5","object":"chat.completion.chunk","created":1776257883,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"8MJCODWpQEKPvMf"}
-
- data: {"id":"chatcmpl-DUu6dbcDDCxFP4pPfEuGDi8hPOPj5","object":"chat.completion.chunk","created":1776257883,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"path"}}]},"finish_reason":null}],"usage":null,"obfuscation":"1LKoNBOtfwPlmf"}
-
- data: {"id":"chatcmpl-DUu6dbcDDCxFP4pPfEuGDi8hPOPj5","object":"chat.completion.chunk","created":1776257883,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"2JXfcIk3YtP4Fu3"}
-
- data: {"id":"chatcmpl-DUu6dbcDDCxFP4pPfEuGDi8hPOPj5","object":"chat.completion.chunk","created":1776257883,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" \"/"}}]},"finish_reason":null}],"usage":null,"obfuscation":"gENbFWsIuVPprx"}
-
- data: {"id":"chatcmpl-DUu6dbcDDCxFP4pPfEuGDi8hPOPj5","object":"chat.completion.chunk","created":1776257883,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"tmp"}}]},"finish_reason":null}],"usage":null,"obfuscation":"6CZmtEUqH5E4ira"}
-
- data: {"id":"chatcmpl-DUu6dbcDDCxFP4pPfEuGDi8hPOPj5","object":"chat.completion.chunk","created":1776257883,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/cr"}}]},"finish_reason":null}],"usage":null,"obfuscation":"xDSBAifacOR0WjB"}
-
- data: {"id":"chatcmpl-DUu6dbcDDCxFP4pPfEuGDi8hPOPj5","object":"chat.completion.chunk","created":1776257883,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ush"}}]},"finish_reason":null}],"usage":null,"obfuscation":"9CyqpqW4Gkem9XS"}
-
- data: {"id":"chatcmpl-DUu6dbcDDCxFP4pPfEuGDi8hPOPj5","object":"chat.completion.chunk","created":1776257883,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"LubFBJ2mYOqZ1"}
-
- data: {"id":"chatcmpl-DUu6dbcDDCxFP4pPfEuGDi8hPOPj5","object":"chat.completion.chunk","created":1776257883,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/Test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"actCnm3hyyEMf"}
-
- data: {"id":"chatcmpl-DUu6dbcDDCxFP4pPfEuGDi8hPOPj5","object":"chat.completion.chunk","created":1776257883,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Coder"}}]},"finish_reason":null}],"usage":null,"obfuscation":"5ANdtr7zM8ao5"}
-
- data: {"id":"chatcmpl-DUu6dbcDDCxFP4pPfEuGDi8hPOPj5","object":"chat.completion.chunk","created":1776257883,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Agent"}}]},"finish_reason":null}],"usage":null,"obfuscation":"IdIx9tJ3KbPKC"}
-
- data: {"id":"chatcmpl-DUu6dbcDDCxFP4pPfEuGDi8hPOPj5","object":"chat.completion.chunk","created":1776257883,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/open"}}]},"finish_reason":null}],"usage":null,"obfuscation":"sMiufd5D8Ujkt"}
-
- data: {"id":"chatcmpl-DUu6dbcDDCxFP4pPfEuGDi8hPOPj5","object":"chat.completion.chunk","created":1776257883,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ai"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu6dbcDDCxFP4pPfEuGDi8hPOPj5","object":"chat.completion.chunk","created":1776257883,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-g"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu6dbcDDCxFP4pPfEuGDi8hPOPj5","object":"chat.completion.chunk","created":1776257883,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"pt"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu6dbcDDCxFP4pPfEuGDi8hPOPj5","object":"chat.completion.chunk","created":1776257883,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-"}}]},"finish_reason":null}],"usage":null,"obfuscation":"R"}
-
- data: {"id":"chatcmpl-DUu6dbcDDCxFP4pPfEuGDi8hPOPj5","object":"chat.completion.chunk","created":1776257883,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"5"}}]},"finish_reason":null}],"usage":null,"obfuscation":"4"}
-
- data: {"id":"chatcmpl-DUu6dbcDDCxFP4pPfEuGDi8hPOPj5","object":"chat.completion.chunk","created":1776257883,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/read"}}]},"finish_reason":null}],"usage":null,"obfuscation":"RdHbHz9h5etS4"}
-
- data: {"id":"chatcmpl-DUu6dbcDDCxFP4pPfEuGDi8hPOPj5","object":"chat.completion.chunk","created":1776257883,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_a"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu6dbcDDCxFP4pPfEuGDi8hPOPj5","object":"chat.completion.chunk","created":1776257883,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_file"}}]},"finish_reason":null}],"usage":null,"obfuscation":"feY3enDVuwFuy"}
-
- data: {"id":"chatcmpl-DUu6dbcDDCxFP4pPfEuGDi8hPOPj5","object":"chat.completion.chunk","created":1776257883,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\","}}]},"finish_reason":null}],"usage":null,"obfuscation":"ssqGK5Eiu8I2k8z"}
-
- data: {"id":"chatcmpl-DUu6dbcDDCxFP4pPfEuGDi8hPOPj5","object":"chat.completion.chunk","created":1776257883,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" \""}}]},"finish_reason":null}],"usage":null,"obfuscation":"xw9DKf8EwAeL0u2"}
-
- data: {"id":"chatcmpl-DUu6dbcDDCxFP4pPfEuGDi8hPOPj5","object":"chat.completion.chunk","created":1776257883,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"depth"}}]},"finish_reason":null}],"usage":null,"obfuscation":"F7xzqGKtJa93V"}
-
- data: {"id":"chatcmpl-DUu6dbcDDCxFP4pPfEuGDi8hPOPj5","object":"chat.completion.chunk","created":1776257883,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"3Fsfl5KaN5eJlOX"}
-
- data: {"id":"chatcmpl-DUu6dbcDDCxFP4pPfEuGDi8hPOPj5","object":"chat.completion.chunk","created":1776257883,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" "}}]},"finish_reason":null}],"usage":null,"obfuscation":"L"}
-
- data: {"id":"chatcmpl-DUu6dbcDDCxFP4pPfEuGDi8hPOPj5","object":"chat.completion.chunk","created":1776257883,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"2"}}]},"finish_reason":null}],"usage":null,"obfuscation":"R"}
-
- data: {"id":"chatcmpl-DUu6dbcDDCxFP4pPfEuGDi8hPOPj5","object":"chat.completion.chunk","created":1776257883,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"}"}}]},"finish_reason":null}],"usage":null,"obfuscation":"t"}
-
- data: {"id":"chatcmpl-DUu6dbcDDCxFP4pPfEuGDi8hPOPj5","object":"chat.completion.chunk","created":1776257883,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu6dbcDDCxFP4pPfEuGDi8hPOPj5","object":"chat.completion.chunk","created":1776257883,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6369,"completion_tokens":173,"total_tokens":6542,"prompt_tokens_details":{"cached_tokens":0,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":128,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"Ch4Z"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 4.421312125s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31007
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/read_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"Read the go mod\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_CIO1zmREyjXAUexomrJaocFg\",\"function\":{\"arguments\":\"{\\\"path\\\": \\\"/tmp/crush-test/TestCoderAgent/openai-gpt-5/read_a_file\\\", \\\"depth\\\": 2}\",\"name\":\"ls\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n- /tmp/crush-test/TestCoderAgent/openai-gpt-5/read_a_file/\\n - go.mod\\n - main.go\\n\",\"tool_call_id\":\"call_CIO1zmREyjXAUexomrJaocFg\",\"role\":\"tool\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":null,"tool_calls":[{"index":0,"id":"call_oCbvYT1188mVJglkWmIyT5pj","type":"function","function":{"name":"view","arguments":""}}],"refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"HmfiC5GhEclZ"}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"GuopYBSwgEpPRl8"}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"file"}}]},"finish_reason":null}],"usage":null,"obfuscation":"6bsDo4NTo3r1UH"}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_path"}}]},"finish_reason":null}],"usage":null,"obfuscation":"RXLSIs57WCwig"}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"N9QccLkir4tbeX2"}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"/"}}]},"finish_reason":null}],"usage":null,"obfuscation":"Rt6zEkamMgm58zZ"}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"tmp"}}]},"finish_reason":null}],"usage":null,"obfuscation":"ZiaL80yVZ1YsHxZ"}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/cr"}}]},"finish_reason":null}],"usage":null,"obfuscation":"LPO7jwZRToG3DeT"}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ush"}}]},"finish_reason":null}],"usage":null,"obfuscation":"yxcj3TLPRFTIYh4"}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"z9Ve3tyPwXSu6"}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/Test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"YnrCmMxxfhBXE"}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Coder"}}]},"finish_reason":null}],"usage":null,"obfuscation":"B69t3ZRrm0O7e"}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Agent"}}]},"finish_reason":null}],"usage":null,"obfuscation":"gFuVIbBlP45z2"}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/open"}}]},"finish_reason":null}],"usage":null,"obfuscation":"P5gbmr87SCVa5"}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ai"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-g"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"pt"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-"}}]},"finish_reason":null}],"usage":null,"obfuscation":"4"}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"5"}}]},"finish_reason":null}],"usage":null,"obfuscation":"v"}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/read"}}]},"finish_reason":null}],"usage":null,"obfuscation":"GOYIgcgPRakg7"}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_a"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_file"}}]},"finish_reason":null}],"usage":null,"obfuscation":"nIM7XotrFDBDL"}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/go"}}]},"finish_reason":null}],"usage":null,"obfuscation":"LjgapO1cx64UyJV"}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".mod"}}]},"finish_reason":null}],"usage":null,"obfuscation":"mI6DdrDUIl6suv"}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\",\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"6WRs8JJHMwW5u"}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"limit"}}]},"finish_reason":null}],"usage":null,"obfuscation":"fFP5pmOSh31cr"}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"Em3kdAUwiXOxnqf"}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"200"}}]},"finish_reason":null}],"usage":null,"obfuscation":"BZe8ayX5h6Y0Jc8"}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"0"}}]},"finish_reason":null}],"usage":null,"obfuscation":"w"}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"}"}}]},"finish_reason":null}],"usage":null,"obfuscation":"I"}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu6iMyxYxa7SWAUsZkVUU26EboFJ","object":"chat.completion.chunk","created":1776257888,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6446,"completion_tokens":40,"total_tokens":6486,"prompt_tokens_details":{"cached_tokens":6272,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":0,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"jcD6"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 1.571665542s
-- id: 3
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31396
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/read_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"Read the go mod\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_CIO1zmREyjXAUexomrJaocFg\",\"function\":{\"arguments\":\"{\\\"path\\\": \\\"/tmp/crush-test/TestCoderAgent/openai-gpt-5/read_a_file\\\", \\\"depth\\\": 2}\",\"name\":\"ls\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n- /tmp/crush-test/TestCoderAgent/openai-gpt-5/read_a_file/\\n - go.mod\\n - main.go\\n\",\"tool_call_id\":\"call_CIO1zmREyjXAUexomrJaocFg\",\"role\":\"tool\"},{\"tool_calls\":[{\"id\":\"call_oCbvYT1188mVJglkWmIyT5pj\",\"function\":{\"arguments\":\"{\\\"file_path\\\":\\\"/tmp/crush-test/TestCoderAgent/openai-gpt-5/read_a_file/go.mod\\\",\\\"limit\\\":2000}\",\"name\":\"view\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n 1|module example.com/testproject\\n 2|\\n 3|go 1.23\\n\\n\",\"tool_call_id\":\"call_oCbvYT1188mVJglkWmIyT5pj\",\"role\":\"tool\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu6jIlXuof02FXifT3s1qjTQFIfl","object":"chat.completion.chunk","created":1776257889,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":"","refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"BPglBC0tmN"}
-
- data: {"id":"chatcmpl-DUu6jIlXuof02FXifT3s1qjTQFIfl","object":"chat.completion.chunk","created":1776257889,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"module"},"finish_reason":null}],"usage":null,"obfuscation":"HRC7Ai"}
-
- data: {"id":"chatcmpl-DUu6jIlXuof02FXifT3s1qjTQFIfl","object":"chat.completion.chunk","created":1776257889,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":" example"},"finish_reason":null}],"usage":null,"obfuscation":"9gPl"}
-
- data: {"id":"chatcmpl-DUu6jIlXuof02FXifT3s1qjTQFIfl","object":"chat.completion.chunk","created":1776257889,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":".com"},"finish_reason":null}],"usage":null,"obfuscation":"eerpo3Ih"}
-
- data: {"id":"chatcmpl-DUu6jIlXuof02FXifT3s1qjTQFIfl","object":"chat.completion.chunk","created":1776257889,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/test"},"finish_reason":null}],"usage":null,"obfuscation":"Nq2Uemn"}
-
- data: {"id":"chatcmpl-DUu6jIlXuof02FXifT3s1qjTQFIfl","object":"chat.completion.chunk","created":1776257889,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"project"},"finish_reason":null}],"usage":null,"obfuscation":"VGE8g"}
-
- data: {"id":"chatcmpl-DUu6jIlXuof02FXifT3s1qjTQFIfl","object":"chat.completion.chunk","created":1776257889,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"\n"},"finish_reason":null}],"usage":null,"obfuscation":"XxKssje7aK"}
-
- data: {"id":"chatcmpl-DUu6jIlXuof02FXifT3s1qjTQFIfl","object":"chat.completion.chunk","created":1776257889,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"go"},"finish_reason":null}],"usage":null,"obfuscation":"fXZzlZin5J"}
-
- data: {"id":"chatcmpl-DUu6jIlXuof02FXifT3s1qjTQFIfl","object":"chat.completion.chunk","created":1776257889,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":" "},"finish_reason":null}],"usage":null,"obfuscation":"oxc1TkI6h7J"}
-
- data: {"id":"chatcmpl-DUu6jIlXuof02FXifT3s1qjTQFIfl","object":"chat.completion.chunk","created":1776257889,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"1"},"finish_reason":null}],"usage":null,"obfuscation":"yHOhTKyzdTP"}
-
- data: {"id":"chatcmpl-DUu6jIlXuof02FXifT3s1qjTQFIfl","object":"chat.completion.chunk","created":1776257889,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"."},"finish_reason":null}],"usage":null,"obfuscation":"MfK9s5VT6Lq"}
-
- data: {"id":"chatcmpl-DUu6jIlXuof02FXifT3s1qjTQFIfl","object":"chat.completion.chunk","created":1776257889,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"23"},"finish_reason":null}],"usage":null,"obfuscation":"ZK2JATEXSs"}
-
- data: {"id":"chatcmpl-DUu6jIlXuof02FXifT3s1qjTQFIfl","object":"chat.completion.chunk","created":1776257889,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":null,"obfuscation":"6iyLi0"}
-
- data: {"id":"chatcmpl-DUu6jIlXuof02FXifT3s1qjTQFIfl","object":"chat.completion.chunk","created":1776257889,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6526,"completion_tokens":14,"total_tokens":6540,"prompt_tokens_details":{"cached_tokens":6400,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":0,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"U7RU"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 738.906917ms
diff --git a/internal/agent/testdata/TestCoderAgent/openai-gpt-5/simple_test.yaml b/internal/agent/testdata/TestCoderAgent/openai-gpt-5/simple_test.yaml
deleted file mode 100644
index becf9465138abe8f16305b2028045cbaa4062ebe..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/openai-gpt-5/simple_test.yaml
+++ /dev/null
@@ -1,83 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 747
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nHello\n \n\n","role":"user"}],"model":"gpt-4o","max_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu6aInjKVkX6LT00KPNjZ0Gb1TG2","object":"chat.completion.chunk","created":1776257880,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"role":"assistant","content":"","refusal":null},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"Ih6EAwz94Qt5Lg"}
-
- data: {"id":"chatcmpl-DUu6aInjKVkX6LT00KPNjZ0Gb1TG2","object":"chat.completion.chunk","created":1776257880,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":"Greeting"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"zzUirghY"}
-
- data: {"id":"chatcmpl-DUu6aInjKVkX6LT00KPNjZ0Gb1TG2","object":"chat.completion.chunk","created":1776257880,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" Message"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"wxDMlPti"}
-
- data: {"id":"chatcmpl-DUu6aInjKVkX6LT00KPNjZ0Gb1TG2","object":"chat.completion.chunk","created":1776257880,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{},"logprobs":null,"finish_reason":"stop"}],"usage":null,"obfuscation":"oajOqsk75J"}
-
- data: {"id":"chatcmpl-DUu6aInjKVkX6LT00KPNjZ0Gb1TG2","object":"chat.completion.chunk","created":1776257880,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[],"usage":{"prompt_tokens":142,"completion_tokens":2,"total_tokens":144,"prompt_tokens_details":{"cached_tokens":0,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":0,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"Un8176gp7eRd67"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 2.400276875s
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30616
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/simple_test\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"Hello\",\"role\":\"user\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu6ak0lUGmX9XRTA78t3OCksv15F","object":"chat.completion.chunk","created":1776257880,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":"","refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"6QgW7XwWG5"}
-
- data: {"id":"chatcmpl-DUu6ak0lUGmX9XRTA78t3OCksv15F","object":"chat.completion.chunk","created":1776257880,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"Hi"},"finish_reason":null}],"usage":null,"obfuscation":"5y6ODLiKxP"}
-
- data: {"id":"chatcmpl-DUu6ak0lUGmX9XRTA78t3OCksv15F","object":"chat.completion.chunk","created":1776257880,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":null,"obfuscation":"8Jy2Ry"}
-
- data: {"id":"chatcmpl-DUu6ak0lUGmX9XRTA78t3OCksv15F","object":"chat.completion.chunk","created":1776257880,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6365,"completion_tokens":74,"total_tokens":6439,"prompt_tokens_details":{"cached_tokens":0,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":64,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"MvMmvH"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 4.051119584s
diff --git a/internal/agent/testdata/TestCoderAgent/openai-gpt-5/sourcegraph_tool.yaml b/internal/agent/testdata/TestCoderAgent/openai-gpt-5/sourcegraph_tool.yaml
deleted file mode 100644
index 78f2e56985f26108aa583f51de90072bbd151465..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/openai-gpt-5/sourcegraph_tool.yaml
+++ /dev/null
@@ -1,412 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 802
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse sourcegraph to search for ''func main'' in Go repositories\n \n\n","role":"user"}],"model":"gpt-4o","max_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu8PgjWhzEBZ68J5cdeqYLGGLCfn","object":"chat.completion.chunk","created":1776257993,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"role":"assistant","content":"","refusal":null},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"TLIhDuNot1v6xm"}
-
- data: {"id":"chatcmpl-DUu8PgjWhzEBZ68J5cdeqYLGGLCfn","object":"chat.completion.chunk","created":1776257993,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":"Searching"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"zzHJdCx"}
-
- data: {"id":"chatcmpl-DUu8PgjWhzEBZ68J5cdeqYLGGLCfn","object":"chat.completion.chunk","created":1776257993,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":" '"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"mcqjdcqPhyB95D"}
-
- data: {"id":"chatcmpl-DUu8PgjWhzEBZ68J5cdeqYLGGLCfn","object":"chat.completion.chunk","created":1776257993,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":"func"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"TcarK8oZfyPg"}
-
- data: {"id":"chatcmpl-DUu8PgjWhzEBZ68J5cdeqYLGGLCfn","object":"chat.completion.chunk","created":1776257993,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":" main"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"PNb5sfnv46S"}
-
- data: {"id":"chatcmpl-DUu8PgjWhzEBZ68J5cdeqYLGGLCfn","object":"chat.completion.chunk","created":1776257993,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":"'"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"3Vpc2Y9IUXz0XEY"}
-
- data: {"id":"chatcmpl-DUu8PgjWhzEBZ68J5cdeqYLGGLCfn","object":"chat.completion.chunk","created":1776257993,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":" in"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"VVXceRwtpom8t"}
-
- data: {"id":"chatcmpl-DUu8PgjWhzEBZ68J5cdeqYLGGLCfn","object":"chat.completion.chunk","created":1776257993,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":" Go"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"H5hyBr4mmUZIY"}
-
- data: {"id":"chatcmpl-DUu8PgjWhzEBZ68J5cdeqYLGGLCfn","object":"chat.completion.chunk","created":1776257993,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":" Re"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"t54vf8ssB1HQh"}
-
- data: {"id":"chatcmpl-DUu8PgjWhzEBZ68J5cdeqYLGGLCfn","object":"chat.completion.chunk","created":1776257993,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":"positories"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"bPh76T"}
-
- data: {"id":"chatcmpl-DUu8PgjWhzEBZ68J5cdeqYLGGLCfn","object":"chat.completion.chunk","created":1776257993,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":" Using"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"BQKXikInH7"}
-
- data: {"id":"chatcmpl-DUu8PgjWhzEBZ68J5cdeqYLGGLCfn","object":"chat.completion.chunk","created":1776257993,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":" Source"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"utFZ1tXLk"}
-
- data: {"id":"chatcmpl-DUu8PgjWhzEBZ68J5cdeqYLGGLCfn","object":"chat.completion.chunk","created":1776257993,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":"graph"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"BOZEYp094Vm"}
-
- data: {"id":"chatcmpl-DUu8PgjWhzEBZ68J5cdeqYLGGLCfn","object":"chat.completion.chunk","created":1776257993,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{},"logprobs":null,"finish_reason":"stop"}],"usage":null,"obfuscation":"9nGlrL9PcC"}
-
- data: {"id":"chatcmpl-DUu8PgjWhzEBZ68J5cdeqYLGGLCfn","object":"chat.completion.chunk","created":1776257993,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[],"usage":{"prompt_tokens":154,"completion_tokens":12,"total_tokens":166,"prompt_tokens_details":{"cached_tokens":0,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":0,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"4GNnu13ZQlgrM"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 531.269ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30676
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/sourcegraph_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use sourcegraph to search for 'func main' in Go repositories\",\"role\":\"user\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu8PKyTvWQ8YyjbSITfZdaLVf4i3","object":"chat.completion.chunk","created":1776257993,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":null,"tool_calls":[{"index":0,"id":"call_Fmhtvwj7YgCbE2050w62UqIA","type":"function","function":{"name":"sourcegraph","arguments":""}}],"refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"YYVm0"}
-
- data: {"id":"chatcmpl-DUu8PKyTvWQ8YyjbSITfZdaLVf4i3","object":"chat.completion.chunk","created":1776257993,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"EPZeIg242CvciZi"}
-
- data: {"id":"chatcmpl-DUu8PKyTvWQ8YyjbSITfZdaLVf4i3","object":"chat.completion.chunk","created":1776257993,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"query"}}]},"finish_reason":null}],"usage":null,"obfuscation":"eged8e2flnbo7"}
-
- data: {"id":"chatcmpl-DUu8PKyTvWQ8YyjbSITfZdaLVf4i3","object":"chat.completion.chunk","created":1776257993,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"uw0bxDX4FY2pV"}
-
- data: {"id":"chatcmpl-DUu8PKyTvWQ8YyjbSITfZdaLVf4i3","object":"chat.completion.chunk","created":1776257993,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"lang"}}]},"finish_reason":null}],"usage":null,"obfuscation":"OXHtiOw68h2wP7"}
-
- data: {"id":"chatcmpl-DUu8PKyTvWQ8YyjbSITfZdaLVf4i3","object":"chat.completion.chunk","created":1776257993,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"O"}
-
- data: {"id":"chatcmpl-DUu8PKyTvWQ8YyjbSITfZdaLVf4i3","object":"chat.completion.chunk","created":1776257993,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"go"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu8PKyTvWQ8YyjbSITfZdaLVf4i3","object":"chat.completion.chunk","created":1776257993,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" \\\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"5G5IZmMewBK77"}
-
- data: {"id":"chatcmpl-DUu8PKyTvWQ8YyjbSITfZdaLVf4i3","object":"chat.completion.chunk","created":1776257993,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"func"}}]},"finish_reason":null}],"usage":null,"obfuscation":"gsBK7sc3jsHGfU"}
-
- data: {"id":"chatcmpl-DUu8PKyTvWQ8YyjbSITfZdaLVf4i3","object":"chat.completion.chunk","created":1776257993,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" main"}}]},"finish_reason":null}],"usage":null,"obfuscation":"QfXviIdPwlZbt"}
-
- data: {"id":"chatcmpl-DUu8PKyTvWQ8YyjbSITfZdaLVf4i3","object":"chat.completion.chunk","created":1776257993,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\\\"\","}}]},"finish_reason":null}],"usage":null,"obfuscation":"nCy2N1370H8"}
-
- data: {"id":"chatcmpl-DUu8PKyTvWQ8YyjbSITfZdaLVf4i3","object":"chat.completion.chunk","created":1776257993,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" \""}}]},"finish_reason":null}],"usage":null,"obfuscation":"BfxGo12KNeibScf"}
-
- data: {"id":"chatcmpl-DUu8PKyTvWQ8YyjbSITfZdaLVf4i3","object":"chat.completion.chunk","created":1776257993,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"count"}}]},"finish_reason":null}],"usage":null,"obfuscation":"7XSkdK8Zo8SLD"}
-
- data: {"id":"chatcmpl-DUu8PKyTvWQ8YyjbSITfZdaLVf4i3","object":"chat.completion.chunk","created":1776257993,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"LNMDddfDOk6ZXJM"}
-
- data: {"id":"chatcmpl-DUu8PKyTvWQ8YyjbSITfZdaLVf4i3","object":"chat.completion.chunk","created":1776257993,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"20"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu8PKyTvWQ8YyjbSITfZdaLVf4i3","object":"chat.completion.chunk","created":1776257993,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":","}}]},"finish_reason":null}],"usage":null,"obfuscation":"1"}
-
- data: {"id":"chatcmpl-DUu8PKyTvWQ8YyjbSITfZdaLVf4i3","object":"chat.completion.chunk","created":1776257993,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" \""}}]},"finish_reason":null}],"usage":null,"obfuscation":"g5DhOAGYpeq4P56"}
-
- data: {"id":"chatcmpl-DUu8PKyTvWQ8YyjbSITfZdaLVf4i3","object":"chat.completion.chunk","created":1776257993,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"context"}}]},"finish_reason":null}],"usage":null,"obfuscation":"LBI9esGeIEe"}
-
- data: {"id":"chatcmpl-DUu8PKyTvWQ8YyjbSITfZdaLVf4i3","object":"chat.completion.chunk","created":1776257993,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_window"}}]},"finish_reason":null}],"usage":null,"obfuscation":"AUClcg0rhaj"}
-
- data: {"id":"chatcmpl-DUu8PKyTvWQ8YyjbSITfZdaLVf4i3","object":"chat.completion.chunk","created":1776257993,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"OlOVWIwqFcKPW2I"}
-
- data: {"id":"chatcmpl-DUu8PKyTvWQ8YyjbSITfZdaLVf4i3","object":"chat.completion.chunk","created":1776257993,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"5"}}]},"finish_reason":null}],"usage":null,"obfuscation":"x"}
-
- data: {"id":"chatcmpl-DUu8PKyTvWQ8YyjbSITfZdaLVf4i3","object":"chat.completion.chunk","created":1776257993,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":","}}]},"finish_reason":null}],"usage":null,"obfuscation":"P"}
-
- data: {"id":"chatcmpl-DUu8PKyTvWQ8YyjbSITfZdaLVf4i3","object":"chat.completion.chunk","created":1776257993,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" \""}}]},"finish_reason":null}],"usage":null,"obfuscation":"pjEmuV1LdgOcsAZ"}
-
- data: {"id":"chatcmpl-DUu8PKyTvWQ8YyjbSITfZdaLVf4i3","object":"chat.completion.chunk","created":1776257993,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"timeout"}}]},"finish_reason":null}],"usage":null,"obfuscation":"RPRrEGRSSoh"}
-
- data: {"id":"chatcmpl-DUu8PKyTvWQ8YyjbSITfZdaLVf4i3","object":"chat.completion.chunk","created":1776257993,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"CE12tnYDl3jhGZ3"}
-
- data: {"id":"chatcmpl-DUu8PKyTvWQ8YyjbSITfZdaLVf4i3","object":"chat.completion.chunk","created":1776257993,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"120"}}]},"finish_reason":null}],"usage":null,"obfuscation":"Gib5ydshwfcKQSK"}
-
- data: {"id":"chatcmpl-DUu8PKyTvWQ8YyjbSITfZdaLVf4i3","object":"chat.completion.chunk","created":1776257993,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"}"}}]},"finish_reason":null}],"usage":null,"obfuscation":"Z"}
-
- data: {"id":"chatcmpl-DUu8PKyTvWQ8YyjbSITfZdaLVf4i3","object":"chat.completion.chunk","created":1776257993,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu8PKyTvWQ8YyjbSITfZdaLVf4i3","object":"chat.completion.chunk","created":1776257993,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6378,"completion_tokens":428,"total_tokens":6806,"prompt_tokens_details":{"cached_tokens":3456,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":384,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"f"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 7.594252875s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 428
- host: sourcegraph.com
- body: '{"query":"query Search($query: String!) { search(query: $query, version: V2, patternType: keyword ) { results { matchCount, limitHit, resultCount, approximateResultCount, missing { name }, timedout { name }, indexUnavailable, results { __typename, ... on FileMatch { repository { name }, file { path, url, content }, lineMatches { preview, lineNumber, offsetAndLengths } } } } } }","variables":{"query":"lang:go \"func main\""}}'
- headers:
- Content-Type:
- - application/json
- User-Agent:
- - crush/1.0
- url: https://sourcegraph.com/.api/graphql
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: '{"data":{"search":{"results":{"matchCount":30,"limitHit":true,"resultCount":30,"approximateResultCount":"30+","missing":[],"timedout":[],"indexUnavailable":false,"results":[{"__typename":"FileMatch","repository":{"name":"github.com/wx-chevalier/Go-Notes"},"file":{"path":"99_参考资料/2017_A Huge Number of Go Examples/codes/first/first/main.go","url":"/r/github.com/wx-chevalier/Go-Notes/-/blob/99_%E5%8F%82%E8%80%83%E8%B5%84%E6%96%99/2017_A%20Huge%20Number%20of%20Go%20Examples/codes/first/first/main.go","content":"\n// package main is a special package\n// it allows Go to create an executable file\npackage main\n\n/*\nThis is a multi-line comment.\n\nimport keyword makes another package available\n for this .go \"file\".\n\nimport \"fmt\" lets you access fmt package''s functionality\n here in this file.\n*/\nimport \"fmt\"\n\n// \"func main\" is special.\n//\n// Go has to know where to start\n//\n// func main creates a starting point for Go\n//\n// After compiling the code,\n// Go runtime will first run this function\nfunc main() {\n\t// after: import \"fmt\"\n\t// Println function of \"fmt\" package becomes available\n\n\t// Look at what it looks like by typing in the console:\n\t// godoc -src fmt Println\n\n\t// Println is just an exported function from\n\t// \"fmt\" package\n\n\t// Exported = First Letter is uppercase\n\tfmt.Println(\"Hello Gopher!\")\n\n\t// Go cannot call Println function by itself.\n\t// That''s why you need to call it here.\n\t// It only calls `func main` automatically.\n\n\t// -----\n\n\t// Go supports Unicode characters in string literals\n\t// And also in source-code: KÖSTEBEK!\n\n\t// EXERCISE: Remove the comments from below --\u003e //\n\t// fmt.Println(\"Merhaba Köstebek!\")\n\n\t// Unnecessary note:\n\t// \"Merhaba Köstebek\" means \"Hello Gopher\"\n\t// in Turkish language\n}\n"},"lineMatches":[{"preview":"// \"func main\" is special.","lineNumber":16,"offsetAndLengths":[[3,11]]},{"preview":"func main() {","lineNumber":24,"offsetAndLengths":[[0,9]]},{"preview":"// func main creates a starting point for Go","lineNumber":20,"offsetAndLengths":[[3,9]]},{"preview":"\t// It only calls `func main` automatically.","lineNumber":39,"offsetAndLengths":[[19,9]]}]},{"__typename":"FileMatch","repository":{"name":"github.com/wx-chevalier/Go-Notes"},"file":{"path":"99_参考资料/2017_A Huge Number of Go Examples/codes/02_write-your-first-program/main.go","url":"/r/github.com/wx-chevalier/Go-Notes/-/blob/99_%E5%8F%82%E8%80%83%E8%B5%84%E6%96%99/2017_A%20Huge%20Number%20of%20Go%20Examples/codes/02_write-your-first-program/main.go","content":"\n// package main is a special package\n// it allows Go to create an executable file\npackage main\n\n/*\nThis is a multi-line comment.\n\nimport keyword makes another package available\n for this .go \"file\".\n\nimport \"fmt\" lets you access fmt package''s functionality\n here in this file.\n*/\nimport \"fmt\"\n\n// \"func main\" is special.\n//\n// Go has to know where to start\n//\n// func main creates a starting point for Go\n//\n// After compiling the code,\n// Go runtime will first run this function\nfunc main() {\n\t// after: import \"fmt\"\n\t// Println function of \"fmt\" package becomes available\n\n\t// Look at what it looks like by typing in the console:\n\t// go doc -src fmt Println\n\n\t// Println is just an exported function from\n\t// \"fmt\" package\n\n\t// Exported = First Letter is uppercase\n\tfmt.Println(\"Hello Gopher!\")\n\n\t// Go cannot call Println function by itself.\n\t// That''s why you need to call it here.\n\t// It only calls `func main` automatically.\n\n\t// -----\n\n\t// Go supports Unicode characters in string literals\n\t// And also in source-code: KÖSTEBEK!\n\t//\n\t// Because: Literal ~= Source Code\n\n\t// EXERCISE: Remove the comments from below --\u003e //\n\t// fmt.Println(\"Merhaba Köstebek!\")\n\n\t// Unnecessary note:\n\t// \"Merhaba Köstebek\" means \"Hello Gopher\"\n\t// in Turkish language\n}\n"},"lineMatches":[{"preview":"// \"func main\" is special.","lineNumber":16,"offsetAndLengths":[[3,11]]},{"preview":"func main() {","lineNumber":24,"offsetAndLengths":[[0,9]]},{"preview":"// func main creates a starting point for Go","lineNumber":20,"offsetAndLengths":[[3,9]]},{"preview":"\t// It only calls `func main` automatically.","lineNumber":39,"offsetAndLengths":[[19,9]]}]},{"__typename":"FileMatch","repository":{"name":"github.com/wx-chevalier/Go-Notes"},"file":{"path":"99_参考资料/2018_《Learn Go》/translation/chinese/02_编写第一个程序/main.go","url":"/r/github.com/wx-chevalier/Go-Notes/-/blob/99_%E5%8F%82%E8%80%83%E8%B5%84%E6%96%99/2018_%E3%80%8ALearn%20Go%E3%80%8B/translation/chinese/02_%E7%BC%96%E5%86%99%E7%AC%AC%E4%B8%80%E4%B8%AA%E7%A8%8B%E5%BA%8F/main.go","content":"// Copyright © 2018 Inanc Gumus\n// Learn Go Programming Course\n// License: https://creativecommons.org/licenses/by-nc-sa/4.0/\n//\n// For more tutorials : https://learngoprogramming.com\n// In-person training : https://www.linkedin.com/in/inancgumus/\n// Follow me on twitter: https://twitter.com/inancgumus\n\n// package main 是一个特殊的包\n// 它允许 Go 创建一个可执行文件\npackage main\n\n/*\n这是一个多行注释\n\nimport 关键字引入另一个包\n 对于这个 .go ”文件“\n\nimport \"fmt\" 允许你使用 fmt 包的各种函数\n 在这个文件\n*/\nimport \"fmt\"\n\n// \"func main\" 是特殊的\n//\n// Go 必须知道从哪里开始执行\n//\n// func main 对于 Go 而言就是开始执行的位置\n//\n// 代码编译后，\n// Go runtime 第一个执行该函数\nfunc main() {\n\t// import \"fmt\" 之后\n\t// \"fmt\" 包中的 Println 函数就可以使用\n\n\t// 在控制台输入下面的命令可以查看它的源代码:\n\t// go doc -src fmt Println\n\n\t// Println 是一个 exported 函数来自\n\t// \"fmt\" 包\n\n\t// Exported = 首字母大写\n\tfmt.Println(\"Hello Gopher!\")\n\n\t// Go 本身不能调用 Println 函数\n\t// 这就是为什么你需要在这里调用它\n\t// Go 只能自动地调用 `func main`\n\n\t// -----\n\n\t// Go 字符串支持 Unicode 字符\n\t// 源代码同时也支持: KÖSTEBEK!\n\t//\n\t// 因为: 字面值 ~= 源代码\n\n\t// 练习: 删除下行注释的 --\u003e //\n\t// fmt.Println(\"Merhaba Köstebek!\")\n\n\t// 不重要的注释:\n\t// \"Merhaba Köstebek\" 的意思是 \"你好 Gopher\"\n\t// 在土耳其语\n}\n"},"lineMatches":[{"preview":"// \"func main\" 是特殊的","lineNumber":23,"offsetAndLengths":[[3,11]]},{"preview":"func main() {","lineNumber":31,"offsetAndLengths":[[0,9]]},{"preview":"\t// Go 只能自动地调用 `func main`","lineNumber":46,"offsetAndLengths":[[16,9]]},{"preview":"// func main 对于 Go 而言就是开始执行的位置","lineNumber":27,"offsetAndLengths":[[3,9]]}]},{"__typename":"FileMatch","repository":{"name":"github.com/wx-chevalier/Go-Notes"},"file":{"path":"99_参考资料/2018_《Learn Go》/first/first/main.go","url":"/r/github.com/wx-chevalier/Go-Notes/-/blob/99_%E5%8F%82%E8%80%83%E8%B5%84%E6%96%99/2018_%E3%80%8ALearn%20Go%E3%80%8B/first/first/main.go","content":"// Copyright © 2018 Inanc Gumus\n// Learn Go Programming Course\n// License: https://creativecommons.org/licenses/by-nc-sa/4.0/\n//\n// For more tutorials : https://learngoprogramming.com\n// In-person training : https://www.linkedin.com/in/inancgumus/\n// Follow me on twitter: https://twitter.com/inancgumus\n\n// package main is a special package\n// it allows Go to create an executable file\npackage main\n\n/*\nThis is a multi-line comment.\n\nimport keyword makes another package available\n for this .go \"file\".\n\nimport \"fmt\" lets you access fmt package''s functionality\n here in this file.\n*/\nimport \"fmt\"\n\n// \"func main\" is special.\n//\n// Go has to know where to start\n//\n// func main creates a starting point for Go\n//\n// After compiling the code,\n// Go runtime will first run this function\nfunc main() {\n\t// after: import \"fmt\"\n\t// Println function of \"fmt\" package becomes available\n\n\t// Look at what it looks like by typing in the console:\n\t// go doc -src fmt Println\n\n\t// Println is just an exported function from\n\t// \"fmt\" package\n\n\t// Exported = First Letter is uppercase\n\tfmt.Println(\"Hello Gopher!\")\n\n\t// Go cannot call Println function by itself.\n\t// That''s why you need to call it here.\n\t// It only calls `func main` automatically.\n\n\t// -----\n\n\t// Go supports Unicode characters in string literals\n\t// And also in source-code: KÖSTEBEK!\n\n\t// EXERCISE: Remove the comments from below --\u003e //\n\t// fmt.Println(\"Merhaba Köstebek!\")\n\n\t// Unnecessary note:\n\t// \"Merhaba Köstebek\" means \"Hello Gopher\"\n\t// in Turkish language\n}\n"},"lineMatches":[{"preview":"// \"func main\" is special.","lineNumber":23,"offsetAndLengths":[[3,11]]},{"preview":"func main() {","lineNumber":31,"offsetAndLengths":[[0,9]]},{"preview":"// func main creates a starting point for Go","lineNumber":27,"offsetAndLengths":[[3,9]]},{"preview":"\t// It only calls `func main` automatically.","lineNumber":46,"offsetAndLengths":[[19,9]]}]},{"__typename":"FileMatch","repository":{"name":"github.com/wx-chevalier/Go-Notes"},"file":{"path":"99_参考资料/2018_《Learn Go》/translation/spanish/02_tu-primer-programa/main.go","url":"/r/github.com/wx-chevalier/Go-Notes/-/blob/99_%E5%8F%82%E8%80%83%E8%B5%84%E6%96%99/2018_%E3%80%8ALearn%20Go%E3%80%8B/translation/spanish/02_tu-primer-programa/main.go","content":"// Copyright © 2021 Inanc Gumus\n// Learn Go Programming Course\n// Licencia: https://creativecommons.org/licenses/by-nc-sa/4.0/\n//\n// Para mas tutoriales: https://learngoprogramming.com\n// Clases particulares: https://www.linkedin.com/in/inancgumus/\n// Sigueme en twitter: https://twitter.com/inancgumus\n\n// package main es un paquete especial\n// le permite a Go crear un archivo ejecutable\npackage main\n\n/*\nEsto es un comentario multilinea.\n\nla palabra import hace que otro paquete este disponible\n para este \"archivo\" .go.\n\nimport \"fmt\" te permite acceder a la funcionaliddad del paquete fmt\n en este archivo.\n*/\nimport \"fmt\"\n\n// \"func main\" es especial.\n//\n// Go tiene que saber por donde empezar\n//\n// func main crea un punto de inicio para Go\n//\n// Después de compilar el codigo,\n// Go ejecutará esta función primero.\nfunc main() {\n\t// después del: import \"fmt\"\n\t// La función Println del paquete \"fmt\" estará disponible\n\n\t// Lee sobre ella escribiendo los siguiente en la consola:\n\t// go doc -src fmt Println\n\n\t// Println es simplemente una función exportada de:\n\t// \"fmt\" package\n\n\t// Para poder exportar una función tendrás que escribir el\n\t// primer caracter del nombre de la función en mayúscula.\n\tfmt.Println(\"Hello Gopher!\")\n\n\t// Go no puede llamar a la función Println por si mismo.\n\t// Por eso la tienes que llamar aqui.\n\t// Solo llama a la `func main` de forma automatica.\n\n\t// -----\n\n\t// Go soporta caracteres unicode en cadenas de texto literal\n\t// y tambien en el codigo fuente: KÖSTEBEK!\n\t//\n\t// Porque: Literal ~= Codigo Fuente\n}\n"},"lineMatches":[{"preview":"// \"func main\" es especial.","lineNumber":23,"offsetAndLengths":[[3,11]]},{"preview":"func main() {","lineNumber":31,"offsetAndLengths":[[0,9]]},{"preview":"// func main crea un punto de inicio para Go","lineNumber":27,"offsetAndLengths":[[3,9]]},{"preview":"\t// Solo llama a la `func main` de forma automatica.","lineNumber":47,"offsetAndLengths":[[21,9]]}]},{"__typename":"FileMatch","repository":{"name":"github.com/wx-chevalier/Go-Notes"},"file":{"path":"99_参考资料/2018_《Learn Go》/02_write-your-first-program/main.go","url":"/r/github.com/wx-chevalier/Go-Notes/-/blob/99_%E5%8F%82%E8%80%83%E8%B5%84%E6%96%99/2018_%E3%80%8ALearn%20Go%E3%80%8B/02_write-your-first-program/main.go","content":"// Copyright © 2018 Inanc Gumus\n// Learn Go Programming Course\n// License: https://creativecommons.org/licenses/by-nc-sa/4.0/\n//\n// For more tutorials : https://learngoprogramming.com\n// In-person training : https://www.linkedin.com/in/inancgumus/\n// Follow me on twitter: https://twitter.com/inancgumus\n\n// package main is a special package\n// it allows Go to create an executable file\npackage main\n\n/*\nThis is a multi-line comment.\n\nimport keyword makes another package available\n for this .go \"file\".\n\nimport \"fmt\" lets you access fmt package''s functionality\n here in this file.\n*/\nimport \"fmt\"\n\n// \"func main\" is special.\n//\n// Go has to know where to start\n//\n// func main creates a starting point for Go\n//\n// After compiling the code,\n// Go runtime will first run this function\nfunc main() {\n\t// after: import \"fmt\"\n\t// Println function of \"fmt\" package becomes available\n\n\t// Look at what it looks like by typing in the console:\n\t// go doc -src fmt Println\n\n\t// Println is just an exported function from\n\t// \"fmt\" package\n\n\t// Exported = First Letter is uppercase\n\tfmt.Println(\"Hello Gopher!\")\n\n\t// Go cannot call Println function by itself.\n\t// That''s why you need to call it here.\n\t// It only calls `func main` automatically.\n\n\t// -----\n\n\t// Go supports Unicode characters in string literals\n\t// And also in source-code: KÖSTEBEK!\n\t//\n\t// Because: Literal ~= Source Code\n\n\t// EXERCISE: Remove the comments from below --\u003e //\n\t// fmt.Println(\"Merhaba Köstebek!\")\n\n\t// Unnecessary note:\n\t// \"Merhaba Köstebek\" means \"Hello Gopher\"\n\t// in Turkish language\n}\n"},"lineMatches":[{"preview":"// \"func main\" is special.","lineNumber":23,"offsetAndLengths":[[3,11]]},{"preview":"func main() {","lineNumber":31,"offsetAndLengths":[[0,9]]},{"preview":"// func main creates a starting point for Go","lineNumber":27,"offsetAndLengths":[[3,9]]},{"preview":"\t// It only calls `func main` automatically.","lineNumber":46,"offsetAndLengths":[[19,9]]}]},{"__typename":"FileMatch","repository":{"name":"github.com/ShixiangWang/self-study"},"file":{"path":"go/learngo/first/first/main.go","url":"/r/github.com/ShixiangWang/self-study/-/blob/go/learngo/first/first/main.go","content":"// Copyright © 2018 Inanc Gumus\n// Learn Go Programming Course\n// License: https://creativecommons.org/licenses/by-nc-sa/4.0/\n//\n// For more tutorials : https://learngoprogramming.com\n// In-person training : https://www.linkedin.com/in/inancgumus/\n// Follow me on twitter: https://twitter.com/inancgumus\n\n// package main is a special package\n// it allows Go to create an executable file\npackage main\n\n/*\nThis is a multi-line comment.\n\nimport keyword makes another package available\n for this .go \"file\".\n\nimport \"fmt\" lets you access fmt package''s functionality\n here in this file.\n*/\nimport \"fmt\"\n\n// \"func main\" is special.\n//\n// Go has to know where to start\n//\n// func main creates a starting point for Go\n//\n// After compiling the code,\n// Go runtime will first run this function\nfunc main() {\n\t// after: import \"fmt\"\n\t// Println function of \"fmt\" package becomes available\n\n\t// Look at what it looks like by typing in the console:\n\t// godoc -src fmt Println\n\n\t// Println is just an exported function from\n\t// \"fmt\" package\n\n\t// Exported = First Letter is uppercase\n\tfmt.Println(\"Hello Gopher!\")\n\n\t// Go cannot call Println function by itself.\n\t// That''s why you need to call it here.\n\t// It only calls `func main` automatically.\n\n\t// -----\n\n\t// Go supports Unicode characters in string literals\n\t// And also in source-code: KÖSTEBEK!\n\n\t// EXERCISE: Remove the comments from below --\u003e //\n\t// fmt.Println(\"Merhaba Köstebek!\")\n\n\t// Unnecessary note:\n\t// \"Merhaba Köstebek\" means \"Hello Gopher\"\n\t// in Turkish language\n}\n"},"lineMatches":[{"preview":"// \"func main\" is special.","lineNumber":23,"offsetAndLengths":[[3,11]]},{"preview":"func main() {","lineNumber":31,"offsetAndLengths":[[0,9]]},{"preview":"// func main creates a starting point for Go","lineNumber":27,"offsetAndLengths":[[3,9]]},{"preview":"\t// It only calls `func main` automatically.","lineNumber":46,"offsetAndLengths":[[19,9]]}]},{"__typename":"FileMatch","repository":{"name":"github.com/ShixiangWang/self-study"},"file":{"path":"go/learngo/02-write-your-first-program/main.go","url":"/r/github.com/ShixiangWang/self-study/-/blob/go/learngo/02-write-your-first-program/main.go","content":"// Copyright © 2018 Inanc Gumus\n// Learn Go Programming Course\n// License: https://creativecommons.org/licenses/by-nc-sa/4.0/\n//\n// For more tutorials : https://learngoprogramming.com\n// In-person training : https://www.linkedin.com/in/inancgumus/\n// Follow me on twitter: https://twitter.com/inancgumus\n\n// package main is a special package\n// it allows Go to create an executable file\npackage main\n\n/*\nThis is a multi-line comment.\n\nimport keyword makes another package available\n for this .go \"file\".\n\nimport \"fmt\" lets you access fmt package''s functionality\n here in this file.\n*/\nimport \"fmt\"\n\n// \"func main\" is special.\n//\n// Go has to know where to start\n//\n// func main creates a starting point for Go\n//\n// After compiling the code,\n// Go runtime will first run this function\nfunc main() {\n\t// after: import \"fmt\"\n\t// Println function of \"fmt\" package becomes available\n\n\t// Look at what it looks like by typing in the console:\n\t// go doc -src fmt Println\n\n\t// Println is just an exported function from\n\t// \"fmt\" package\n\n\t// Exported = First Letter is uppercase\n\tfmt.Println(\"Hello Gopher!\")\n\n\t// Go cannot call Println function by itself.\n\t// That''s why you need to call it here.\n\t// It only calls `func main` automatically.\n\n\t// -----\n\n\t// Go supports Unicode characters in string literals\n\t// And also in source-code: KÖSTEBEK!\n\t//\n\t// Because: Literal ~= Source Code\n\n\t// EXERCISE: Remove the comments from below --\u003e //\n\t// fmt.Println(\"Merhaba Köstebek!\")\n\n\t// Unnecessary note:\n\t// \"Merhaba Köstebek\" means \"Hello Gopher\"\n\t// in Turkish language\n}\n"},"lineMatches":[{"preview":"// \"func main\" is special.","lineNumber":23,"offsetAndLengths":[[3,11]]},{"preview":"func main() {","lineNumber":31,"offsetAndLengths":[[0,9]]}]}]}}}}'
- headers:
- Content-Type:
- - application/json
- status: 200 OK
- code: 200
- duration: 1.339439458s
-- id: 3
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 43532
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/sourcegraph_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use sourcegraph to search for 'func main' in Go repositories\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_Fmhtvwj7YgCbE2050w62UqIA\",\"function\":{\"arguments\":\"{\\\"query\\\":\\\"lang:go \\\\\\\"func main\\\\\\\"\\\", \\\"count\\\":20, \\\"context_window\\\":5, \\\"timeout\\\":120}\",\"name\":\"sourcegraph\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"# Sourcegraph Search Results\\n\\nFound 30 matches across 30 results\\n(Result limit reached, try a more specific query)\\n\\n## Result 1: github.com/wx-chevalier/Go-Notes/99_参考资料/2017_A Huge Number of Go Examples/codes/first/first/main.go\\n\\nURL: /r/github.com/wx-chevalier/Go-Notes/-/blob/99_%E5%8F%82%E8%80%83%E8%B5%84%E6%96%99/2017_A%20Huge%20Number%20of%20Go%20Examples/codes/first/first/main.go\\n\\n```\\n11| \\n12| import \\\"fmt\\\" lets you access fmt package's functionality\\n13| here in this file.\\n14| */\\n15| import \\\"fmt\\\"\\n16| // \\\"func main\\\" is special.\\n17| // \\\"func main\\\" is special.\\n18| //\\n19| // Go has to know where to start\\n20| //\\n21| // func main creates a starting point for Go\\n```\\n\\n```\\n19| // Go has to know where to start\\n20| //\\n21| // func main creates a starting point for Go\\n22| //\\n23| // After compiling the code,\\n24| func main() {\\n25| func main() {\\n26| \\t// after: import \\\"fmt\\\"\\n27| \\t// Println function of \\\"fmt\\\" package becomes available\\n28| \\n29| \\t// Look at what it looks like by typing in the console:\\n```\\n\\n```\\n15| import \\\"fmt\\\"\\n16| \\n17| // \\\"func main\\\" is special.\\n18| //\\n19| // Go has to know where to start\\n20| // func main creates a starting point for Go\\n21| // func main creates a starting point for Go\\n22| //\\n23| // After compiling the code,\\n24| // Go runtime will first run this function\\n25| func main() {\\n```\\n\\n```\\n34| \\n35| \\t// Exported = First Letter is uppercase\\n36| \\tfmt.Println(\\\"Hello Gopher!\\\")\\n37| \\n38| \\t// Go cannot call Println function by itself.\\n39| \\t// It only calls `func main` automatically.\\n40| \\t// It only calls `func main` automatically.\\n41| \\n42| \\t// -----\\n43| \\n44| \\t// Go supports Unicode characters in string literals\\n```\\n\\n## Result 2: github.com/wx-chevalier/Go-Notes/99_参考资料/2017_A Huge Number of Go Examples/codes/02_write-your-first-program/main.go\\n\\nURL: /r/github.com/wx-chevalier/Go-Notes/-/blob/99_%E5%8F%82%E8%80%83%E8%B5%84%E6%96%99/2017_A%20Huge%20Number%20of%20Go%20Examples/codes/02_write-your-first-program/main.go\\n\\n```\\n11| \\n12| import \\\"fmt\\\" lets you access fmt package's functionality\\n13| here in this file.\\n14| */\\n15| import \\\"fmt\\\"\\n16| // \\\"func main\\\" is special.\\n17| // \\\"func main\\\" is special.\\n18| //\\n19| // Go has to know where to start\\n20| //\\n21| // func main creates a starting point for Go\\n```\\n\\n```\\n19| // Go has to know where to start\\n20| //\\n21| // func main creates a starting point for Go\\n22| //\\n23| // After compiling the code,\\n24| func main() {\\n25| func main() {\\n26| \\t// after: import \\\"fmt\\\"\\n27| \\t// Println function of \\\"fmt\\\" package becomes available\\n28| \\n29| \\t// Look at what it looks like by typing in the console:\\n```\\n\\n```\\n15| import \\\"fmt\\\"\\n16| \\n17| // \\\"func main\\\" is special.\\n18| //\\n19| // Go has to know where to start\\n20| // func main creates a starting point for Go\\n21| // func main creates a starting point for Go\\n22| //\\n23| // After compiling the code,\\n24| // Go runtime will first run this function\\n25| func main() {\\n```\\n\\n```\\n34| \\n35| \\t// Exported = First Letter is uppercase\\n36| \\tfmt.Println(\\\"Hello Gopher!\\\")\\n37| \\n38| \\t// Go cannot call Println function by itself.\\n39| \\t// It only calls `func main` automatically.\\n40| \\t// It only calls `func main` automatically.\\n41| \\n42| \\t// -----\\n43| \\n44| \\t// Go supports Unicode characters in string literals\\n```\\n\\n## Result 3: github.com/wx-chevalier/Go-Notes/99_参考资料/2018_《Learn Go》/translation/chinese/02_编写第一个程序/main.go\\n\\nURL: /r/github.com/wx-chevalier/Go-Notes/-/blob/99_%E5%8F%82%E8%80%83%E8%B5%84%E6%96%99/2018_%E3%80%8ALearn%20Go%E3%80%8B/translation/chinese/02_%E7%BC%96%E5%86%99%E7%AC%AC%E4%B8%80%E4%B8%AA%E7%A8%8B%E5%BA%8F/main.go\\n\\n```\\n18| \\n19| import \\\"fmt\\\" 允许你使用 fmt 包的各种函数\\n20| 在这个文件\\n21| */\\n22| import \\\"fmt\\\"\\n23| // \\\"func main\\\" 是特殊的\\n24| // \\\"func main\\\" 是特殊的\\n25| //\\n26| // Go 必须知道从哪里开始执行\\n27| //\\n28| // func main 对于 Go 而言就是开始执行的位置\\n```\\n\\n```\\n26| // Go 必须知道从哪里开始执行\\n27| //\\n28| // func main 对于 Go 而言就是开始执行的位置\\n29| //\\n30| // 代码编译后，\\n31| func main() {\\n32| func main() {\\n33| \\t// import \\\"fmt\\\" 之后\\n34| \\t// \\\"fmt\\\" 包中的 Println 函数就可以使用\\n35| \\n36| \\t// 在控制台输入下面的命令可以查看它的源代码:\\n```\\n\\n```\\n41| \\n42| \\t// Exported = 首字母大写\\n43| \\tfmt.Println(\\\"Hello Gopher!\\\")\\n44| \\n45| \\t// Go 本身不能调用 Println 函数\\n46| \\t// Go 只能自动地调用 `func main`\\n47| \\t// Go 只能自动地调用 `func main`\\n48| \\n49| \\t// -----\\n50| \\n51| \\t// Go 字符串支持 Unicode 字符\\n```\\n\\n```\\n22| import \\\"fmt\\\"\\n23| \\n24| // \\\"func main\\\" 是特殊的\\n25| //\\n26| // Go 必须知道从哪里开始执行\\n27| // func main 对于 Go 而言就是开始执行的位置\\n28| // func main 对于 Go 而言就是开始执行的位置\\n29| //\\n30| // 代码编译后，\\n31| // Go runtime 第一个执行该函数\\n32| func main() {\\n```\\n\\n## Result 4: github.com/wx-chevalier/Go-Notes/99_参考资料/2018_《Learn Go》/first/first/main.go\\n\\nURL: /r/github.com/wx-chevalier/Go-Notes/-/blob/99_%E5%8F%82%E8%80%83%E8%B5%84%E6%96%99/2018_%E3%80%8ALearn%20Go%E3%80%8B/first/first/main.go\\n\\n```\\n18| \\n19| import \\\"fmt\\\" lets you access fmt package's functionality\\n20| here in this file.\\n21| */\\n22| import \\\"fmt\\\"\\n23| // \\\"func main\\\" is special.\\n24| // \\\"func main\\\" is special.\\n25| //\\n26| // Go has to know where to start\\n27| //\\n28| // func main creates a starting point for Go\\n```\\n\\n```\\n26| // Go has to know where to start\\n27| //\\n28| // func main creates a starting point for Go\\n29| //\\n30| // After compiling the code,\\n31| func main() {\\n32| func main() {\\n33| \\t// after: import \\\"fmt\\\"\\n34| \\t// Println function of \\\"fmt\\\" package becomes available\\n35| \\n36| \\t// Look at what it looks like by typing in the console:\\n```\\n\\n```\\n22| import \\\"fmt\\\"\\n23| \\n24| // \\\"func main\\\" is special.\\n25| //\\n26| // Go has to know where to start\\n27| // func main creates a starting point for Go\\n28| // func main creates a starting point for Go\\n29| //\\n30| // After compiling the code,\\n31| // Go runtime will first run this function\\n32| func main() {\\n```\\n\\n```\\n41| \\n42| \\t// Exported = First Letter is uppercase\\n43| \\tfmt.Println(\\\"Hello Gopher!\\\")\\n44| \\n45| \\t// Go cannot call Println function by itself.\\n46| \\t// It only calls `func main` automatically.\\n47| \\t// It only calls `func main` automatically.\\n48| \\n49| \\t// -----\\n50| \\n51| \\t// Go supports Unicode characters in string literals\\n```\\n\\n## Result 5: github.com/wx-chevalier/Go-Notes/99_参考资料/2018_《Learn Go》/translation/spanish/02_tu-primer-programa/main.go\\n\\nURL: /r/github.com/wx-chevalier/Go-Notes/-/blob/99_%E5%8F%82%E8%80%83%E8%B5%84%E6%96%99/2018_%E3%80%8ALearn%20Go%E3%80%8B/translation/spanish/02_tu-primer-programa/main.go\\n\\n```\\n18| \\n19| import \\\"fmt\\\" te permite acceder a la funcionaliddad del paquete fmt\\n20| en este archivo.\\n21| */\\n22| import \\\"fmt\\\"\\n23| // \\\"func main\\\" es especial.\\n24| // \\\"func main\\\" es especial.\\n25| //\\n26| // Go tiene que saber por donde empezar\\n27| //\\n28| // func main crea un punto de inicio para Go\\n```\\n\\n```\\n26| // Go tiene que saber por donde empezar\\n27| //\\n28| // func main crea un punto de inicio para Go\\n29| //\\n30| // Después de compilar el codigo,\\n31| func main() {\\n32| func main() {\\n33| \\t// después del: import \\\"fmt\\\"\\n34| \\t// La función Println del paquete \\\"fmt\\\" estará disponible\\n35| \\n36| \\t// Lee sobre ella escribiendo los siguiente en la consola:\\n```\\n\\n```\\n22| import \\\"fmt\\\"\\n23| \\n24| // \\\"func main\\\" es especial.\\n25| //\\n26| // Go tiene que saber por donde empezar\\n27| // func main crea un punto de inicio para Go\\n28| // func main crea un punto de inicio para Go\\n29| //\\n30| // Después de compilar el codigo,\\n31| // Go ejecutará esta función primero.\\n32| func main() {\\n```\\n\\n```\\n42| \\t// Para poder exportar una función tendrás que escribir el\\n43| \\t// primer caracter del nombre de la función en mayúscula.\\n44| \\tfmt.Println(\\\"Hello Gopher!\\\")\\n45| \\n46| \\t// Go no puede llamar a la función Println por si mismo.\\n47| \\t// Solo llama a la `func main` de forma automatica.\\n48| \\t// Solo llama a la `func main` de forma automatica.\\n49| \\n50| \\t// -----\\n51| \\n52| \\t// Go soporta caracteres unicode en cadenas de texto literal\\n```\\n\\n## Result 6: github.com/wx-chevalier/Go-Notes/99_参考资料/2018_《Learn Go》/02_write-your-first-program/main.go\\n\\nURL: /r/github.com/wx-chevalier/Go-Notes/-/blob/99_%E5%8F%82%E8%80%83%E8%B5%84%E6%96%99/2018_%E3%80%8ALearn%20Go%E3%80%8B/02_write-your-first-program/main.go\\n\\n```\\n18| \\n19| import \\\"fmt\\\" lets you access fmt package's functionality\\n20| here in this file.\\n21| */\\n22| import \\\"fmt\\\"\\n23| // \\\"func main\\\" is special.\\n24| // \\\"func main\\\" is special.\\n25| //\\n26| // Go has to know where to start\\n27| //\\n28| // func main creates a starting point for Go\\n```\\n\\n```\\n26| // Go has to know where to start\\n27| //\\n28| // func main creates a starting point for Go\\n29| //\\n30| // After compiling the code,\\n31| func main() {\\n32| func main() {\\n33| \\t// after: import \\\"fmt\\\"\\n34| \\t// Println function of \\\"fmt\\\" package becomes available\\n35| \\n36| \\t// Look at what it looks like by typing in the console:\\n```\\n\\n```\\n22| import \\\"fmt\\\"\\n23| \\n24| // \\\"func main\\\" is special.\\n25| //\\n26| // Go has to know where to start\\n27| // func main creates a starting point for Go\\n28| // func main creates a starting point for Go\\n29| //\\n30| // After compiling the code,\\n31| // Go runtime will first run this function\\n32| func main() {\\n```\\n\\n```\\n41| \\n42| \\t// Exported = First Letter is uppercase\\n43| \\tfmt.Println(\\\"Hello Gopher!\\\")\\n44| \\n45| \\t// Go cannot call Println function by itself.\\n46| \\t// It only calls `func main` automatically.\\n47| \\t// It only calls `func main` automatically.\\n48| \\n49| \\t// -----\\n50| \\n51| \\t// Go supports Unicode characters in string literals\\n```\\n\\n## Result 7: github.com/ShixiangWang/self-study/go/learngo/first/first/main.go\\n\\nURL: /r/github.com/ShixiangWang/self-study/-/blob/go/learngo/first/first/main.go\\n\\n```\\n18| \\n19| import \\\"fmt\\\" lets you access fmt package's functionality\\n20| here in this file.\\n21| */\\n22| import \\\"fmt\\\"\\n23| // \\\"func main\\\" is special.\\n24| // \\\"func main\\\" is special.\\n25| //\\n26| // Go has to know where to start\\n27| //\\n28| // func main creates a starting point for Go\\n```\\n\\n```\\n26| // Go has to know where to start\\n27| //\\n28| // func main creates a starting point for Go\\n29| //\\n30| // After compiling the code,\\n31| func main() {\\n32| func main() {\\n33| \\t// after: import \\\"fmt\\\"\\n34| \\t// Println function of \\\"fmt\\\" package becomes available\\n35| \\n36| \\t// Look at what it looks like by typing in the console:\\n```\\n\\n```\\n22| import \\\"fmt\\\"\\n23| \\n24| // \\\"func main\\\" is special.\\n25| //\\n26| // Go has to know where to start\\n27| // func main creates a starting point for Go\\n28| // func main creates a starting point for Go\\n29| //\\n30| // After compiling the code,\\n31| // Go runtime will first run this function\\n32| func main() {\\n```\\n\\n```\\n41| \\n42| \\t// Exported = First Letter is uppercase\\n43| \\tfmt.Println(\\\"Hello Gopher!\\\")\\n44| \\n45| \\t// Go cannot call Println function by itself.\\n46| \\t// It only calls `func main` automatically.\\n47| \\t// It only calls `func main` automatically.\\n48| \\n49| \\t// -----\\n50| \\n51| \\t// Go supports Unicode characters in string literals\\n```\\n\\n## Result 8: github.com/ShixiangWang/self-study/go/learngo/02-write-your-first-program/main.go\\n\\nURL: /r/github.com/ShixiangWang/self-study/-/blob/go/learngo/02-write-your-first-program/main.go\\n\\n```\\n18| \\n19| import \\\"fmt\\\" lets you access fmt package's functionality\\n20| here in this file.\\n21| */\\n22| import \\\"fmt\\\"\\n23| // \\\"func main\\\" is special.\\n24| // \\\"func main\\\" is special.\\n25| //\\n26| // Go has to know where to start\\n27| //\\n28| // func main creates a starting point for Go\\n```\\n\\n```\\n26| // Go has to know where to start\\n27| //\\n28| // func main creates a starting point for Go\\n29| //\\n30| // After compiling the code,\\n31| func main() {\\n32| func main() {\\n33| \\t// after: import \\\"fmt\\\"\\n34| \\t// Println function of \\\"fmt\\\" package becomes available\\n35| \\n36| \\t// Look at what it looks like by typing in the console:\\n```\\n\\n\",\"tool_call_id\":\"call_Fmhtvwj7YgCbE2050w62UqIA\",\"role\":\"tool\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":"","refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"OmgBmxmqKJ"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"30"},"finish_reason":null}],"usage":null,"obfuscation":"W1eLVwSAF6"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":" results"},"finish_reason":null}],"usage":null,"obfuscation":"FeR9"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":" ("},"finish_reason":null}],"usage":null,"obfuscation":"yvmHNo5ySd"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"limit"},"finish_reason":null}],"usage":null,"obfuscation":"mm9DVgQ"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":" reached"},"finish_reason":null}],"usage":null,"obfuscation":"ybg6"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":")."},"finish_reason":null}],"usage":null,"obfuscation":"MyPGOMOmTh"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":" Examples"},"finish_reason":null}],"usage":null,"obfuscation":"Pfz"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":":\n"},"finish_reason":null}],"usage":null,"obfuscation":"gmiWw3L7K"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"-"},"finish_reason":null}],"usage":null,"obfuscation":"AeSmMl60rY5"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":" github"},"finish_reason":null}],"usage":null,"obfuscation":"62mdh"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":".com"},"finish_reason":null}],"usage":null,"obfuscation":"f2DowdXt"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/w"},"finish_reason":null}],"usage":null,"obfuscation":"46x7eSEOAg"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"x"},"finish_reason":null}],"usage":null,"obfuscation":"3ZQgdMhMLJa"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"-che"},"finish_reason":null}],"usage":null,"obfuscation":"iEkQGEWZ"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"val"},"finish_reason":null}],"usage":null,"obfuscation":"jOHSabxVw"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"ier"},"finish_reason":null}],"usage":null,"obfuscation":"tdAsCDreh"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/"},"finish_reason":null}],"usage":null,"obfuscation":"wpsHizOr07h"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"Go"},"finish_reason":null}],"usage":null,"obfuscation":"zXw9VixZiW"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"-"},"finish_reason":null}],"usage":null,"obfuscation":"th78iKLCCZi"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"Notes"},"finish_reason":null}],"usage":null,"obfuscation":"poOIFyk"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/..."},"finish_reason":null}],"usage":null,"obfuscation":"kxc0dWPi"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/"},"finish_reason":null}],"usage":null,"obfuscation":"uVVABA9cbaq"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"codes"},"finish_reason":null}],"usage":null,"obfuscation":"QA1JUxE"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/"},"finish_reason":null}],"usage":null,"obfuscation":"c1RX0h3QpLY"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"first"},"finish_reason":null}],"usage":null,"obfuscation":"hhT0hmr"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/"},"finish_reason":null}],"usage":null,"obfuscation":"87yohuhhOd1"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"first"},"finish_reason":null}],"usage":null,"obfuscation":"eskoB8R"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/main"},"finish_reason":null}],"usage":null,"obfuscation":"vdA8mh6"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":".go"},"finish_reason":null}],"usage":null,"obfuscation":"1xRAPjf6d"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"\n"},"finish_reason":null}],"usage":null,"obfuscation":"FoMYYOYMSi"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"-"},"finish_reason":null}],"usage":null,"obfuscation":"nlsOA3eJIzt"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":" github"},"finish_reason":null}],"usage":null,"obfuscation":"xu1bf"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":".com"},"finish_reason":null}],"usage":null,"obfuscation":"tACCeBwh"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/w"},"finish_reason":null}],"usage":null,"obfuscation":"QQNrafb9id"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"x"},"finish_reason":null}],"usage":null,"obfuscation":"9IJm5fAZUeB"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"-che"},"finish_reason":null}],"usage":null,"obfuscation":"jRCfxM1C"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"val"},"finish_reason":null}],"usage":null,"obfuscation":"fRXtS7JYu"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"ier"},"finish_reason":null}],"usage":null,"obfuscation":"wlmELn1pm"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/"},"finish_reason":null}],"usage":null,"obfuscation":"nh34tgOhnib"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"Go"},"finish_reason":null}],"usage":null,"obfuscation":"Wg5AB65b2u"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"-"},"finish_reason":null}],"usage":null,"obfuscation":"eIRfR9Xp79G"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"Notes"},"finish_reason":null}],"usage":null,"obfuscation":"xonFsly"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/..."},"finish_reason":null}],"usage":null,"obfuscation":"P07BOMLZ"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/"},"finish_reason":null}],"usage":null,"obfuscation":"CUiDkpTLYeN"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"02"},"finish_reason":null}],"usage":null,"obfuscation":"uBinJoD2yv"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"_write"},"finish_reason":null}],"usage":null,"obfuscation":"MQjkNf"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"-your"},"finish_reason":null}],"usage":null,"obfuscation":"ZcBv3Lb"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"-first"},"finish_reason":null}],"usage":null,"obfuscation":"UeBoLM"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"-program"},"finish_reason":null}],"usage":null,"obfuscation":"DZkl"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/main"},"finish_reason":null}],"usage":null,"obfuscation":"n2RLdMY"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":".go"},"finish_reason":null}],"usage":null,"obfuscation":"of3CHwC96"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"\n"},"finish_reason":null}],"usage":null,"obfuscation":"yfENFkrjCD"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"-"},"finish_reason":null}],"usage":null,"obfuscation":"qj3Rdtf8equ"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":" github"},"finish_reason":null}],"usage":null,"obfuscation":"xdBOa"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":".com"},"finish_reason":null}],"usage":null,"obfuscation":"LcfSgNN3"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/w"},"finish_reason":null}],"usage":null,"obfuscation":"zT5ArYCXkH"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"x"},"finish_reason":null}],"usage":null,"obfuscation":"L3EakKStMXl"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"-che"},"finish_reason":null}],"usage":null,"obfuscation":"Y5d841uF"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"val"},"finish_reason":null}],"usage":null,"obfuscation":"5bJ8jh5xe"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"ier"},"finish_reason":null}],"usage":null,"obfuscation":"r6xzrsoxQ"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/"},"finish_reason":null}],"usage":null,"obfuscation":"e6H8gqLKkqX"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"Go"},"finish_reason":null}],"usage":null,"obfuscation":"HXFT5Jv1vg"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"-"},"finish_reason":null}],"usage":null,"obfuscation":"nmcqtRHFTNc"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"Notes"},"finish_reason":null}],"usage":null,"obfuscation":"njo9G2X"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/..."},"finish_reason":null}],"usage":null,"obfuscation":"FvpWIcW6"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/"},"finish_reason":null}],"usage":null,"obfuscation":"WDclFhjAb4D"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"translation"},"finish_reason":null}],"usage":null,"obfuscation":"b"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/ch"},"finish_reason":null}],"usage":null,"obfuscation":"qRlicCRGK"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"inese"},"finish_reason":null}],"usage":null,"obfuscation":"z4Ij9Xx"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/"},"finish_reason":null}],"usage":null,"obfuscation":"paNUAl0IqX8"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"02"},"finish_reason":null}],"usage":null,"obfuscation":"I9WACvH1PB"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"_"},"finish_reason":null}],"usage":null,"obfuscation":"Sfhlm1e9R4y"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"编"},"finish_reason":null}],"usage":null,"obfuscation":"NLlW9H1aBCX"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"写"},"finish_reason":null}],"usage":null,"obfuscation":"MqMqrun3BuX"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"第"},"finish_reason":null}],"usage":null,"obfuscation":"dRiNZSg6y8K"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"一个"},"finish_reason":null}],"usage":null,"obfuscation":"veu5kts2LR"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"程序"},"finish_reason":null}],"usage":null,"obfuscation":"aNDeRuknOF"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/main"},"finish_reason":null}],"usage":null,"obfuscation":"opjJSHT"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":".go"},"finish_reason":null}],"usage":null,"obfuscation":"JKSvHZGRS"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"\n"},"finish_reason":null}],"usage":null,"obfuscation":"vqVliBizS9"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"-"},"finish_reason":null}],"usage":null,"obfuscation":"RQZaAUFtgpf"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":" github"},"finish_reason":null}],"usage":null,"obfuscation":"SUigF"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":".com"},"finish_reason":null}],"usage":null,"obfuscation":"wffb89XB"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/Sh"},"finish_reason":null}],"usage":null,"obfuscation":"jMqAuwbXo"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"ixi"},"finish_reason":null}],"usage":null,"obfuscation":"tqIgciOCx"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"ang"},"finish_reason":null}],"usage":null,"obfuscation":"AwdfdV4Hc"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"W"},"finish_reason":null}],"usage":null,"obfuscation":"lemVEbpziCw"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"ang"},"finish_reason":null}],"usage":null,"obfuscation":"eGayqdDM8"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/self"},"finish_reason":null}],"usage":null,"obfuscation":"Thho1TB"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"-study"},"finish_reason":null}],"usage":null,"obfuscation":"zmIJ9o"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/..."},"finish_reason":null}],"usage":null,"obfuscation":"7DnUanmO"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/"},"finish_reason":null}],"usage":null,"obfuscation":"uyNcIaCpzD8"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"first"},"finish_reason":null}],"usage":null,"obfuscation":"btSWFWQ"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/"},"finish_reason":null}],"usage":null,"obfuscation":"1rIfb7DVG5L"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"first"},"finish_reason":null}],"usage":null,"obfuscation":"uuc3Nk6"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"/main"},"finish_reason":null}],"usage":null,"obfuscation":"UGPJB9L"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":".go"},"finish_reason":null}],"usage":null,"obfuscation":"fcHk4idMp"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":null,"obfuscation":"TpGlFx"}
-
- data: {"id":"chatcmpl-DUu8ZglolyiwRhWMewilv3FgClkUT","object":"chat.completion.chunk","created":1776258003,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":10237,"completion_tokens":618,"total_tokens":10855,"prompt_tokens_details":{"cached_tokens":6272,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":512,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"8Oxxfo2Jwc6nX1V"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 11.872341625s
diff --git a/internal/agent/testdata/TestCoderAgent/openai-gpt-5/update_a_file.yaml b/internal/agent/testdata/TestCoderAgent/openai-gpt-5/update_a_file.yaml
deleted file mode 100644
index 25ed9c065ade014becf605e71bf021289b31e1d5..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/openai-gpt-5/update_a_file.yaml
+++ /dev/null
@@ -1,553 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 811
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nupdate the main.go file by changing the print to say hello from crush\n \n\n","role":"user"}],"model":"gpt-4o","max_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu6kvaBnw75LTfXpSDqulFyMN2rq","object":"chat.completion.chunk","created":1776257890,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"role":"assistant","content":"","refusal":null},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"vxS8DuIry7SlEa"}
-
- data: {"id":"chatcmpl-DUu6kvaBnw75LTfXpSDqulFyMN2rq","object":"chat.completion.chunk","created":1776257890,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":"Update"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"Qbznj4V5xk"}
-
- data: {"id":"chatcmpl-DUu6kvaBnw75LTfXpSDqulFyMN2rq","object":"chat.completion.chunk","created":1776257890,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" print"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"wNiaID5A6p"}
-
- data: {"id":"chatcmpl-DUu6kvaBnw75LTfXpSDqulFyMN2rq","object":"chat.completion.chunk","created":1776257890,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" message"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"qG3mkG7V"}
-
- data: {"id":"chatcmpl-DUu6kvaBnw75LTfXpSDqulFyMN2rq","object":"chat.completion.chunk","created":1776257890,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" in"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"y74AD9fsadJFa"}
-
- data: {"id":"chatcmpl-DUu6kvaBnw75LTfXpSDqulFyMN2rq","object":"chat.completion.chunk","created":1776257890,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" main"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"vgW1Y1epwFK"}
-
- data: {"id":"chatcmpl-DUu6kvaBnw75LTfXpSDqulFyMN2rq","object":"chat.completion.chunk","created":1776257890,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":".go"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"1NSzZyNShRz1a"}
-
- data: {"id":"chatcmpl-DUu6kvaBnw75LTfXpSDqulFyMN2rq","object":"chat.completion.chunk","created":1776257890,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" to"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"ptGkueYFA1B2o"}
-
- data: {"id":"chatcmpl-DUu6kvaBnw75LTfXpSDqulFyMN2rq","object":"chat.completion.chunk","created":1776257890,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" \""},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"txuhK9Uh7GcD7"}
-
- data: {"id":"chatcmpl-DUu6kvaBnw75LTfXpSDqulFyMN2rq","object":"chat.completion.chunk","created":1776257890,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":"hello"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"pD7Yo0C1wzV"}
-
- data: {"id":"chatcmpl-DUu6kvaBnw75LTfXpSDqulFyMN2rq","object":"chat.completion.chunk","created":1776257890,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" from"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"zpgxh1k4fQF"}
-
- data: {"id":"chatcmpl-DUu6kvaBnw75LTfXpSDqulFyMN2rq","object":"chat.completion.chunk","created":1776257890,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":" crush"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"3jWMQjThdP"}
-
- data: {"id":"chatcmpl-DUu6kvaBnw75LTfXpSDqulFyMN2rq","object":"chat.completion.chunk","created":1776257890,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{"content":"\""},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"LbDPFqNW5l6W1w"}
-
- data: {"id":"chatcmpl-DUu6kvaBnw75LTfXpSDqulFyMN2rq","object":"chat.completion.chunk","created":1776257890,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[{"index":0,"delta":{},"logprobs":null,"finish_reason":"stop"}],"usage":null,"obfuscation":"4PaKFljQoZ"}
-
- data: {"id":"chatcmpl-DUu6kvaBnw75LTfXpSDqulFyMN2rq","object":"chat.completion.chunk","created":1776257890,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_2608ccf94f","choices":[],"usage":{"prompt_tokens":155,"completion_tokens":12,"total_tokens":167,"prompt_tokens_details":{"cached_tokens":0,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":0,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"ceuoCJksB71I1"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 1.119495167s
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30682
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/update_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"update the main.go file by changing the print to say hello from crush\",\"role\":\"user\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu6kWAUWCaAILhzI2C72eEmkUIyK","object":"chat.completion.chunk","created":1776257890,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":null,"tool_calls":[{"index":0,"id":"call_fVC8ZP9c6nI5ZmyOkzi3PKR0","type":"function","function":{"name":"ls","arguments":""}}],"refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"7NigTXIF3hWx6i"}
-
- data: {"id":"chatcmpl-DUu6kWAUWCaAILhzI2C72eEmkUIyK","object":"chat.completion.chunk","created":1776257890,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"2IaMrcUjMoJ6rRf"}
-
- data: {"id":"chatcmpl-DUu6kWAUWCaAILhzI2C72eEmkUIyK","object":"chat.completion.chunk","created":1776257890,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"path"}}]},"finish_reason":null}],"usage":null,"obfuscation":"CeCI0eXyoZF8TK"}
-
- data: {"id":"chatcmpl-DUu6kWAUWCaAILhzI2C72eEmkUIyK","object":"chat.completion.chunk","created":1776257890,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"OMAP6UBfoL7I7ed"}
-
- data: {"id":"chatcmpl-DUu6kWAUWCaAILhzI2C72eEmkUIyK","object":"chat.completion.chunk","created":1776257890,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"/"}}]},"finish_reason":null}],"usage":null,"obfuscation":"6ZxIRlJck1EOL0N"}
-
- data: {"id":"chatcmpl-DUu6kWAUWCaAILhzI2C72eEmkUIyK","object":"chat.completion.chunk","created":1776257890,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"tmp"}}]},"finish_reason":null}],"usage":null,"obfuscation":"w117deHvYIeFfuL"}
-
- data: {"id":"chatcmpl-DUu6kWAUWCaAILhzI2C72eEmkUIyK","object":"chat.completion.chunk","created":1776257890,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/cr"}}]},"finish_reason":null}],"usage":null,"obfuscation":"jHTFibFr6Pdjvtk"}
-
- data: {"id":"chatcmpl-DUu6kWAUWCaAILhzI2C72eEmkUIyK","object":"chat.completion.chunk","created":1776257890,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ush"}}]},"finish_reason":null}],"usage":null,"obfuscation":"bbov95HiVVNkTZa"}
-
- data: {"id":"chatcmpl-DUu6kWAUWCaAILhzI2C72eEmkUIyK","object":"chat.completion.chunk","created":1776257890,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"jGTyLOUATcHwH"}
-
- data: {"id":"chatcmpl-DUu6kWAUWCaAILhzI2C72eEmkUIyK","object":"chat.completion.chunk","created":1776257890,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/Test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"Gtc1uEmkwbuTB"}
-
- data: {"id":"chatcmpl-DUu6kWAUWCaAILhzI2C72eEmkUIyK","object":"chat.completion.chunk","created":1776257890,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Coder"}}]},"finish_reason":null}],"usage":null,"obfuscation":"6qryTyaITnv7N"}
-
- data: {"id":"chatcmpl-DUu6kWAUWCaAILhzI2C72eEmkUIyK","object":"chat.completion.chunk","created":1776257890,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Agent"}}]},"finish_reason":null}],"usage":null,"obfuscation":"szmDo8qV3kwKQ"}
-
- data: {"id":"chatcmpl-DUu6kWAUWCaAILhzI2C72eEmkUIyK","object":"chat.completion.chunk","created":1776257890,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/open"}}]},"finish_reason":null}],"usage":null,"obfuscation":"1DhlsyYKbnt8Z"}
-
- data: {"id":"chatcmpl-DUu6kWAUWCaAILhzI2C72eEmkUIyK","object":"chat.completion.chunk","created":1776257890,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ai"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu6kWAUWCaAILhzI2C72eEmkUIyK","object":"chat.completion.chunk","created":1776257890,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-g"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu6kWAUWCaAILhzI2C72eEmkUIyK","object":"chat.completion.chunk","created":1776257890,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"pt"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu6kWAUWCaAILhzI2C72eEmkUIyK","object":"chat.completion.chunk","created":1776257890,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-"}}]},"finish_reason":null}],"usage":null,"obfuscation":"9"}
-
- data: {"id":"chatcmpl-DUu6kWAUWCaAILhzI2C72eEmkUIyK","object":"chat.completion.chunk","created":1776257890,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"5"}}]},"finish_reason":null}],"usage":null,"obfuscation":"K"}
-
- data: {"id":"chatcmpl-DUu6kWAUWCaAILhzI2C72eEmkUIyK","object":"chat.completion.chunk","created":1776257890,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/update"}}]},"finish_reason":null}],"usage":null,"obfuscation":"6Zizn9gtYHB"}
-
- data: {"id":"chatcmpl-DUu6kWAUWCaAILhzI2C72eEmkUIyK","object":"chat.completion.chunk","created":1776257890,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_a"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu6kWAUWCaAILhzI2C72eEmkUIyK","object":"chat.completion.chunk","created":1776257890,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_file"}}]},"finish_reason":null}],"usage":null,"obfuscation":"9wehDODbZ5o1R"}
-
- data: {"id":"chatcmpl-DUu6kWAUWCaAILhzI2C72eEmkUIyK","object":"chat.completion.chunk","created":1776257890,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\",\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"CcD1CoEf2IyAP"}
-
- data: {"id":"chatcmpl-DUu6kWAUWCaAILhzI2C72eEmkUIyK","object":"chat.completion.chunk","created":1776257890,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"depth"}}]},"finish_reason":null}],"usage":null,"obfuscation":"wDjNoY2RuUawO"}
-
- data: {"id":"chatcmpl-DUu6kWAUWCaAILhzI2C72eEmkUIyK","object":"chat.completion.chunk","created":1776257890,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"hhf5Z7Ufo2AYOq4"}
-
- data: {"id":"chatcmpl-DUu6kWAUWCaAILhzI2C72eEmkUIyK","object":"chat.completion.chunk","created":1776257890,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"2"}}]},"finish_reason":null}],"usage":null,"obfuscation":"P"}
-
- data: {"id":"chatcmpl-DUu6kWAUWCaAILhzI2C72eEmkUIyK","object":"chat.completion.chunk","created":1776257890,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"}"}}]},"finish_reason":null}],"usage":null,"obfuscation":"S"}
-
- data: {"id":"chatcmpl-DUu6kWAUWCaAILhzI2C72eEmkUIyK","object":"chat.completion.chunk","created":1776257890,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu6kWAUWCaAILhzI2C72eEmkUIyK","object":"chat.completion.chunk","created":1776257890,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6379,"completion_tokens":43,"total_tokens":6422,"prompt_tokens_details":{"cached_tokens":3456,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":0,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"FID4"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 2.679926166s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31064
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/update_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"update the main.go file by changing the print to say hello from crush\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_fVC8ZP9c6nI5ZmyOkzi3PKR0\",\"function\":{\"arguments\":\"{\\\"path\\\":\\\"/tmp/crush-test/TestCoderAgent/openai-gpt-5/update_a_file\\\",\\\"depth\\\":2}\",\"name\":\"ls\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n- /tmp/crush-test/TestCoderAgent/openai-gpt-5/update_a_file/\\n - go.mod\\n - main.go\\n\",\"tool_call_id\":\"call_fVC8ZP9c6nI5ZmyOkzi3PKR0\",\"role\":\"tool\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu6nGs036a3Q9jzFdJsL67Ky8b9n","object":"chat.completion.chunk","created":1776257893,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":null,"tool_calls":[{"index":0,"id":"call_8K3cqZYsa8xSduqeAoASkOGl","type":"function","function":{"name":"view","arguments":""}}],"refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"MOaXl9vUitTj"}
-
- data: {"id":"chatcmpl-DUu6nGs036a3Q9jzFdJsL67Ky8b9n","object":"chat.completion.chunk","created":1776257893,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"QNsFf4uiIYTa2lV"}
-
- data: {"id":"chatcmpl-DUu6nGs036a3Q9jzFdJsL67Ky8b9n","object":"chat.completion.chunk","created":1776257893,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"file"}}]},"finish_reason":null}],"usage":null,"obfuscation":"vIFNspxsPaeLo7"}
-
- data: {"id":"chatcmpl-DUu6nGs036a3Q9jzFdJsL67Ky8b9n","object":"chat.completion.chunk","created":1776257893,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_path"}}]},"finish_reason":null}],"usage":null,"obfuscation":"ODTicF5h6mleJ"}
-
- data: {"id":"chatcmpl-DUu6nGs036a3Q9jzFdJsL67Ky8b9n","object":"chat.completion.chunk","created":1776257893,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"ciOz0OD4cXXECRN"}
-
- data: {"id":"chatcmpl-DUu6nGs036a3Q9jzFdJsL67Ky8b9n","object":"chat.completion.chunk","created":1776257893,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"/"}}]},"finish_reason":null}],"usage":null,"obfuscation":"WJRC5BFWUiN37AE"}
-
- data: {"id":"chatcmpl-DUu6nGs036a3Q9jzFdJsL67Ky8b9n","object":"chat.completion.chunk","created":1776257893,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"tmp"}}]},"finish_reason":null}],"usage":null,"obfuscation":"heD9ArEFsBsgvW7"}
-
- data: {"id":"chatcmpl-DUu6nGs036a3Q9jzFdJsL67Ky8b9n","object":"chat.completion.chunk","created":1776257893,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/cr"}}]},"finish_reason":null}],"usage":null,"obfuscation":"t42GQ1q5hjR3yQo"}
-
- data: {"id":"chatcmpl-DUu6nGs036a3Q9jzFdJsL67Ky8b9n","object":"chat.completion.chunk","created":1776257893,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ush"}}]},"finish_reason":null}],"usage":null,"obfuscation":"weVvid8a5HXZlR1"}
-
- data: {"id":"chatcmpl-DUu6nGs036a3Q9jzFdJsL67Ky8b9n","object":"chat.completion.chunk","created":1776257893,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"PVK9MCexgPjtK"}
-
- data: {"id":"chatcmpl-DUu6nGs036a3Q9jzFdJsL67Ky8b9n","object":"chat.completion.chunk","created":1776257893,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/Test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"IBmu8r7FHWb5J"}
-
- data: {"id":"chatcmpl-DUu6nGs036a3Q9jzFdJsL67Ky8b9n","object":"chat.completion.chunk","created":1776257893,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Coder"}}]},"finish_reason":null}],"usage":null,"obfuscation":"JVbYpdLKIpWTN"}
-
- data: {"id":"chatcmpl-DUu6nGs036a3Q9jzFdJsL67Ky8b9n","object":"chat.completion.chunk","created":1776257893,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Agent"}}]},"finish_reason":null}],"usage":null,"obfuscation":"XFElBzgW9073R"}
-
- data: {"id":"chatcmpl-DUu6nGs036a3Q9jzFdJsL67Ky8b9n","object":"chat.completion.chunk","created":1776257893,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/open"}}]},"finish_reason":null}],"usage":null,"obfuscation":"oas3V3q6oiczI"}
-
- data: {"id":"chatcmpl-DUu6nGs036a3Q9jzFdJsL67Ky8b9n","object":"chat.completion.chunk","created":1776257893,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ai"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu6nGs036a3Q9jzFdJsL67Ky8b9n","object":"chat.completion.chunk","created":1776257893,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-g"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu6nGs036a3Q9jzFdJsL67Ky8b9n","object":"chat.completion.chunk","created":1776257893,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"pt"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu6nGs036a3Q9jzFdJsL67Ky8b9n","object":"chat.completion.chunk","created":1776257893,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-"}}]},"finish_reason":null}],"usage":null,"obfuscation":"R"}
-
- data: {"id":"chatcmpl-DUu6nGs036a3Q9jzFdJsL67Ky8b9n","object":"chat.completion.chunk","created":1776257893,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"5"}}]},"finish_reason":null}],"usage":null,"obfuscation":"i"}
-
- data: {"id":"chatcmpl-DUu6nGs036a3Q9jzFdJsL67Ky8b9n","object":"chat.completion.chunk","created":1776257893,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/update"}}]},"finish_reason":null}],"usage":null,"obfuscation":"fs92AkSD0ol"}
-
- data: {"id":"chatcmpl-DUu6nGs036a3Q9jzFdJsL67Ky8b9n","object":"chat.completion.chunk","created":1776257893,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_a"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu6nGs036a3Q9jzFdJsL67Ky8b9n","object":"chat.completion.chunk","created":1776257893,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_file"}}]},"finish_reason":null}],"usage":null,"obfuscation":"mSy676EEViEgW"}
-
- data: {"id":"chatcmpl-DUu6nGs036a3Q9jzFdJsL67Ky8b9n","object":"chat.completion.chunk","created":1776257893,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/main"}}]},"finish_reason":null}],"usage":null,"obfuscation":"WxyzCVliVmRMF"}
-
- data: {"id":"chatcmpl-DUu6nGs036a3Q9jzFdJsL67Ky8b9n","object":"chat.completion.chunk","created":1776257893,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".go"}}]},"finish_reason":null}],"usage":null,"obfuscation":"NJXI6N6ZVdHmoUn"}
-
- data: {"id":"chatcmpl-DUu6nGs036a3Q9jzFdJsL67Ky8b9n","object":"chat.completion.chunk","created":1776257893,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"}"}}]},"finish_reason":null}],"usage":null,"obfuscation":"QqnY9ikSGq9D0nn"}
-
- data: {"id":"chatcmpl-DUu6nGs036a3Q9jzFdJsL67Ky8b9n","object":"chat.completion.chunk","created":1776257893,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu6nGs036a3Q9jzFdJsL67Ky8b9n","object":"chat.completion.chunk","created":1776257893,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6456,"completion_tokens":35,"total_tokens":6491,"prompt_tokens_details":{"cached_tokens":6272,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":0,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"OPEm"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 735.781833ms
-- id: 3
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31512
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/update_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"update the main.go file by changing the print to say hello from crush\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_fVC8ZP9c6nI5ZmyOkzi3PKR0\",\"function\":{\"arguments\":\"{\\\"path\\\":\\\"/tmp/crush-test/TestCoderAgent/openai-gpt-5/update_a_file\\\",\\\"depth\\\":2}\",\"name\":\"ls\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n- /tmp/crush-test/TestCoderAgent/openai-gpt-5/update_a_file/\\n - go.mod\\n - main.go\\n\",\"tool_call_id\":\"call_fVC8ZP9c6nI5ZmyOkzi3PKR0\",\"role\":\"tool\"},{\"tool_calls\":[{\"id\":\"call_8K3cqZYsa8xSduqeAoASkOGl\",\"function\":{\"arguments\":\"{\\\"file_path\\\":\\\"/tmp/crush-test/TestCoderAgent/openai-gpt-5/update_a_file/main.go\\\"}\",\"name\":\"view\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\tfmt.Println(\\\"Hello, World!\\\")\\n 7|}\\n\\n\",\"tool_call_id\":\"call_8K3cqZYsa8xSduqeAoASkOGl\",\"role\":\"tool\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":null,"tool_calls":[{"index":0,"id":"call_mlB5XrcdPc1VVwnSdwYg78Hp","type":"function","function":{"name":"edit","arguments":""}}],"refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"Yc2ggPGf6Pup"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"lB0sv6C2yxvsmuT"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"file"}}]},"finish_reason":null}],"usage":null,"obfuscation":"b649fd9AnkMi1d"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_path"}}]},"finish_reason":null}],"usage":null,"obfuscation":"0h5FLzNzylbzo"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"3iegMEC75ol7B0o"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"/"}}]},"finish_reason":null}],"usage":null,"obfuscation":"JuCS8YGHUnrkBcN"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"tmp"}}]},"finish_reason":null}],"usage":null,"obfuscation":"1SBWxwNdBHJ6DYp"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/cr"}}]},"finish_reason":null}],"usage":null,"obfuscation":"TOrlrZslJZgW43x"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ush"}}]},"finish_reason":null}],"usage":null,"obfuscation":"yFccWHNQfTEIY0K"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"FqaRPUQ10XoLM"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/Test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"ikSWvzWtwOU6s"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Coder"}}]},"finish_reason":null}],"usage":null,"obfuscation":"M7mjWeBSVGIlp"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Agent"}}]},"finish_reason":null}],"usage":null,"obfuscation":"1qdNn9chTQj7a"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/open"}}]},"finish_reason":null}],"usage":null,"obfuscation":"NBqLsgZdzkbgk"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ai"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-g"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"pt"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-"}}]},"finish_reason":null}],"usage":null,"obfuscation":"c"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"5"}}]},"finish_reason":null}],"usage":null,"obfuscation":"I"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/update"}}]},"finish_reason":null}],"usage":null,"obfuscation":"VaQPEOquJxp"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_a"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_file"}}]},"finish_reason":null}],"usage":null,"obfuscation":"g9UKmmNOyTCMM"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/main"}}]},"finish_reason":null}],"usage":null,"obfuscation":"NOGwb6BXdsZCX"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".go"}}]},"finish_reason":null}],"usage":null,"obfuscation":"qHroOD8qM2LEzW4"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\",\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"SsPBg9Ia8Vkj3"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"old"}}]},"finish_reason":null}],"usage":null,"obfuscation":"Er4G9VN2JJ23CQn"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_string"}}]},"finish_reason":null}],"usage":null,"obfuscation":"5i6QmfsR8fJ"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"uDDI3wOCvuRU3"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"func"}}]},"finish_reason":null}],"usage":null,"obfuscation":"HcVqTsOiFZFoSm"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" main"}}]},"finish_reason":null}],"usage":null,"obfuscation":"Fdy9HeQ7p4gow"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"()"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" {\\"}}]},"finish_reason":null}],"usage":null,"obfuscation":"Y1N5dvoFJ16wTU"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"n"}}]},"finish_reason":null}],"usage":null,"obfuscation":"W"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\\t"}}]},"finish_reason":null}],"usage":null,"obfuscation":"W98wQHSQ4qySVz1"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"fmt"}}]},"finish_reason":null}],"usage":null,"obfuscation":"hnGNEkxjyUQf69a"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".Println"}}]},"finish_reason":null}],"usage":null,"obfuscation":"vFcWq9aJYl"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"(\\\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"TAVZJJt4wq7AU"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Hello"}}]},"finish_reason":null}],"usage":null,"obfuscation":"REZnbhCskstFo"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":","}}]},"finish_reason":null}],"usage":null,"obfuscation":"V"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" World"}}]},"finish_reason":null}],"usage":null,"obfuscation":"dP3on4gC9vDo"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"!\\"}}]},"finish_reason":null}],"usage":null,"obfuscation":"7yzdIk7oN8qlvzL"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\")"}}]},"finish_reason":null}],"usage":null,"obfuscation":"Iaf1un1JlHaLaj9"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\\"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"n"}}]},"finish_reason":null}],"usage":null,"obfuscation":"p"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"}\\"}}]},"finish_reason":null}],"usage":null,"obfuscation":"mTC2EviYfwFy2nn"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"n"}}]},"finish_reason":null}],"usage":null,"obfuscation":"q"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\",\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"3618WzVe6yZNy"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"new"}}]},"finish_reason":null}],"usage":null,"obfuscation":"CJmHjClY8EbK367"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_string"}}]},"finish_reason":null}],"usage":null,"obfuscation":"l9kpQMdfDa3"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"UZaoDiHgZVokS"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"func"}}]},"finish_reason":null}],"usage":null,"obfuscation":"XXZtsFOtzrf4Pf"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" main"}}]},"finish_reason":null}],"usage":null,"obfuscation":"uZYyDi1t1pIlE"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"()"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" {\\"}}]},"finish_reason":null}],"usage":null,"obfuscation":"zweTbJ4k4Zo16J"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"n"}}]},"finish_reason":null}],"usage":null,"obfuscation":"8"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\\t"}}]},"finish_reason":null}],"usage":null,"obfuscation":"B7zMJoIHRolppBj"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"fmt"}}]},"finish_reason":null}],"usage":null,"obfuscation":"AJwLFofih0N2zYr"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".Println"}}]},"finish_reason":null}],"usage":null,"obfuscation":"sfIjS9JfVE"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"(\\\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"E7jFQPJ7H7Ret"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"hello"}}]},"finish_reason":null}],"usage":null,"obfuscation":"wdOBPyQ31b27q"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" from"}}]},"finish_reason":null}],"usage":null,"obfuscation":"8LwW4hzNSK9oN"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" crush"}}]},"finish_reason":null}],"usage":null,"obfuscation":"o5jDu5JlwBDd"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\\\")"}}]},"finish_reason":null}],"usage":null,"obfuscation":"ClvlMRoNzbqqK"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\\"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"n"}}]},"finish_reason":null}],"usage":null,"obfuscation":"k"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"}\\"}}]},"finish_reason":null}],"usage":null,"obfuscation":"uBLPsxhuURjKoH4"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"n"}}]},"finish_reason":null}],"usage":null,"obfuscation":"2"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"}"}}]},"finish_reason":null}],"usage":null,"obfuscation":"tqmaG0oUJUHCcek"}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu6oZXzFMaDo11qNBamzYi1KQFn0","object":"chat.completion.chunk","created":1776257894,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6553,"completion_tokens":788,"total_tokens":7341,"prompt_tokens_details":{"cached_tokens":6400,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":704,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"l"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 14.115683291s
-- id: 4
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 32085
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/update_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"update the main.go file by changing the print to say hello from crush\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_fVC8ZP9c6nI5ZmyOkzi3PKR0\",\"function\":{\"arguments\":\"{\\\"path\\\":\\\"/tmp/crush-test/TestCoderAgent/openai-gpt-5/update_a_file\\\",\\\"depth\\\":2}\",\"name\":\"ls\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n- /tmp/crush-test/TestCoderAgent/openai-gpt-5/update_a_file/\\n - go.mod\\n - main.go\\n\",\"tool_call_id\":\"call_fVC8ZP9c6nI5ZmyOkzi3PKR0\",\"role\":\"tool\"},{\"tool_calls\":[{\"id\":\"call_8K3cqZYsa8xSduqeAoASkOGl\",\"function\":{\"arguments\":\"{\\\"file_path\\\":\\\"/tmp/crush-test/TestCoderAgent/openai-gpt-5/update_a_file/main.go\\\"}\",\"name\":\"view\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\tfmt.Println(\\\"Hello, World!\\\")\\n 7|}\\n\\n\",\"tool_call_id\":\"call_8K3cqZYsa8xSduqeAoASkOGl\",\"role\":\"tool\"},{\"tool_calls\":[{\"id\":\"call_mlB5XrcdPc1VVwnSdwYg78Hp\",\"function\":{\"arguments\":\"{\\\"file_path\\\":\\\"/tmp/crush-test/TestCoderAgent/openai-gpt-5/update_a_file/main.go\\\",\\\"old_string\\\":\\\"func main() {\\\\n\\\\tfmt.Println(\\\\\\\"Hello, World!\\\\\\\")\\\\n}\\\\n\\\",\\\"new_string\\\":\\\"func main() {\\\\n\\\\tfmt.Println(\\\\\\\"hello from crush\\\\\\\")\\\\n}\\\\n\\\"}\",\"name\":\"edit\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\nContent replaced in file: /tmp/crush-test/TestCoderAgent/openai-gpt-5/update_a_file/main.go\\n\\n\",\"tool_call_id\":\"call_mlB5XrcdPc1VVwnSdwYg78Hp\",\"role\":\"tool\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":null,"tool_calls":[{"index":0,"id":"call_783MnDcur0LereCQrd31ySqV","type":"function","function":{"name":"bash","arguments":""}}],"refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"LMxQ9Lvjk9Gk"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"Jajn4RzhcjIR9un"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"command"}}]},"finish_reason":null}],"usage":null,"obfuscation":"n4BmphAxye6"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"Vwwu52nFic09N"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"go"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" run"}}]},"finish_reason":null}],"usage":null,"obfuscation":"swYT7pVtUFvIaD"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" main"}}]},"finish_reason":null}],"usage":null,"obfuscation":"TdryZjGfAqOa0"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".go"}}]},"finish_reason":null}],"usage":null,"obfuscation":"Hrbt2cUQWcPpuyA"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\",\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"QYiY4ZY40c6ee"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"description"}}]},"finish_reason":null}],"usage":null,"obfuscation":"tH1bCya"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"1WNfjQWjWKP4C"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Run"}}]},"finish_reason":null}],"usage":null,"obfuscation":"JHuWsOMEfFJ8d8A"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" Go"}}]},"finish_reason":null}],"usage":null,"obfuscation":"y1gYKnbWwr0tow8"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" program"}}]},"finish_reason":null}],"usage":null,"obfuscation":"Q6fKxq8rn8"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\",\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"6ybRDDEYb1YLC"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"working"}}]},"finish_reason":null}],"usage":null,"obfuscation":"HBR9hbKmd6O"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_dir"}}]},"finish_reason":null}],"usage":null,"obfuscation":"JCBFbQfcck6ugE"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"ZlrjIhgFIK2iYeo"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"/"}}]},"finish_reason":null}],"usage":null,"obfuscation":"CunfOrm9ULUcipH"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"tmp"}}]},"finish_reason":null}],"usage":null,"obfuscation":"MNL7udoBSFHkIzR"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/cr"}}]},"finish_reason":null}],"usage":null,"obfuscation":"sQpBILnf0p4ny6B"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ush"}}]},"finish_reason":null}],"usage":null,"obfuscation":"39WMrcq0mb0JphW"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"qPotaLR07CITk"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/Test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"9WVVBmz1laYyV"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Coder"}}]},"finish_reason":null}],"usage":null,"obfuscation":"S1d6Inair1LGT"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Agent"}}]},"finish_reason":null}],"usage":null,"obfuscation":"UJ9DpeOYGZuOS"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/open"}}]},"finish_reason":null}],"usage":null,"obfuscation":"ZeCUy2S1O7BdJ"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ai"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-g"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"pt"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-"}}]},"finish_reason":null}],"usage":null,"obfuscation":"Z"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"5"}}]},"finish_reason":null}],"usage":null,"obfuscation":"l"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/update"}}]},"finish_reason":null}],"usage":null,"obfuscation":"jJtxMJ4ytv0"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_a"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_file"}}]},"finish_reason":null}],"usage":null,"obfuscation":"lntxkrbuWagQy"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"}"}}]},"finish_reason":null}],"usage":null,"obfuscation":"SjsxKaaFndean2l"}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu73DEpnY3pFVttBbqQRCQ76Fpgb","object":"chat.completion.chunk","created":1776257909,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6672,"completion_tokens":47,"total_tokens":6719,"prompt_tokens_details":{"cached_tokens":6528,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":0,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"o7T1"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 740.600417ms
-- id: 5
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 32531
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/update_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"update the main.go file by changing the print to say hello from crush\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_fVC8ZP9c6nI5ZmyOkzi3PKR0\",\"function\":{\"arguments\":\"{\\\"path\\\":\\\"/tmp/crush-test/TestCoderAgent/openai-gpt-5/update_a_file\\\",\\\"depth\\\":2}\",\"name\":\"ls\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n- /tmp/crush-test/TestCoderAgent/openai-gpt-5/update_a_file/\\n - go.mod\\n - main.go\\n\",\"tool_call_id\":\"call_fVC8ZP9c6nI5ZmyOkzi3PKR0\",\"role\":\"tool\"},{\"tool_calls\":[{\"id\":\"call_8K3cqZYsa8xSduqeAoASkOGl\",\"function\":{\"arguments\":\"{\\\"file_path\\\":\\\"/tmp/crush-test/TestCoderAgent/openai-gpt-5/update_a_file/main.go\\\"}\",\"name\":\"view\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\tfmt.Println(\\\"Hello, World!\\\")\\n 7|}\\n\\n\",\"tool_call_id\":\"call_8K3cqZYsa8xSduqeAoASkOGl\",\"role\":\"tool\"},{\"tool_calls\":[{\"id\":\"call_mlB5XrcdPc1VVwnSdwYg78Hp\",\"function\":{\"arguments\":\"{\\\"file_path\\\":\\\"/tmp/crush-test/TestCoderAgent/openai-gpt-5/update_a_file/main.go\\\",\\\"old_string\\\":\\\"func main() {\\\\n\\\\tfmt.Println(\\\\\\\"Hello, World!\\\\\\\")\\\\n}\\\\n\\\",\\\"new_string\\\":\\\"func main() {\\\\n\\\\tfmt.Println(\\\\\\\"hello from crush\\\\\\\")\\\\n}\\\\n\\\"}\",\"name\":\"edit\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\nContent replaced in file: /tmp/crush-test/TestCoderAgent/openai-gpt-5/update_a_file/main.go\\n\\n\",\"tool_call_id\":\"call_mlB5XrcdPc1VVwnSdwYg78Hp\",\"role\":\"tool\"},{\"tool_calls\":[{\"id\":\"call_783MnDcur0LereCQrd31ySqV\",\"function\":{\"arguments\":\"{\\\"command\\\":\\\"go run main.go\\\",\\\"description\\\":\\\"Run Go program\\\",\\\"working_dir\\\":\\\"/tmp/crush-test/TestCoderAgent/openai-gpt-5/update_a_file\\\"}\",\"name\":\"bash\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"hello from crush\\n\\n\\n/tmp/crush-test/TestCoderAgent/openai-gpt-5/update_a_file\",\"tool_call_id\":\"call_783MnDcur0LereCQrd31ySqV\",\"role\":\"tool\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu75zH0MzNqUA9xlL9EdofkNO0eu","object":"chat.completion.chunk","created":1776257911,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":"","refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"8xLjuE4vQZ"}
-
- data: {"id":"chatcmpl-DUu75zH0MzNqUA9xlL9EdofkNO0eu","object":"chat.completion.chunk","created":1776257911,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"Done"},"finish_reason":null}],"usage":null,"obfuscation":"az5ZEKE7"}
-
- data: {"id":"chatcmpl-DUu75zH0MzNqUA9xlL9EdofkNO0eu","object":"chat.completion.chunk","created":1776257911,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":null,"obfuscation":"1MJ3rs"}
-
- data: {"id":"chatcmpl-DUu75zH0MzNqUA9xlL9EdofkNO0eu","object":"chat.completion.chunk","created":1776257911,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6755,"completion_tokens":4,"total_tokens":6759,"prompt_tokens_details":{"cached_tokens":6656,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":0,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"2k5iO"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 1.683021625s
diff --git a/internal/agent/testdata/TestCoderAgent/openai-gpt-5/write_tool.yaml b/internal/agent/testdata/TestCoderAgent/openai-gpt-5/write_tool.yaml
deleted file mode 100644
index fb601782c2c465700d247380acafa80f001455ab..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/openai-gpt-5/write_tool.yaml
+++ /dev/null
@@ -1,216 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 851
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse write to create a new file called config.json with content ''{\"name\": \"test\", \"version\": \"1.0.0\"}''\n \n\n","role":"user"}],"model":"gpt-4o","max_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu8lVyNkTTXsfkIjl5sXmDORr187","object":"chat.completion.chunk","created":1776258015,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"role":"assistant","content":"","refusal":null},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"8U8nsmGrF8Fv1d"}
-
- data: {"id":"chatcmpl-DUu8lVyNkTTXsfkIjl5sXmDORr187","object":"chat.completion.chunk","created":1776258015,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":"Create"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"FES1t3Wwo6"}
-
- data: {"id":"chatcmpl-DUu8lVyNkTTXsfkIjl5sXmDORr187","object":"chat.completion.chunk","created":1776258015,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":" config"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"u7eObrGnf"}
-
- data: {"id":"chatcmpl-DUu8lVyNkTTXsfkIjl5sXmDORr187","object":"chat.completion.chunk","created":1776258015,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":".json"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"EAC3vpR592v"}
-
- data: {"id":"chatcmpl-DUu8lVyNkTTXsfkIjl5sXmDORr187","object":"chat.completion.chunk","created":1776258015,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":" with"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"rOIAeBXA7R5"}
-
- data: {"id":"chatcmpl-DUu8lVyNkTTXsfkIjl5sXmDORr187","object":"chat.completion.chunk","created":1776258015,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":" test"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"gmnjJ1S0yuP"}
-
- data: {"id":"chatcmpl-DUu8lVyNkTTXsfkIjl5sXmDORr187","object":"chat.completion.chunk","created":1776258015,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":" version"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"jHPoBFcS"}
-
- data: {"id":"chatcmpl-DUu8lVyNkTTXsfkIjl5sXmDORr187","object":"chat.completion.chunk","created":1776258015,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{"content":" set"},"logprobs":null,"finish_reason":null}],"usage":null,"obfuscation":"RZw1hvk9C6ty"}
-
- data: {"id":"chatcmpl-DUu8lVyNkTTXsfkIjl5sXmDORr187","object":"chat.completion.chunk","created":1776258015,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[{"index":0,"delta":{},"logprobs":null,"finish_reason":"stop"}],"usage":null,"obfuscation":"CEG2uvyf8b"}
-
- data: {"id":"chatcmpl-DUu8lVyNkTTXsfkIjl5sXmDORr187","object":"chat.completion.chunk","created":1776258015,"model":"gpt-4o-2024-08-06","service_tier":"default","system_fingerprint":"fp_d08c293973","choices":[],"usage":{"prompt_tokens":169,"completion_tokens":7,"total_tokens":176,"prompt_tokens_details":{"cached_tokens":0,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":0,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"Wrdx8Q5Hl3ci7s"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 575.75125ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30719
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/write_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use write to create a new file called config.json with content '{\\\"name\\\": \\\"test\\\", \\\"version\\\": \\\"1.0.0\\\"}'\",\"role\":\"user\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":null,"tool_calls":[{"index":0,"id":"call_MqgU2oanL8jcyDZxTkMMo64j","type":"function","function":{"name":"write","arguments":""}}],"refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"5OnRNV6JRGl"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"UHGKlwMN0E0A4xI"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"file"}}]},"finish_reason":null}],"usage":null,"obfuscation":"qwYQz8DcIBOlXq"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_path"}}]},"finish_reason":null}],"usage":null,"obfuscation":"YtbWZI1y9PIPy"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"5q6Itxk8biTvt7q"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"/"}}]},"finish_reason":null}],"usage":null,"obfuscation":"1IMJscj3yzs4sR7"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"tmp"}}]},"finish_reason":null}],"usage":null,"obfuscation":"s5RT8RYId6dXiFA"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/cr"}}]},"finish_reason":null}],"usage":null,"obfuscation":"tvZoZ2PwveO1BFJ"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ush"}}]},"finish_reason":null}],"usage":null,"obfuscation":"SPvslpPeq95KlXg"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"qr1MotSY95556"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/Test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"i2fHbUSXxJ65F"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Coder"}}]},"finish_reason":null}],"usage":null,"obfuscation":"OD1F9DgkAPnWl"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"Agent"}}]},"finish_reason":null}],"usage":null,"obfuscation":"OiPwT1YGtBgAe"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/open"}}]},"finish_reason":null}],"usage":null,"obfuscation":"BrTaayHysaW6K"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"ai"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-g"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"pt"}}]},"finish_reason":null}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"-"}}]},"finish_reason":null}],"usage":null,"obfuscation":"K"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"5"}}]},"finish_reason":null}],"usage":null,"obfuscation":"V"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/write"}}]},"finish_reason":null}],"usage":null,"obfuscation":"IfaSLeivAoU9"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"_tool"}}]},"finish_reason":null}],"usage":null,"obfuscation":"PEu4dOlq3eLNv"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"/config"}}]},"finish_reason":null}],"usage":null,"obfuscation":"4kJNo5fTbeq"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":".json"}}]},"finish_reason":null}],"usage":null,"obfuscation":"daeicKfqc9Brn"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\",\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"gW2aIQLfos4Y1"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"content"}}]},"finish_reason":null}],"usage":null,"obfuscation":"a7dWWQ45uHM"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\":\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"N7EVdPCaIWqcn"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\\\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"uSjOa5d9gNk8x"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"name"}}]},"finish_reason":null}],"usage":null,"obfuscation":"P7aPka70s8EzOD"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\\\":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"rS4gtcgZMAOaH"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" \\\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"jbEPcqRxcfsg7"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"test"}}]},"finish_reason":null}],"usage":null,"obfuscation":"fx6TPw75BEt9YX"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\\\","}}]},"finish_reason":null}],"usage":null,"obfuscation":"d8LC7hwVaKSiR"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" \\\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"zx7jYnoee7dPJ"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"version"}}]},"finish_reason":null}],"usage":null,"obfuscation":"Badn8UZZVTY"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\\\":"}}]},"finish_reason":null}],"usage":null,"obfuscation":"OHUXjBM1yV0c8"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":" \\\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"dKTEAgQdGizFr"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"1"}}]},"finish_reason":null}],"usage":null,"obfuscation":"O"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"."}}]},"finish_reason":null}],"usage":null,"obfuscation":"E"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"0"}}]},"finish_reason":null}],"usage":null,"obfuscation":"2"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"."}}]},"finish_reason":null}],"usage":null,"obfuscation":"H"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"0"}}]},"finish_reason":null}],"usage":null,"obfuscation":"4"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\\\""}}]},"finish_reason":null}],"usage":null,"obfuscation":"vVGC8kX2e9q6my"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"}"}}]},"finish_reason":null}],"usage":null,"obfuscation":"s"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"\"}"}}]},"finish_reason":null}],"usage":null,"obfuscation":"4aBB5WrvoEGlRtp"}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}],"usage":null,"obfuscation":""}
-
- data: {"id":"chatcmpl-DUu8mqaju5u98MOiDjwMVr1TPbL5S","object":"chat.completion.chunk","created":1776258016,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6393,"completion_tokens":316,"total_tokens":6709,"prompt_tokens_details":{"cached_tokens":3456,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":256,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"6"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 10.758426625s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31208
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openai-gpt-5/write_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use write to create a new file called config.json with content '{\\\"name\\\": \\\"test\\\", \\\"version\\\": \\\"1.0.0\\\"}'\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_MqgU2oanL8jcyDZxTkMMo64j\",\"function\":{\"arguments\":\"{\\\"file_path\\\":\\\"/tmp/crush-test/TestCoderAgent/openai-gpt-5/write_tool/config.json\\\",\\\"content\\\":\\\"{\\\\\\\"name\\\\\\\": \\\\\\\"test\\\\\\\", \\\\\\\"version\\\\\\\": \\\\\\\"1.0.0\\\\\\\"}\\\"}\",\"name\":\"write\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\nFile successfully written: /tmp/crush-test/TestCoderAgent/openai-gpt-5/write_tool/config.json\\n\",\"tool_call_id\":\"call_MqgU2oanL8jcyDZxTkMMo64j\",\"role\":\"tool\"}],\"model\":\"gpt-5\",\"max_completion_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.openai.com/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"chatcmpl-DUu8yqM8OBbxW9HYtlmGai2JBzI1k","object":"chat.completion.chunk","created":1776258028,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"role":"assistant","content":"","refusal":null},"finish_reason":null}],"usage":null,"obfuscation":"nZHNZqiV1x"}
-
- data: {"id":"chatcmpl-DUu8yqM8OBbxW9HYtlmGai2JBzI1k","object":"chat.completion.chunk","created":1776258028,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{"content":"Done"},"finish_reason":null}],"usage":null,"obfuscation":"rnbJlYUh"}
-
- data: {"id":"chatcmpl-DUu8yqM8OBbxW9HYtlmGai2JBzI1k","object":"chat.completion.chunk","created":1776258028,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":null,"obfuscation":"ijAw3z"}
-
- data: {"id":"chatcmpl-DUu8yqM8OBbxW9HYtlmGai2JBzI1k","object":"chat.completion.chunk","created":1776258028,"model":"gpt-5-2025-08-07","service_tier":"default","system_fingerprint":null,"choices":[],"usage":{"prompt_tokens":6486,"completion_tokens":4,"total_tokens":6490,"prompt_tokens_details":{"cached_tokens":6272,"audio_tokens":0},"completion_tokens_details":{"reasoning_tokens":0,"audio_tokens":0,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"obfuscation":"udxSS"}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream; charset=utf-8
- status: 200 OK
- code: 200
- duration: 1.444174958s
diff --git a/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/bash_tool.yaml b/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/bash_tool.yaml
deleted file mode 100644
index e96a14bab16967dac0c4135bd9960b69f1f167f7..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/bash_tool.yaml
+++ /dev/null
@@ -1,287 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 887
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse bash to create a file named test.txt with content ''hello bash''. do not print its timestamp\n \n\n","role":"user"}],"model":"qwen/qwen3-next-80b-a3b-instruct","max_tokens":40,"stream_options":{"include_usage":true},"usage":{"include":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258422-gEiJdRzuFc7MzOlqcEsC","object":"chat.completion.chunk","created":1776258422,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":"Create","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258422-gEiJdRzuFc7MzOlqcEsC","object":"chat.completion.chunk","created":1776258422,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":" test","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258422-gEiJdRzuFc7MzOlqcEsC","object":"chat.completion.chunk","created":1776258422,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":".txt with hello","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258422-gEiJdRzuFc7MzOlqcEsC","object":"chat.completion.chunk","created":1776258422,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":" bash using","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258422-gEiJdRzuFc7MzOlqcEsC","object":"chat.completion.chunk","created":1776258422,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":" bash","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258422-gEiJdRzuFc7MzOlqcEsC","object":"chat.completion.chunk","created":1776258422,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}]}
-
- data: {"id":"gen-1776258422-gEiJdRzuFc7MzOlqcEsC","object":"chat.completion.chunk","created":1776258422,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}],"usage":{"prompt_tokens":167,"completion_tokens":8,"total_tokens":175,"cost":0.0000225225,"is_byok":false,"prompt_tokens_details":{"cached_tokens":0,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00003465,"upstream_inference_prompt_cost":0.00002505,"upstream_inference_completions_cost":0.0000096},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 2.484926s
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30852
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/bash_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"use bash to create a file named test.txt with content 'hello bash'. do not print its timestamp\",\"role\":\"user\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258422-26GHdfwXhHq49lFk9DSs","object":"chat.completion.chunk","created":1776258422,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"id":"bash:0","type":"function","function":{"name":"bash","arguments":""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258422-26GHdfwXhHq49lFk9DSs","object":"chat.completion.chunk","created":1776258422,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258422-26GHdfwXhHq49lFk9DSs","object":"chat.completion.chunk","created":1776258422,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"description"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258422-26GHdfwXhHq49lFk9DSs","object":"chat.completion.chunk","created":1776258422,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258422-26GHdfwXhHq49lFk9DSs","object":"chat.completion.chunk","created":1776258422,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258422-26GHdfwXhHq49lFk9DSs","object":"chat.completion.chunk","created":1776258422,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"Create"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258422-26GHdfwXhHq49lFk9DSs","object":"chat.completion.chunk","created":1776258422,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" test"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258422-26GHdfwXhHq49lFk9DSs","object":"chat.completion.chunk","created":1776258422,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":".txt"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258422-26GHdfwXhHq49lFk9DSs","object":"chat.completion.chunk","created":1776258422,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" file"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258422-26GHdfwXhHq49lFk9DSs","object":"chat.completion.chunk","created":1776258422,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" with"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258422-26GHdfwXhHq49lFk9DSs","object":"chat.completion.chunk","created":1776258422,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" content"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258422-26GHdfwXhHq49lFk9DSs","object":"chat.completion.chunk","created":1776258422,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\","}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258422-26GHdfwXhHq49lFk9DSs","object":"chat.completion.chunk","created":1776258422,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258422-26GHdfwXhHq49lFk9DSs","object":"chat.completion.chunk","created":1776258422,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"command"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258422-26GHdfwXhHq49lFk9DSs","object":"chat.completion.chunk","created":1776258422,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258422-26GHdfwXhHq49lFk9DSs","object":"chat.completion.chunk","created":1776258422,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258422-26GHdfwXhHq49lFk9DSs","object":"chat.completion.chunk","created":1776258422,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"echo"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258422-26GHdfwXhHq49lFk9DSs","object":"chat.completion.chunk","created":1776258422,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" '"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258422-26GHdfwXhHq49lFk9DSs","object":"chat.completion.chunk","created":1776258422,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"hello"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258422-26GHdfwXhHq49lFk9DSs","object":"chat.completion.chunk","created":1776258422,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" bash"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258422-26GHdfwXhHq49lFk9DSs","object":"chat.completion.chunk","created":1776258422,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"'"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258422-26GHdfwXhHq49lFk9DSs","object":"chat.completion.chunk","created":1776258422,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" >"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258422-26GHdfwXhHq49lFk9DSs","object":"chat.completion.chunk","created":1776258422,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" test"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258422-26GHdfwXhHq49lFk9DSs","object":"chat.completion.chunk","created":1776258422,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":".txt"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258422-26GHdfwXhHq49lFk9DSs","object":"chat.completion.chunk","created":1776258422,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\"}"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258422-26GHdfwXhHq49lFk9DSs","object":"chat.completion.chunk","created":1776258422,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}]}
-
- data: {"id":"gen-1776258422-26GHdfwXhHq49lFk9DSs","object":"chat.completion.chunk","created":1776258422,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}],"usage":{"prompt_tokens":6502,"completion_tokens":34,"total_tokens":6536,"cost":0.0039862,"is_byok":false,"prompt_tokens_details":{"cached_tokens":6502,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0039862,"upstream_inference_prompt_cost":0.0039012,"upstream_inference_completions_cost":0.000085},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 4.024270542s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31092
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/bash_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use bash to create a file named test.txt with content 'hello bash'. do not print its timestamp\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"bash:0\",\"function\":{\"arguments\":\"{\\\"description\\\": \\\"Create test.txt file with content\\\", \\\"command\\\": \\\"echo 'hello bash' > test.txt\\\"}\",\"name\":\"bash\"},\"type\":\"function\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"role\":\"assistant\"},{\"content\":\"no output\",\"tool_call_id\":\"bash:0\",\"role\":\"tool\"}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258426-I37Xbe1ggwblVNlc718r","object":"chat.completion.chunk","created":1776258426,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"id":"view:1","type":"function","function":{"name":"view","arguments":""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258426-I37Xbe1ggwblVNlc718r","object":"chat.completion.chunk","created":1776258426,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258426-I37Xbe1ggwblVNlc718r","object":"chat.completion.chunk","created":1776258426,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"file"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258426-I37Xbe1ggwblVNlc718r","object":"chat.completion.chunk","created":1776258426,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"_path"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258426-I37Xbe1ggwblVNlc718r","object":"chat.completion.chunk","created":1776258426,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258426-I37Xbe1ggwblVNlc718r","object":"chat.completion.chunk","created":1776258426,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \"/"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258426-I37Xbe1ggwblVNlc718r","object":"chat.completion.chunk","created":1776258426,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"tmp"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258426-I37Xbe1ggwblVNlc718r","object":"chat.completion.chunk","created":1776258426,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/c"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258426-I37Xbe1ggwblVNlc718r","object":"chat.completion.chunk","created":1776258426,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"rush"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258426-I37Xbe1ggwblVNlc718r","object":"chat.completion.chunk","created":1776258426,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-test"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258426-I37Xbe1ggwblVNlc718r","object":"chat.completion.chunk","created":1776258426,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/Test"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258426-I37Xbe1ggwblVNlc718r","object":"chat.completion.chunk","created":1776258426,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"Coder"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258426-I37Xbe1ggwblVNlc718r","object":"chat.completion.chunk","created":1776258426,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"Agent"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258426-I37Xbe1ggwblVNlc718r","object":"chat.completion.chunk","created":1776258426,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/open"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258426-I37Xbe1ggwblVNlc718r","object":"chat.completion.chunk","created":1776258426,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"router"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258426-I37Xbe1ggwblVNlc718r","object":"chat.completion.chunk","created":1776258426,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-k"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258426-I37Xbe1ggwblVNlc718r","object":"chat.completion.chunk","created":1776258426,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"imi"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258426-I37Xbe1ggwblVNlc718r","object":"chat.completion.chunk","created":1776258426,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-k"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258426-I37Xbe1ggwblVNlc718r","object":"chat.completion.chunk","created":1776258426,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"2"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258426-I37Xbe1ggwblVNlc718r","object":"chat.completion.chunk","created":1776258426,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/bash"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258426-I37Xbe1ggwblVNlc718r","object":"chat.completion.chunk","created":1776258426,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"_tool"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258426-I37Xbe1ggwblVNlc718r","object":"chat.completion.chunk","created":1776258426,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/test"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258426-I37Xbe1ggwblVNlc718r","object":"chat.completion.chunk","created":1776258426,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":".txt"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258426-I37Xbe1ggwblVNlc718r","object":"chat.completion.chunk","created":1776258426,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\"}"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258426-I37Xbe1ggwblVNlc718r","object":"chat.completion.chunk","created":1776258426,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}]}
-
- data: {"id":"gen-1776258426-I37Xbe1ggwblVNlc718r","object":"chat.completion.chunk","created":1776258426,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}],"usage":{"prompt_tokens":6555,"completion_tokens":32,"total_tokens":6587,"cost":0.004013,"is_byok":false,"prompt_tokens_details":{"cached_tokens":6400,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.004013,"upstream_inference_prompt_cost":0.003933,"upstream_inference_completions_cost":0.00008},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 2.922342916s
-- id: 3
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31382
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/bash_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use bash to create a file named test.txt with content 'hello bash'. do not print its timestamp\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"bash:0\",\"function\":{\"arguments\":\"{\\\"description\\\": \\\"Create test.txt file with content\\\", \\\"command\\\": \\\"echo 'hello bash' > test.txt\\\"}\",\"name\":\"bash\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"no output\",\"tool_call_id\":\"bash:0\",\"role\":\"tool\"},{\"tool_calls\":[{\"id\":\"view:1\",\"function\":{\"arguments\":\"{\\\"file_path\\\": \\\"/tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/bash_tool/test.txt\\\"}\",\"name\":\"view\"},\"type\":\"function\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"role\":\"assistant\"},{\"content\":\"\\n 1|hello bash\\n\\n\",\"tool_call_id\":\"view:1\",\"role\":\"tool\"}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258429-LaoGFYgO1QmvL9F6qD9N","object":"chat.completion.chunk","created":1776258429,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"Done","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258429-LaoGFYgO1QmvL9F6qD9N","object":"chat.completion.chunk","created":1776258429,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}]}
-
- data: {"id":"gen-1776258429-LaoGFYgO1QmvL9F6qD9N","object":"chat.completion.chunk","created":1776258429,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}],"usage":{"prompt_tokens":6615,"completion_tokens":2,"total_tokens":6617,"cost":0.003974,"is_byok":false,"prompt_tokens_details":{"cached_tokens":6400,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.003974,"upstream_inference_prompt_cost":0.003969,"upstream_inference_completions_cost":0.000005},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 1.192156583s
diff --git a/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/download_tool.yaml b/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/download_tool.yaml
deleted file mode 100644
index a46509102dcaca37dcaac185c470c31add162610..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/download_tool.yaml
+++ /dev/null
@@ -1,286 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 908
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\ndownload the file from https://example-files.online-convert.com/document/txt/example.txt and save it as example.txt\n \n\n","role":"user"}],"model":"qwen/qwen3-next-80b-a3b-instruct","max_tokens":40,"stream_options":{"include_usage":true},"usage":{"include":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258083-34CApIFOELFlfXiEavBa","object":"chat.completion.chunk","created":1776258083,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"AtlasCloud","choices":[{"index":0,"delta":{"content":"Download","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-34CApIFOELFlfXiEavBa","object":"chat.completion.chunk","created":1776258083,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"AtlasCloud","choices":[{"index":0,"delta":{"content":" example","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-34CApIFOELFlfXiEavBa","object":"chat.completion.chunk","created":1776258083,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"AtlasCloud","choices":[{"index":0,"delta":{"content":".txt","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-34CApIFOELFlfXiEavBa","object":"chat.completion.chunk","created":1776258083,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"AtlasCloud","choices":[{"index":0,"delta":{"content":" from","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-34CApIFOELFlfXiEavBa","object":"chat.completion.chunk","created":1776258083,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"AtlasCloud","choices":[{"index":0,"delta":{"content":" example","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-34CApIFOELFlfXiEavBa","object":"chat.completion.chunk","created":1776258083,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"AtlasCloud","choices":[{"index":0,"delta":{"content":"-files","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-34CApIFOELFlfXiEavBa","object":"chat.completion.chunk","created":1776258083,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"AtlasCloud","choices":[{"index":0,"delta":{"content":".online","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-34CApIFOELFlfXiEavBa","object":"chat.completion.chunk","created":1776258083,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"AtlasCloud","choices":[{"index":0,"delta":{"content":"-","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-34CApIFOELFlfXiEavBa","object":"chat.completion.chunk","created":1776258083,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"AtlasCloud","choices":[{"index":0,"delta":{"content":"convert","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-34CApIFOELFlfXiEavBa","object":"chat.completion.chunk","created":1776258083,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"AtlasCloud","choices":[{"index":0,"delta":{"content":".com","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-34CApIFOELFlfXiEavBa","object":"chat.completion.chunk","created":1776258083,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"AtlasCloud","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}]}
-
- data: {"id":"gen-1776258083-34CApIFOELFlfXiEavBa","object":"chat.completion.chunk","created":1776258083,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"AtlasCloud","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}],"usage":{"prompt_tokens":166,"completion_tokens":11,"total_tokens":177,"cost":0.0000414,"is_byok":false,"prompt_tokens_details":{"cached_tokens":0,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0000414,"upstream_inference_prompt_cost":0.0000249,"upstream_inference_completions_cost":0.0000165},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 1.67017125s
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30877
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/download_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"download the file from https://example-files.online-convert.com/document/txt/example.txt and save it as example.txt\",\"role\":\"user\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"id":"download:0","type":"function","function":{"name":"download","arguments":""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"url"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"https"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"://"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"example"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-files"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":".online"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-con"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"vert"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":".com"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/document"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"txt"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/example"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":".txt"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\","}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"file"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"_path"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \"/"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"tmp"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/c"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"rush"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-test"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/Test"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"Coder"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"Agent"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/open"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"router"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-k"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"imi"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-k"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"2"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/download"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"_tool"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/example"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":".txt"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\"}"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}]}
-
- data: {"id":"gen-1776258083-diI2MY1jcYquNzpskGW6","object":"chat.completion.chunk","created":1776258083,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}],"usage":{"prompt_tokens":6505,"completion_tokens":50,"total_tokens":6555,"cost":0.004028,"is_byok":false,"prompt_tokens_details":{"cached_tokens":6144,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.004028,"upstream_inference_prompt_cost":0.003903,"upstream_inference_completions_cost":0.000125},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 2.179737042s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 0
- host: example-files.online-convert.com
- headers:
- User-Agent:
- - crush/1.0
- url: https://example-files.online-convert.com/document/txt/example.txt
- method: GET
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |-
- TXT test file
- Purpose: Provide example of this file type
- Document file type: TXT
- Version: 1.0
- Remark:
-
- Example content:
- The names "John Doe" for males, "Jane Doe" or "Jane Roe" for females, or "Jonnie Doe" and "Janie Doe" for children, or just "Doe" non-gender-specifically are used as placeholder names for a party whose true identity is unknown or must be withheld in a legal action, case, or discussion. The names are also used to refer to acorpse or hospital patient whose identity is unknown. This practice is widely used in the United States and Canada, but is rarely used in other English-speaking countries including the United Kingdom itself, from where the use of "John Doe" in a legal context originates. The names Joe Bloggs or John Smith are used in the UK instead, as well as in Australia and New Zealand.
-
- John Doe is sometimes used to refer to a typical male in other contexts as well, in a similar manner to John Q. Public, known in Great Britain as Joe Public, John Smith or Joe Bloggs. For example, the first name listed on a form is often John Doe, along with a fictional address or other fictional information to provide an example of how to fill in the form. The name is also used frequently in popular culture, for example in the Frank Capra film Meet John Doe. John Doe was also the name of a 2002 American television series.
-
- Similarly, a child or baby whose identity is unknown may be referred to as Baby Doe. A notorious murder case in Kansas City, Missouri, referred to the baby victim as Precious Doe. Other unidentified female murder victims are Cali Doe and Princess Doe. Additional persons may be called James Doe, Judy Doe, etc. However, to avoid possible confusion, if two anonymous or unknown parties are cited in a specific case or action, the surnames Doe and Roe may be used simultaneously; for example, "John Doe v. Jane Roe". If several anonymous parties are referenced, they may simply be labelled John Doe #1, John Doe #2, etc. (the U.S. Operation Delego cited 21 (numbered) "John Doe"s) or labelled with other variants of Doe / Roe / Poe / etc. Other early alternatives such as John Stiles and Richard Miles are now rarely used, and Mary Major has been used in some American federal cases.
-
-
-
- File created by https://www.online-convert.com
- More example files: https://www.online-convert.com/file-type
- Text of Example content: Wikipedia (https://en.wikipedia.org/wiki/John_Doe)
- License: Attribution-ShareAlike 4.0 (https://creativecommons.org/licenses/by-sa/4.0/)
-
- Feel free to use and share the file according to the license above.
- headers:
- Content-Type:
- - text/plain; charset=UTF-8
- status: 200 OK
- code: 200
- duration: 1.454876583s
-- id: 3
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31284
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/download_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"download the file from https://example-files.online-convert.com/document/txt/example.txt and save it as example.txt\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"download:0\",\"function\":{\"arguments\":\"{\\\"url\\\": \\\"https://example-files.online-convert.com/document/txt/example.txt\\\", \\\"file_path\\\": \\\"/tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/download_tool/example.txt\\\"}\",\"name\":\"download\"},\"type\":\"function\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"role\":\"assistant\"},{\"content\":\"Successfully downloaded 2574 bytes to example.txt (Content-Type: text/plain; charset=UTF-8)\",\"tool_call_id\":\"download:0\",\"role\":\"tool\"}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258088-ei1AEOqDnMIQZsHHiNs3","object":"chat.completion.chunk","created":1776258088,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"Done","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258088-ei1AEOqDnMIQZsHHiNs3","object":"chat.completion.chunk","created":1776258088,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}]}
-
- data: {"id":"gen-1776258088-ei1AEOqDnMIQZsHHiNs3","object":"chat.completion.chunk","created":1776258088,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}],"usage":{"prompt_tokens":6592,"completion_tokens":2,"total_tokens":6594,"cost":0.0039602,"is_byok":false,"prompt_tokens_details":{"cached_tokens":6400,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0039602,"upstream_inference_prompt_cost":0.0039552,"upstream_inference_completions_cost":0.000005},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 1.325491916s
diff --git a/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/fetch_tool.yaml b/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/fetch_tool.yaml
deleted file mode 100644
index c97bc6122e11a8d16407396973b4ce18f59997a8..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/fetch_tool.yaml
+++ /dev/null
@@ -1,289 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 929
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nfetch the content from https://example-files.online-convert.com/website/html/example.html and tell me if it contains the word ''John Doe''\n \n\n","role":"user"}],"model":"qwen/qwen3-next-80b-a3b-instruct","max_tokens":40,"stream_options":{"include_usage":true},"usage":{"include":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258090-Iz0T3rmv80ztgYAqt4UB","object":"chat.completion.chunk","created":1776258090,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":"Check","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258090-Iz0T3rmv80ztgYAqt4UB","object":"chat.completion.chunk","created":1776258090,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":" if example","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258090-Iz0T3rmv80ztgYAqt4UB","object":"chat.completion.chunk","created":1776258090,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":".html contains John","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258090-Iz0T3rmv80ztgYAqt4UB","object":"chat.completion.chunk","created":1776258090,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":" Doe","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258090-Iz0T3rmv80ztgYAqt4UB","object":"chat.completion.chunk","created":1776258090,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}]}
-
- data: {"id":"gen-1776258090-Iz0T3rmv80ztgYAqt4UB","object":"chat.completion.chunk","created":1776258090,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}],"usage":{"prompt_tokens":175,"completion_tokens":7,"total_tokens":182,"cost":0.0000225225,"is_byok":false,"prompt_tokens_details":{"cached_tokens":0,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00003465,"upstream_inference_prompt_cost":0.00002625,"upstream_inference_completions_cost":0.0000084},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 674.561292ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30895
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/fetch_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"fetch the content from https://example-files.online-convert.com/website/html/example.html and tell me if it contains the word 'John Doe'\",\"role\":\"user\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258090-WqGBLmGG7zAGycyqKhP9","object":"chat.completion.chunk","created":1776258090,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"id":"fetch:0","type":"function","function":{"name":"fetch","arguments":""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258090-WqGBLmGG7zAGycyqKhP9","object":"chat.completion.chunk","created":1776258090,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258090-WqGBLmGG7zAGycyqKhP9","object":"chat.completion.chunk","created":1776258090,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"url"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258090-WqGBLmGG7zAGycyqKhP9","object":"chat.completion.chunk","created":1776258090,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258090-WqGBLmGG7zAGycyqKhP9","object":"chat.completion.chunk","created":1776258090,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258090-WqGBLmGG7zAGycyqKhP9","object":"chat.completion.chunk","created":1776258090,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"https"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258090-WqGBLmGG7zAGycyqKhP9","object":"chat.completion.chunk","created":1776258090,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"://"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258090-WqGBLmGG7zAGycyqKhP9","object":"chat.completion.chunk","created":1776258090,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"example"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258090-WqGBLmGG7zAGycyqKhP9","object":"chat.completion.chunk","created":1776258090,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-files"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258090-WqGBLmGG7zAGycyqKhP9","object":"chat.completion.chunk","created":1776258090,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":".online"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258090-WqGBLmGG7zAGycyqKhP9","object":"chat.completion.chunk","created":1776258090,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-con"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258090-WqGBLmGG7zAGycyqKhP9","object":"chat.completion.chunk","created":1776258090,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"vert"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258090-WqGBLmGG7zAGycyqKhP9","object":"chat.completion.chunk","created":1776258090,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":".com"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258090-WqGBLmGG7zAGycyqKhP9","object":"chat.completion.chunk","created":1776258090,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258090-WqGBLmGG7zAGycyqKhP9","object":"chat.completion.chunk","created":1776258090,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"website"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258090-WqGBLmGG7zAGycyqKhP9","object":"chat.completion.chunk","created":1776258090,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/html"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258090-WqGBLmGG7zAGycyqKhP9","object":"chat.completion.chunk","created":1776258090,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/example"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258090-WqGBLmGG7zAGycyqKhP9","object":"chat.completion.chunk","created":1776258090,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":".html"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258090-WqGBLmGG7zAGycyqKhP9","object":"chat.completion.chunk","created":1776258090,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\","}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258090-WqGBLmGG7zAGycyqKhP9","object":"chat.completion.chunk","created":1776258090,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258090-WqGBLmGG7zAGycyqKhP9","object":"chat.completion.chunk","created":1776258090,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"format"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258090-WqGBLmGG7zAGycyqKhP9","object":"chat.completion.chunk","created":1776258090,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258090-WqGBLmGG7zAGycyqKhP9","object":"chat.completion.chunk","created":1776258090,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258090-WqGBLmGG7zAGycyqKhP9","object":"chat.completion.chunk","created":1776258090,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"text"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258090-WqGBLmGG7zAGycyqKhP9","object":"chat.completion.chunk","created":1776258090,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\"}"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258090-WqGBLmGG7zAGycyqKhP9","object":"chat.completion.chunk","created":1776258090,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}]}
-
- data: {"id":"gen-1776258090-WqGBLmGG7zAGycyqKhP9","object":"chat.completion.chunk","created":1776258090,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}],"usage":{"prompt_tokens":6512,"completion_tokens":33,"total_tokens":6545,"cost":0.0039897,"is_byok":false,"prompt_tokens_details":{"cached_tokens":6144,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0039897,"upstream_inference_prompt_cost":0.0039072,"upstream_inference_completions_cost":0.0000825},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 2.538716916s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 0
- host: example-files.online-convert.com
- headers:
- User-Agent:
- - crush/1.0
- url: https://example-files.online-convert.com/website/html/example.html
- method: GET
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |-
-
-
-
-
-
- HTML test file
-
- Purpose: Provide example of this file type
- Document file type: HTML
- Version: 1.0
- Remark:
-
-
-
- Example content:
- The names "John Doe" for males, "Jane Doe" or "Jane Roe" for females, or "Jonnie Doe" and "Janie Doe" for children, or just "Doe" non-gender-specifically are used as placeholder names for a party whose true identity is unknown or must be withheld in a legal action, case, or discussion. The names are also used to refer to acorpse or hospital patient whose identity is unknown. This practice is widely used in the United States and Canada, but is rarely used in other English-speaking countries including the United Kingdom itself, from where the use of "John Doe" in a legal context originates. The names Joe Bloggs or John Smith are used in the UK instead, as well as in Australia and New Zealand.
-
- John Doe is sometimes used to refer to a typical male in other contexts as well, in a similar manner to John Q. Public, known in Great Britain as Joe Public, John Smith or Joe Bloggs. For example, the first name listed on a form is often John Doe, along with a fictional address or other fictional information to provide an example of how to fill in the form. The name is also used frequently in popular culture, for example in the Frank Capra film Meet John Doe. John Doe was also the name of a 2002 American television series.
-
- Similarly, a child or baby whose identity is unknown may be referred to as Baby Doe. A notorious murder case in Kansas City, Missouri, referred to the baby victim as Precious Doe. Other unidentified female murder victims are Cali Doe and Princess Doe. Additional persons may be called James Doe, Judy Doe, etc. However, to avoid possible confusion, if two anonymous or unknown parties are cited in a specific case or action, the surnames Doe and Roe may be used simultaneously; for example, "John Doe v. Jane Roe". If several anonymous parties are referenced, they may simply be labelled John Doe #1, John Doe #2, etc. (the U.S. Operation Delego cited 21 (numbered) "John Doe"s) or labelled with other variants of Doe / Roe / Poe / etc. Other early alternatives such as John Stiles and Richard Miles are now rarely used, and Mary Major has been used in some American federal cases.
-
-
-
- File created by https://www.online-convert.com
- More example files: https://www.online-convert.com/file-type
- Text of "Example content": Wikipedia
- License: Attribution-ShareAlike 3.0 Unported
-
-
-
- Feel free to use and share the file according to license above.
-
-
-
- headers:
- Content-Type:
- - text/html; charset=UTF-8
- status: 200 OK
- code: 200
- duration: 475.292375ms
-- id: 3
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 33634
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/fetch_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"fetch the content from https://example-files.online-convert.com/website/html/example.html and tell me if it contains the word 'John Doe'\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"fetch:0\",\"function\":{\"arguments\":\"{\\\"url\\\": \\\"https://example-files.online-convert.com/website/html/example.html\\\", \\\"format\\\": \\\"text\\\"}\",\"name\":\"fetch\"},\"type\":\"function\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"role\":\"assistant\"},{\"content\":\"HTML test file Purpose: Provide example of this file type Document file type: HTML Version: 1.0 Remark: Example content: The names \\\"John Doe\\\" for males, \\\"Jane Doe\\\" or \\\"Jane Roe\\\" for females, or \\\"Jonnie Doe\\\" and \\\"Janie Doe\\\" for children, or just \\\"Doe\\\" non-gender-specifically are used as placeholder names for a party whose true identity is unknown or must be withheld in a legal action, case, or discussion. The names are also used to refer to acorpse or hospital patient whose identity is unknown. This practice is widely used in the United States and Canada, but is rarely used in other English-speaking countries including the United Kingdom itself, from where the use of \\\"John Doe\\\" in a legal context originates. The names Joe Bloggs or John Smith are used in the UK instead, as well as in Australia and New Zealand. John Doe is sometimes used to refer to a typical male in other contexts as well, in a similar manner to John Q. Public, known in Great Britain as Joe Public, John Smith or Joe Bloggs. For example, the first name listed on a form is often John Doe, along with a fictional address or other fictional information to provide an example of how to fill in the form. The name is also used frequently in popular culture, for example in the Frank Capra film Meet John Doe. John Doe was also the name of a 2002 American television series. Similarly, a child or baby whose identity is unknown may be referred to as Baby Doe. A notorious murder case in Kansas City, Missouri, referred to the baby victim as Precious Doe. Other unidentified female murder victims are Cali Doe and Princess Doe. Additional persons may be called James Doe, Judy Doe, etc. However, to avoid possible confusion, if two anonymous or unknown parties are cited in a specific case or action, the surnames Doe and Roe may be used simultaneously; for example, \\\"John Doe v. Jane Roe\\\". If several anonymous parties are referenced, they may simply be labelled John Doe #1, John Doe #2, etc. (the U.S. Operation Delego cited 21 (numbered) \\\"John Doe\\\"s) or labelled with other variants of Doe / Roe / Poe / etc. Other early alternatives such as John Stiles and Richard Miles are now rarely used, and Mary Major has been used in some American federal cases. File created by https://www.online-convert.com More example files: https://www.online-convert.com/file-type Text of \\\"Example content\\\": Wikipedia License: Attribution-ShareAlike 3.0 Unported Feel free to use and share the file according to license above.\",\"tool_call_id\":\"fetch:0\",\"role\":\"tool\"}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258095-2hieluq6MtDo3AX0wdmC","object":"chat.completion.chunk","created":1776258095,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"Yes","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258095-2hieluq6MtDo3AX0wdmC","object":"chat.completion.chunk","created":1776258095,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":",","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258095-2hieluq6MtDo3AX0wdmC","object":"chat.completion.chunk","created":1776258095,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" the","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258095-2hieluq6MtDo3AX0wdmC","object":"chat.completion.chunk","created":1776258095,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" content","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258095-2hieluq6MtDo3AX0wdmC","object":"chat.completion.chunk","created":1776258095,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" contains","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258095-2hieluq6MtDo3AX0wdmC","object":"chat.completion.chunk","created":1776258095,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" the","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258095-2hieluq6MtDo3AX0wdmC","object":"chat.completion.chunk","created":1776258095,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" words","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258095-2hieluq6MtDo3AX0wdmC","object":"chat.completion.chunk","created":1776258095,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" \"","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258095-2hieluq6MtDo3AX0wdmC","object":"chat.completion.chunk","created":1776258095,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"John","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258095-2hieluq6MtDo3AX0wdmC","object":"chat.completion.chunk","created":1776258095,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" Doe","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258095-2hieluq6MtDo3AX0wdmC","object":"chat.completion.chunk","created":1776258095,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"\"","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258095-2hieluq6MtDo3AX0wdmC","object":"chat.completion.chunk","created":1776258095,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" multiple","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258095-2hieluq6MtDo3AX0wdmC","object":"chat.completion.chunk","created":1776258095,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" times","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258095-2hieluq6MtDo3AX0wdmC","object":"chat.completion.chunk","created":1776258095,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":".","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258095-2hieluq6MtDo3AX0wdmC","object":"chat.completion.chunk","created":1776258095,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}]}
-
- data: {"id":"gen-1776258095-2hieluq6MtDo3AX0wdmC","object":"chat.completion.chunk","created":1776258095,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}],"usage":{"prompt_tokens":7112,"completion_tokens":15,"total_tokens":7127,"cost":0.0043047,"is_byok":false,"prompt_tokens_details":{"cached_tokens":6400,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0043047,"upstream_inference_prompt_cost":0.0042672,"upstream_inference_completions_cost":0.0000375},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 1.832176292s
diff --git a/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/glob_tool.yaml b/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/glob_tool.yaml
deleted file mode 100644
index fffd1b7936b80ae907d7d63211e8a686919b6879..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/glob_tool.yaml
+++ /dev/null
@@ -1,154 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 848
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse glob to find all .go files in the current directory\n \n\n","role":"user"}],"model":"qwen/qwen3-next-80b-a3b-instruct","max_tokens":40,"stream_options":{"include_usage":true},"usage":{"include":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258097-qFYkYYtoO3xmTURMwe2b","object":"chat.completion.chunk","created":1776258097,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Google","choices":[{"index":0,"delta":{"content":"Find","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258097-qFYkYYtoO3xmTURMwe2b","object":"chat.completion.chunk","created":1776258097,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Google","choices":[{"index":0,"delta":{"content":" all .go files","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258097-qFYkYYtoO3xmTURMwe2b","object":"chat.completion.chunk","created":1776258097,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Google","choices":[{"index":0,"delta":{"content":" in current directory using","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258097-qFYkYYtoO3xmTURMwe2b","object":"chat.completion.chunk","created":1776258097,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Google","choices":[{"index":0,"delta":{"content":" glob","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258097-qFYkYYtoO3xmTURMwe2b","object":"chat.completion.chunk","created":1776258097,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Google","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}]}
-
- data: {"id":"gen-1776258097-qFYkYYtoO3xmTURMwe2b","object":"chat.completion.chunk","created":1776258097,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Google","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}],"usage":{"prompt_tokens":155,"completion_tokens":11,"total_tokens":166,"cost":0.00003645,"is_byok":false,"prompt_tokens_details":{"cached_tokens":0,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00003645,"upstream_inference_prompt_cost":0.00002325,"upstream_inference_completions_cost":0.0000132},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 883.711667ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30813
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/glob_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"use glob to find all .go files in the current directory\",\"role\":\"user\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258097-zjeEKO1xlBjemUzjPt1B","object":"chat.completion.chunk","created":1776258097,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"id":"glob:0","type":"function","function":{"name":"glob","arguments":""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258097-zjeEKO1xlBjemUzjPt1B","object":"chat.completion.chunk","created":1776258097,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258097-zjeEKO1xlBjemUzjPt1B","object":"chat.completion.chunk","created":1776258097,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"pattern"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258097-zjeEKO1xlBjemUzjPt1B","object":"chat.completion.chunk","created":1776258097,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258097-zjeEKO1xlBjemUzjPt1B","object":"chat.completion.chunk","created":1776258097,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \"*."}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258097-zjeEKO1xlBjemUzjPt1B","object":"chat.completion.chunk","created":1776258097,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"go"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258097-zjeEKO1xlBjemUzjPt1B","object":"chat.completion.chunk","created":1776258097,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\"}"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258097-zjeEKO1xlBjemUzjPt1B","object":"chat.completion.chunk","created":1776258097,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}]}
-
- data: {"id":"gen-1776258097-zjeEKO1xlBjemUzjPt1B","object":"chat.completion.chunk","created":1776258097,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}],"usage":{"prompt_tokens":6495,"completion_tokens":15,"total_tokens":6510,"cost":0.0039345,"is_byok":false,"prompt_tokens_details":{"cached_tokens":6144,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0039345,"upstream_inference_prompt_cost":0.003897,"upstream_inference_completions_cost":0.0000375},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 2.216590917s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31031
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/glob_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use glob to find all .go files in the current directory\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"glob:0\",\"function\":{\"arguments\":\"{\\\"pattern\\\": \\\"*.go\\\"}\",\"name\":\"glob\"},\"type\":\"function\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"role\":\"assistant\"},{\"content\":\"/tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/glob_tool/main.go\",\"tool_call_id\":\"glob:0\",\"role\":\"tool\"}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258099-qNZH1Mhv658RKzp7CNtT","object":"chat.completion.chunk","created":1776258099,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"main","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258099-qNZH1Mhv658RKzp7CNtT","object":"chat.completion.chunk","created":1776258099,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":".go","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258099-qNZH1Mhv658RKzp7CNtT","object":"chat.completion.chunk","created":1776258099,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}]}
-
- data: {"id":"gen-1776258099-qNZH1Mhv658RKzp7CNtT","object":"chat.completion.chunk","created":1776258099,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}],"usage":{"prompt_tokens":6544,"completion_tokens":3,"total_tokens":6547,"cost":0.0039339,"is_byok":false,"prompt_tokens_details":{"cached_tokens":6400,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0039339,"upstream_inference_prompt_cost":0.0039264,"upstream_inference_completions_cost":0.0000075},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 2.019938292s
diff --git a/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/grep_tool.yaml b/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/grep_tool.yaml
deleted file mode 100644
index 8f940584271228520062ebffa7ed8f52eeb46507..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/grep_tool.yaml
+++ /dev/null
@@ -1,172 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 846
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse grep to search for the word ''package'' in go files\n \n\n","role":"user"}],"model":"qwen/qwen3-next-80b-a3b-instruct","max_tokens":40,"stream_options":{"include_usage":true},"usage":{"include":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258101-2DN1WhDUvsStYUSP0aQr","object":"chat.completion.chunk","created":1776258101,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":"Search","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258101-2DN1WhDUvsStYUSP0aQr","object":"chat.completion.chunk","created":1776258101,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":" for '","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258101-2DN1WhDUvsStYUSP0aQr","object":"chat.completion.chunk","created":1776258101,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":"package' in Go","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258101-2DN1WhDUvsStYUSP0aQr","object":"chat.completion.chunk","created":1776258101,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":" files using grep","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258101-2DN1WhDUvsStYUSP0aQr","object":"chat.completion.chunk","created":1776258101,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}]}
-
- data: {"id":"gen-1776258101-2DN1WhDUvsStYUSP0aQr","object":"chat.completion.chunk","created":1776258101,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}],"usage":{"prompt_tokens":160,"completion_tokens":10,"total_tokens":170,"cost":0.0000234,"is_byok":false,"prompt_tokens_details":{"cached_tokens":0,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.000036,"upstream_inference_prompt_cost":0.000024,"upstream_inference_completions_cost":0.000012},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 1.59618s
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30811
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/grep_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"use grep to search for the word 'package' in go files\",\"role\":\"user\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258101-z6GogS2n8ZjAefW4nLfZ","object":"chat.completion.chunk","created":1776258101,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"id":"grep:0","type":"function","function":{"name":"grep","arguments":""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258101-z6GogS2n8ZjAefW4nLfZ","object":"chat.completion.chunk","created":1776258101,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258101-z6GogS2n8ZjAefW4nLfZ","object":"chat.completion.chunk","created":1776258101,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"pattern"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258101-z6GogS2n8ZjAefW4nLfZ","object":"chat.completion.chunk","created":1776258101,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258101-z6GogS2n8ZjAefW4nLfZ","object":"chat.completion.chunk","created":1776258101,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258101-z6GogS2n8ZjAefW4nLfZ","object":"chat.completion.chunk","created":1776258101,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"package"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258101-z6GogS2n8ZjAefW4nLfZ","object":"chat.completion.chunk","created":1776258101,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\","}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258101-z6GogS2n8ZjAefW4nLfZ","object":"chat.completion.chunk","created":1776258101,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258101-z6GogS2n8ZjAefW4nLfZ","object":"chat.completion.chunk","created":1776258101,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"include"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258101-z6GogS2n8ZjAefW4nLfZ","object":"chat.completion.chunk","created":1776258101,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258101-z6GogS2n8ZjAefW4nLfZ","object":"chat.completion.chunk","created":1776258101,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \"*."}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258101-z6GogS2n8ZjAefW4nLfZ","object":"chat.completion.chunk","created":1776258101,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"go"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258101-z6GogS2n8ZjAefW4nLfZ","object":"chat.completion.chunk","created":1776258101,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\"}"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258101-z6GogS2n8ZjAefW4nLfZ","object":"chat.completion.chunk","created":1776258101,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}]}
-
- data: {"id":"gen-1776258101-z6GogS2n8ZjAefW4nLfZ","object":"chat.completion.chunk","created":1776258101,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}],"usage":{"prompt_tokens":6496,"completion_tokens":22,"total_tokens":6518,"cost":0.0039526,"is_byok":false,"prompt_tokens_details":{"cached_tokens":6144,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0039526,"upstream_inference_prompt_cost":0.0038976,"upstream_inference_completions_cost":0.000055},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 2.309879042s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31107
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/grep_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use grep to search for the word 'package' in go files\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"grep:0\",\"function\":{\"arguments\":\"{\\\"pattern\\\": \\\"package\\\", \\\"include\\\": \\\"*.go\\\"}\",\"name\":\"grep\"},\"type\":\"function\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"role\":\"assistant\"},{\"content\":\"Found 1 matches\\n/tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/grep_tool/main.go:\\n Line 1, Char 1: package main\\n\",\"tool_call_id\":\"grep:0\",\"role\":\"tool\"}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258104-mKizMydVUExXcqARngMn","object":"chat.completion.chunk","created":1776258104,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"main","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258104-mKizMydVUExXcqARngMn","object":"chat.completion.chunk","created":1776258104,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":".go","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258104-mKizMydVUExXcqARngMn","object":"chat.completion.chunk","created":1776258104,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":":","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258104-mKizMydVUExXcqARngMn","object":"chat.completion.chunk","created":1776258104,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"1","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258104-mKizMydVUExXcqARngMn","object":"chat.completion.chunk","created":1776258104,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}]}
-
- data: {"id":"gen-1776258104-mKizMydVUExXcqARngMn","object":"chat.completion.chunk","created":1776258104,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}],"usage":{"prompt_tokens":6571,"completion_tokens":5,"total_tokens":6576,"cost":0.0039551,"is_byok":false,"prompt_tokens_details":{"cached_tokens":6400,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0039551,"upstream_inference_prompt_cost":0.0039426,"upstream_inference_completions_cost":0.0000125},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 1.494037625s
diff --git a/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/ls_tool.yaml b/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/ls_tool.yaml
deleted file mode 100644
index dba40afa6dda00988c0ee9e035c22a478e7c0883..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/ls_tool.yaml
+++ /dev/null
@@ -1,208 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 842
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse ls to list the files in the current directory\n \n\n","role":"user"}],"model":"qwen/qwen3-next-80b-a3b-instruct","max_tokens":40,"stream_options":{"include_usage":true},"usage":{"include":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258105-JakSMN7G85CTKecg6k5G","object":"chat.completion.chunk","created":1776258105,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"DeepInfra","choices":[{"index":0,"delta":{"content":"List","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258105-JakSMN7G85CTKecg6k5G","object":"chat.completion.chunk","created":1776258105,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"DeepInfra","choices":[{"index":0,"delta":{"content":" files in current directory","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258105-JakSMN7G85CTKecg6k5G","object":"chat.completion.chunk","created":1776258105,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"DeepInfra","choices":[{"index":0,"delta":{"content":" using ls","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258105-JakSMN7G85CTKecg6k5G","object":"chat.completion.chunk","created":1776258105,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"DeepInfra","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}]}
-
- data: {"id":"gen-1776258105-JakSMN7G85CTKecg6k5G","object":"chat.completion.chunk","created":1776258105,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"DeepInfra","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}],"usage":{"prompt_tokens":153,"completion_tokens":8,"total_tokens":161,"cost":0.00002257,"is_byok":false,"prompt_tokens_details":{"cached_tokens":0,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00002257,"upstream_inference_prompt_cost":0.00001377,"upstream_inference_completions_cost":0.0000088},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 874.434167ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30805
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/ls_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"use ls to list the files in the current directory\",\"role\":\"user\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258105-Ku2lKE2plz9uYkJMbFNx","object":"chat.completion.chunk","created":1776258105,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"id":"ls:0","type":"function","function":{"name":"ls","arguments":""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258105-Ku2lKE2plz9uYkJMbFNx","object":"chat.completion.chunk","created":1776258105,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258105-Ku2lKE2plz9uYkJMbFNx","object":"chat.completion.chunk","created":1776258105,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"path"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258105-Ku2lKE2plz9uYkJMbFNx","object":"chat.completion.chunk","created":1776258105,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258105-Ku2lKE2plz9uYkJMbFNx","object":"chat.completion.chunk","created":1776258105,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \"/"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258105-Ku2lKE2plz9uYkJMbFNx","object":"chat.completion.chunk","created":1776258105,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"tmp"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258105-Ku2lKE2plz9uYkJMbFNx","object":"chat.completion.chunk","created":1776258105,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/c"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258105-Ku2lKE2plz9uYkJMbFNx","object":"chat.completion.chunk","created":1776258105,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"rush"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258105-Ku2lKE2plz9uYkJMbFNx","object":"chat.completion.chunk","created":1776258105,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-test"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258105-Ku2lKE2plz9uYkJMbFNx","object":"chat.completion.chunk","created":1776258105,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/Test"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258105-Ku2lKE2plz9uYkJMbFNx","object":"chat.completion.chunk","created":1776258105,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"Coder"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258105-Ku2lKE2plz9uYkJMbFNx","object":"chat.completion.chunk","created":1776258105,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"Agent"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258105-Ku2lKE2plz9uYkJMbFNx","object":"chat.completion.chunk","created":1776258105,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/open"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258105-Ku2lKE2plz9uYkJMbFNx","object":"chat.completion.chunk","created":1776258105,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"router"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258105-Ku2lKE2plz9uYkJMbFNx","object":"chat.completion.chunk","created":1776258105,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-k"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258105-Ku2lKE2plz9uYkJMbFNx","object":"chat.completion.chunk","created":1776258105,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"imi"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258105-Ku2lKE2plz9uYkJMbFNx","object":"chat.completion.chunk","created":1776258105,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-k"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258105-Ku2lKE2plz9uYkJMbFNx","object":"chat.completion.chunk","created":1776258105,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"2"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258105-Ku2lKE2plz9uYkJMbFNx","object":"chat.completion.chunk","created":1776258105,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258105-Ku2lKE2plz9uYkJMbFNx","object":"chat.completion.chunk","created":1776258105,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"ls"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258105-Ku2lKE2plz9uYkJMbFNx","object":"chat.completion.chunk","created":1776258105,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"_tool"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258105-Ku2lKE2plz9uYkJMbFNx","object":"chat.completion.chunk","created":1776258105,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\"}"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258105-Ku2lKE2plz9uYkJMbFNx","object":"chat.completion.chunk","created":1776258105,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}]}
-
- data: {"id":"gen-1776258105-Ku2lKE2plz9uYkJMbFNx","object":"chat.completion.chunk","created":1776258105,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}],"usage":{"prompt_tokens":6493,"completion_tokens":31,"total_tokens":6524,"cost":0.0039733,"is_byok":false,"prompt_tokens_details":{"cached_tokens":6144,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0039733,"upstream_inference_prompt_cost":0.0038958,"upstream_inference_completions_cost":0.0000775},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 3.195814292s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31089
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/ls_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use ls to list the files in the current directory\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"ls:0\",\"function\":{\"arguments\":\"{\\\"path\\\": \\\"/tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/ls_tool\\\"}\",\"name\":\"ls\"},\"type\":\"function\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"role\":\"assistant\"},{\"content\":\"\\n- /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/ls_tool/\\n - go.mod\\n - main.go\\n\",\"tool_call_id\":\"ls:0\",\"role\":\"tool\"}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258109-gkBQlsUBFrg2vwa8BdIm","object":"chat.completion.chunk","created":1776258109,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"The","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258109-gkBQlsUBFrg2vwa8BdIm","object":"chat.completion.chunk","created":1776258109,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" current","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258109-gkBQlsUBFrg2vwa8BdIm","object":"chat.completion.chunk","created":1776258109,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" directory","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258109-gkBQlsUBFrg2vwa8BdIm","object":"chat.completion.chunk","created":1776258109,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" contains","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258109-gkBQlsUBFrg2vwa8BdIm","object":"chat.completion.chunk","created":1776258109,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":":\n","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258109-gkBQlsUBFrg2vwa8BdIm","object":"chat.completion.chunk","created":1776258109,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"-","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258109-gkBQlsUBFrg2vwa8BdIm","object":"chat.completion.chunk","created":1776258109,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" go","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258109-gkBQlsUBFrg2vwa8BdIm","object":"chat.completion.chunk","created":1776258109,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":".mod","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258109-gkBQlsUBFrg2vwa8BdIm","object":"chat.completion.chunk","created":1776258109,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"\n","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258109-gkBQlsUBFrg2vwa8BdIm","object":"chat.completion.chunk","created":1776258109,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"-","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258109-gkBQlsUBFrg2vwa8BdIm","object":"chat.completion.chunk","created":1776258109,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" main","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258109-gkBQlsUBFrg2vwa8BdIm","object":"chat.completion.chunk","created":1776258109,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":".go","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258109-gkBQlsUBFrg2vwa8BdIm","object":"chat.completion.chunk","created":1776258109,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}]}
-
- data: {"id":"gen-1776258109-gkBQlsUBFrg2vwa8BdIm","object":"chat.completion.chunk","created":1776258109,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}],"usage":{"prompt_tokens":6570,"completion_tokens":13,"total_tokens":6583,"cost":0.0039745,"is_byok":false,"prompt_tokens_details":{"cached_tokens":6400,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0039745,"upstream_inference_prompt_cost":0.003942,"upstream_inference_completions_cost":0.0000325},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 1.938460542s
diff --git a/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/multiedit_tool.yaml b/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/multiedit_tool.yaml
deleted file mode 100644
index e434f41ef81e8e75479b7f8de19dc3fb9114be68..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/multiedit_tool.yaml
+++ /dev/null
@@ -1,736 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 921
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse multiedit to change ''Hello, World!'' to ''Hello, Crush!'' and add a comment ''// Greeting'' above the fmt.Println line in main.go\n \n\n","role":"user"}],"model":"qwen/qwen3-next-80b-a3b-instruct","max_tokens":40,"stream_options":{"include_usage":true},"usage":{"include":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258111-goltDQrdyo9oIvehdopW","object":"chat.completion.chunk","created":1776258111,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"AtlasCloud","choices":[{"index":0,"delta":{"content":"Use","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-goltDQrdyo9oIvehdopW","object":"chat.completion.chunk","created":1776258111,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"AtlasCloud","choices":[{"index":0,"delta":{"content":" mult","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-goltDQrdyo9oIvehdopW","object":"chat.completion.chunk","created":1776258111,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"AtlasCloud","choices":[{"index":0,"delta":{"content":"ied","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-goltDQrdyo9oIvehdopW","object":"chat.completion.chunk","created":1776258111,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"AtlasCloud","choices":[{"index":0,"delta":{"content":"it","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-goltDQrdyo9oIvehdopW","object":"chat.completion.chunk","created":1776258111,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"AtlasCloud","choices":[{"index":0,"delta":{"content":" to","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-goltDQrdyo9oIvehdopW","object":"chat.completion.chunk","created":1776258111,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"AtlasCloud","choices":[{"index":0,"delta":{"content":" update","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-goltDQrdyo9oIvehdopW","object":"chat.completion.chunk","created":1776258111,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"AtlasCloud","choices":[{"index":0,"delta":{"content":" greeting","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-goltDQrdyo9oIvehdopW","object":"chat.completion.chunk","created":1776258111,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"AtlasCloud","choices":[{"index":0,"delta":{"content":" and","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-goltDQrdyo9oIvehdopW","object":"chat.completion.chunk","created":1776258111,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"AtlasCloud","choices":[{"index":0,"delta":{"content":" add","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-goltDQrdyo9oIvehdopW","object":"chat.completion.chunk","created":1776258111,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"AtlasCloud","choices":[{"index":0,"delta":{"content":" comment","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-goltDQrdyo9oIvehdopW","object":"chat.completion.chunk","created":1776258111,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"AtlasCloud","choices":[{"index":0,"delta":{"content":" in","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-goltDQrdyo9oIvehdopW","object":"chat.completion.chunk","created":1776258111,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"AtlasCloud","choices":[{"index":0,"delta":{"content":" main","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-goltDQrdyo9oIvehdopW","object":"chat.completion.chunk","created":1776258111,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"AtlasCloud","choices":[{"index":0,"delta":{"content":".go","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-goltDQrdyo9oIvehdopW","object":"chat.completion.chunk","created":1776258111,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"AtlasCloud","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}]}
-
- data: {"id":"gen-1776258111-goltDQrdyo9oIvehdopW","object":"chat.completion.chunk","created":1776258111,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"AtlasCloud","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}],"usage":{"prompt_tokens":176,"completion_tokens":14,"total_tokens":190,"cost":0.0000474,"is_byok":false,"prompt_tokens_details":{"cached_tokens":0,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0000474,"upstream_inference_prompt_cost":0.0000264,"upstream_inference_completions_cost":0.000021},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 1.152553833s
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30891
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/multiedit_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"use multiedit to change 'Hello, World!' to 'Hello, Crush!' and add a comment '// Greeting' above the fmt.Println line in main.go\",\"role\":\"user\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"I'll","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" help","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" you","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" use","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" mult","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"ied","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"it","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" to","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" make","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" those","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" changes","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" to","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" main","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":".go","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":".","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" First","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":",","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" let","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" me","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" read","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" the","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" current","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" file","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" to","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" see","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" its","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" structure","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" and","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" then","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" apply","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" the","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" changes","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":".","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"id":"view:0","type":"function","function":{"name":"view","arguments":""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"file"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"_path"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \"/"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"tmp"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/c"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"rush"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-test"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/Test"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"Coder"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"Agent"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/open"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"router"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-k"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"imi"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-k"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"2"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/mult"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"ied"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"it"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"_tool"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/main"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":".go"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\"}"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}]}
-
- data: {"id":"gen-1776258111-0snDnTg7Q3UiQLl1yGDL","object":"chat.completion.chunk","created":1776258111,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}],"usage":{"prompt_tokens":6517,"completion_tokens":67,"total_tokens":6584,"cost":0.0040777,"is_byok":false,"prompt_tokens_details":{"cached_tokens":6144,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0040777,"upstream_inference_prompt_cost":0.0039102,"upstream_inference_completions_cost":0.0001675},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 1.682060708s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31423
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/multiedit_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use multiedit to change 'Hello, World!' to 'Hello, Crush!' and add a comment '// Greeting' above the fmt.Println line in main.go\",\"role\":\"user\"},{\"content\":\"I'll help you use multiedit to make those changes to main.go. First, let me read the current file to see its structure and then apply the changes.\",\"tool_calls\":[{\"id\":\"view:0\",\"function\":{\"arguments\":\"{\\\"file_path\\\": \\\"/tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/multiedit_tool/main.go\\\"}\",\"name\":\"view\"},\"type\":\"function\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"role\":\"assistant\"},{\"content\":\"\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\tfmt.Println(\\\"Hello, World!\\\")\\n 7|}\\n\\n\",\"tool_call_id\":\"view:0\",\"role\":\"tool\"}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"Now","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" I'll","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" use","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" mult","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"ied","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"it","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" to","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" make","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" both","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" changes","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" at","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" once","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" -","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" adding","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" the","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" comment","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" and","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" changing","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" the","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" string","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":":","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"id":"multiedit:1","type":"function","function":{"name":"multiedit","arguments":""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"file"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"_path"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \"/"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"tmp"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/c"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"rush"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-test"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/Test"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"Coder"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"Agent"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/open"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"router"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-k"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"imi"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-k"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"2"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/mult"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"ied"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"it"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"_tool"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/main"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":".go"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\","}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"ed"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"its"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" [{\""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"old"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"_string"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"func"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" main"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"()"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" {\\"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"n"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\\t"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"fmt"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":".Println"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"(\\\""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"Hello"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":","}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" World"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"!\\"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\")\\"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"n"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"}\","}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"new"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"_string"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"func"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" main"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"()"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" {\\"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"n"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\\t"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"//"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" G"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"reeting"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\\n"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\\t"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"fmt"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":".Println"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"(\\\""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"Hello"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":","}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" Crush"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"!\\"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\")\\"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"n"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"}\""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"}]}"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}]}
-
- data: {"id":"gen-1776258116-rGzHJlKQgReeIS0q6jnx","object":"chat.completion.chunk","created":1776258116,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}],"usage":{"prompt_tokens":6652,"completion_tokens":109,"total_tokens":6761,"cost":0.0042637,"is_byok":false,"prompt_tokens_details":{"cached_tokens":6400,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0042637,"upstream_inference_prompt_cost":0.0039912,"upstream_inference_completions_cost":0.0002725},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 1.67491125s
-- id: 3
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 32116
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/multiedit_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use multiedit to change 'Hello, World!' to 'Hello, Crush!' and add a comment '// Greeting' above the fmt.Println line in main.go\",\"role\":\"user\"},{\"content\":\"I'll help you use multiedit to make those changes to main.go. First, let me read the current file to see its structure and then apply the changes.\",\"tool_calls\":[{\"id\":\"view:0\",\"function\":{\"arguments\":\"{\\\"file_path\\\": \\\"/tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/multiedit_tool/main.go\\\"}\",\"name\":\"view\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\tfmt.Println(\\\"Hello, World!\\\")\\n 7|}\\n\\n\",\"tool_call_id\":\"view:0\",\"role\":\"tool\"},{\"content\":\"Now I'll use multiedit to make both changes at once - adding the comment and changing the string:\",\"tool_calls\":[{\"id\":\"multiedit:1\",\"function\":{\"arguments\":\"{\\\"file_path\\\": \\\"/tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/multiedit_tool/main.go\\\", \\\"edits\\\": [{\\\"old_string\\\": \\\"func main() {\\\\n\\\\tfmt.Println(\\\\\\\"Hello, World!\\\\\\\")\\\\n}\\\", \\\"new_string\\\": \\\"func main() {\\\\n\\\\t// Greeting\\\\n\\\\tfmt.Println(\\\\\\\"Hello, Crush!\\\\\\\")\\\\n}\\\"}]}\",\"name\":\"multiedit\"},\"type\":\"function\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"role\":\"assistant\"},{\"content\":\"\\nApplied 1 edits to file: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/multiedit_tool/main.go\\n\\n\",\"tool_call_id\":\"multiedit:1\",\"role\":\"tool\"}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258124-c576nyhO9ergah1jD8sX","object":"chat.completion.chunk","created":1776258124,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"id":"view:2","type":"function","function":{"name":"view","arguments":""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258124-c576nyhO9ergah1jD8sX","object":"chat.completion.chunk","created":1776258124,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258124-c576nyhO9ergah1jD8sX","object":"chat.completion.chunk","created":1776258124,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"file"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258124-c576nyhO9ergah1jD8sX","object":"chat.completion.chunk","created":1776258124,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"_path"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258124-c576nyhO9ergah1jD8sX","object":"chat.completion.chunk","created":1776258124,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258124-c576nyhO9ergah1jD8sX","object":"chat.completion.chunk","created":1776258124,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \"/"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258124-c576nyhO9ergah1jD8sX","object":"chat.completion.chunk","created":1776258124,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"tmp"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258124-c576nyhO9ergah1jD8sX","object":"chat.completion.chunk","created":1776258124,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/c"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258124-c576nyhO9ergah1jD8sX","object":"chat.completion.chunk","created":1776258124,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"rush"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258124-c576nyhO9ergah1jD8sX","object":"chat.completion.chunk","created":1776258124,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-test"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258124-c576nyhO9ergah1jD8sX","object":"chat.completion.chunk","created":1776258124,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/Test"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258124-c576nyhO9ergah1jD8sX","object":"chat.completion.chunk","created":1776258124,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"Coder"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258124-c576nyhO9ergah1jD8sX","object":"chat.completion.chunk","created":1776258124,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"Agent"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258124-c576nyhO9ergah1jD8sX","object":"chat.completion.chunk","created":1776258124,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/open"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258124-c576nyhO9ergah1jD8sX","object":"chat.completion.chunk","created":1776258124,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"router"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258124-c576nyhO9ergah1jD8sX","object":"chat.completion.chunk","created":1776258124,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-k"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258124-c576nyhO9ergah1jD8sX","object":"chat.completion.chunk","created":1776258124,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"imi"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258124-c576nyhO9ergah1jD8sX","object":"chat.completion.chunk","created":1776258124,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-k"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258124-c576nyhO9ergah1jD8sX","object":"chat.completion.chunk","created":1776258124,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"2"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258124-c576nyhO9ergah1jD8sX","object":"chat.completion.chunk","created":1776258124,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/mult"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258124-c576nyhO9ergah1jD8sX","object":"chat.completion.chunk","created":1776258124,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"ied"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258124-c576nyhO9ergah1jD8sX","object":"chat.completion.chunk","created":1776258124,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"it"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258124-c576nyhO9ergah1jD8sX","object":"chat.completion.chunk","created":1776258124,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"_tool"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258124-c576nyhO9ergah1jD8sX","object":"chat.completion.chunk","created":1776258124,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/main"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258124-c576nyhO9ergah1jD8sX","object":"chat.completion.chunk","created":1776258124,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":".go"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258124-c576nyhO9ergah1jD8sX","object":"chat.completion.chunk","created":1776258124,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\"}"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258124-c576nyhO9ergah1jD8sX","object":"chat.completion.chunk","created":1776258124,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}]}
-
- data: {"id":"gen-1776258124-c576nyhO9ergah1jD8sX","object":"chat.completion.chunk","created":1776258124,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}],"usage":{"prompt_tokens":6813,"completion_tokens":34,"total_tokens":6847,"cost":0.0041728,"is_byok":false,"prompt_tokens_details":{"cached_tokens":6400,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0041728,"upstream_inference_prompt_cost":0.0040878,"upstream_inference_completions_cost":0.000085},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 3.518259625s
-- id: 4
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 32548
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/multiedit_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use multiedit to change 'Hello, World!' to 'Hello, Crush!' and add a comment '// Greeting' above the fmt.Println line in main.go\",\"role\":\"user\"},{\"content\":\"I'll help you use multiedit to make those changes to main.go. First, let me read the current file to see its structure and then apply the changes.\",\"tool_calls\":[{\"id\":\"view:0\",\"function\":{\"arguments\":\"{\\\"file_path\\\": \\\"/tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/multiedit_tool/main.go\\\"}\",\"name\":\"view\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\tfmt.Println(\\\"Hello, World!\\\")\\n 7|}\\n\\n\",\"tool_call_id\":\"view:0\",\"role\":\"tool\"},{\"content\":\"Now I'll use multiedit to make both changes at once - adding the comment and changing the string:\",\"tool_calls\":[{\"id\":\"multiedit:1\",\"function\":{\"arguments\":\"{\\\"file_path\\\": \\\"/tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/multiedit_tool/main.go\\\", \\\"edits\\\": [{\\\"old_string\\\": \\\"func main() {\\\\n\\\\tfmt.Println(\\\\\\\"Hello, World!\\\\\\\")\\\\n}\\\", \\\"new_string\\\": \\\"func main() {\\\\n\\\\t// Greeting\\\\n\\\\tfmt.Println(\\\\\\\"Hello, Crush!\\\\\\\")\\\\n}\\\"}]}\",\"name\":\"multiedit\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\nApplied 1 edits to file: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/multiedit_tool/main.go\\n\\n\",\"tool_call_id\":\"multiedit:1\",\"role\":\"tool\"},{\"tool_calls\":[{\"id\":\"view:2\",\"function\":{\"arguments\":\"{\\\"file_path\\\": \\\"/tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/multiedit_tool/main.go\\\"}\",\"name\":\"view\"},\"type\":\"function\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"role\":\"assistant\"},{\"content\":\"\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\t// Greeting\\n 7|\\tfmt.Println(\\\"Hello, Crush!\\\")\\n 8|}\\n\\n\",\"tool_call_id\":\"view:2\",\"role\":\"tool\"}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"Done","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"!","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" I've","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" successfully","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" used","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" mult","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"ied","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"it","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" to","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":":\n","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"1","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":".","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" Add","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" the","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" comment","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" `","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"//","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" G","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"reeting","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"`","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" above","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" the","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" fmt","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":".Println","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" line","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"\n","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"2","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":".","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" Change","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" the","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" string","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" from","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" '","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"Hello","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":",","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" World","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"!'","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" to","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" '","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"Hello","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":",","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" Crush","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"!'","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}]}
-
- data: {"id":"gen-1776258128-eA3FB1HfMFdQLTJ3LA9g","object":"chat.completion.chunk","created":1776258128,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}],"usage":{"prompt_tokens":6924,"completion_tokens":44,"total_tokens":6968,"cost":0.0042644,"is_byok":false,"prompt_tokens_details":{"cached_tokens":6656,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0042644,"upstream_inference_prompt_cost":0.0041544,"upstream_inference_completions_cost":0.00011},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 1.754073666s
diff --git a/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/parallel_tool_calls.yaml b/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/parallel_tool_calls.yaml
deleted file mode 100644
index 84161cf73ed879cde9705aa476cc53a6ccf2daf4..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/parallel_tool_calls.yaml
+++ /dev/null
@@ -1,226 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 927
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse glob to find all .go files and use ls to list the current directory, it is very important that you run both tool calls in parallel\n \n\n","role":"user"}],"model":"qwen/qwen3-next-80b-a3b-instruct","max_tokens":40,"stream_options":{"include_usage":true},"usage":{"include":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258151-AsOfoCF0UigLUkEqX8dE","object":"chat.completion.chunk","created":1776258151,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Parasail","choices":[{"index":0,"delta":{"content":"Run","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258151-AsOfoCF0UigLUkEqX8dE","object":"chat.completion.chunk","created":1776258151,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Parasail","choices":[{"index":0,"delta":{"content":" glob","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258151-AsOfoCF0UigLUkEqX8dE","object":"chat.completion.chunk","created":1776258151,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Parasail","choices":[{"index":0,"delta":{"content":" and","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258151-AsOfoCF0UigLUkEqX8dE","object":"chat.completion.chunk","created":1776258151,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Parasail","choices":[{"index":0,"delta":{"content":" ls","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258151-AsOfoCF0UigLUkEqX8dE","object":"chat.completion.chunk","created":1776258151,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Parasail","choices":[{"index":0,"delta":{"content":" in","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258151-AsOfoCF0UigLUkEqX8dE","object":"chat.completion.chunk","created":1776258151,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Parasail","choices":[{"index":0,"delta":{"content":" parallel","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258151-AsOfoCF0UigLUkEqX8dE","object":"chat.completion.chunk","created":1776258151,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Parasail","choices":[{"index":0,"delta":{"content":" to","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258151-AsOfoCF0UigLUkEqX8dE","object":"chat.completion.chunk","created":1776258151,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Parasail","choices":[{"index":0,"delta":{"content":" find","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258151-AsOfoCF0UigLUkEqX8dE","object":"chat.completion.chunk","created":1776258151,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Parasail","choices":[{"index":0,"delta":{"content":" .","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258151-AsOfoCF0UigLUkEqX8dE","object":"chat.completion.chunk","created":1776258151,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Parasail","choices":[{"index":0,"delta":{"content":"go","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258151-AsOfoCF0UigLUkEqX8dE","object":"chat.completion.chunk","created":1776258151,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Parasail","choices":[{"index":0,"delta":{"content":" files","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258151-AsOfoCF0UigLUkEqX8dE","object":"chat.completion.chunk","created":1776258151,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Parasail","choices":[{"index":0,"delta":{"content":" and","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258151-AsOfoCF0UigLUkEqX8dE","object":"chat.completion.chunk","created":1776258151,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Parasail","choices":[{"index":0,"delta":{"content":" list","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258151-AsOfoCF0UigLUkEqX8dE","object":"chat.completion.chunk","created":1776258151,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Parasail","choices":[{"index":0,"delta":{"content":" directory","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258151-AsOfoCF0UigLUkEqX8dE","object":"chat.completion.chunk","created":1776258151,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Parasail","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}]}
-
- data: {"id":"gen-1776258151-AsOfoCF0UigLUkEqX8dE","object":"chat.completion.chunk","created":1776258151,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Parasail","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}],"usage":{"prompt_tokens":172,"completion_tokens":15,"total_tokens":187,"cost":0.0000337,"is_byok":false,"prompt_tokens_details":{"cached_tokens":0,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0000337,"upstream_inference_prompt_cost":0.0000172,"upstream_inference_completions_cost":0.0000165},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 892.2825ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30902
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/parallel_tool_calls\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"use glob to find all .go files and use ls to list the current directory, it is very important that you run both tool calls in parallel\",\"role\":\"user\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258151-GDK6cyhEDGytzNZYooML","object":"chat.completion.chunk","created":1776258151,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"id":"glob:0","type":"function","function":{"name":"glob","arguments":""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258151-GDK6cyhEDGytzNZYooML","object":"chat.completion.chunk","created":1776258151,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258151-GDK6cyhEDGytzNZYooML","object":"chat.completion.chunk","created":1776258151,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"pattern"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258151-GDK6cyhEDGytzNZYooML","object":"chat.completion.chunk","created":1776258151,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258151-GDK6cyhEDGytzNZYooML","object":"chat.completion.chunk","created":1776258151,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \"*."}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258151-GDK6cyhEDGytzNZYooML","object":"chat.completion.chunk","created":1776258151,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"go"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258151-GDK6cyhEDGytzNZYooML","object":"chat.completion.chunk","created":1776258151,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\"}"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258151-GDK6cyhEDGytzNZYooML","object":"chat.completion.chunk","created":1776258151,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":1,"id":"ls:1","type":"function","function":{"name":"ls","arguments":""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258151-GDK6cyhEDGytzNZYooML","object":"chat.completion.chunk","created":1776258151,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":1,"function":{"arguments":"{}"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258151-GDK6cyhEDGytzNZYooML","object":"chat.completion.chunk","created":1776258151,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}]}
-
- data: {"id":"gen-1776258151-GDK6cyhEDGytzNZYooML","object":"chat.completion.chunk","created":1776258151,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}],"usage":{"prompt_tokens":6513,"completion_tokens":24,"total_tokens":6537,"cost":0.0039678,"is_byok":false,"prompt_tokens_details":{"cached_tokens":6144,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0039678,"upstream_inference_prompt_cost":0.0039078,"upstream_inference_completions_cost":0.00006},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 2.692870584s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31356
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/parallel_tool_calls\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use glob to find all .go files and use ls to list the current directory, it is very important that you run both tool calls in parallel\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"glob:0\",\"function\":{\"arguments\":\"{\\\"pattern\\\": \\\"*.go\\\"}\",\"name\":\"glob\"},\"type\":\"function\"},{\"id\":\"ls:1\",\"function\":{\"arguments\":\"{}\",\"name\":\"ls\"},\"type\":\"function\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"role\":\"assistant\"},{\"content\":\"/tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/parallel_tool_calls/main.go\",\"tool_call_id\":\"glob:0\",\"role\":\"tool\"},{\"content\":\"\\n- /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/parallel_tool_calls/\\n - go.mod\\n - main.go\\n\",\"tool_call_id\":\"ls:1\",\"role\":\"tool\"}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258154-NNWHgVGytwMikj8hg6zv","object":"chat.completion.chunk","created":1776258154,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"Done","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258154-NNWHgVGytwMikj8hg6zv","object":"chat.completion.chunk","created":1776258154,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":".","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258154-NNWHgVGytwMikj8hg6zv","object":"chat.completion.chunk","created":1776258154,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" Found","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258154-NNWHgVGytwMikj8hg6zv","object":"chat.completion.chunk","created":1776258154,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" ","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258154-NNWHgVGytwMikj8hg6zv","object":"chat.completion.chunk","created":1776258154,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"1","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258154-NNWHgVGytwMikj8hg6zv","object":"chat.completion.chunk","created":1776258154,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" .","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258154-NNWHgVGytwMikj8hg6zv","object":"chat.completion.chunk","created":1776258154,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"go","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258154-NNWHgVGytwMikj8hg6zv","object":"chat.completion.chunk","created":1776258154,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" file","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258154-NNWHgVGytwMikj8hg6zv","object":"chat.completion.chunk","created":1776258154,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":":","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258154-NNWHgVGytwMikj8hg6zv","object":"chat.completion.chunk","created":1776258154,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" `","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258154-NNWHgVGytwMikj8hg6zv","object":"chat.completion.chunk","created":1776258154,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"main","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258154-NNWHgVGytwMikj8hg6zv","object":"chat.completion.chunk","created":1776258154,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":".go","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258154-NNWHgVGytwMikj8hg6zv","object":"chat.completion.chunk","created":1776258154,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"`.","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258154-NNWHgVGytwMikj8hg6zv","object":"chat.completion.chunk","created":1776258154,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" The","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258154-NNWHgVGytwMikj8hg6zv","object":"chat.completion.chunk","created":1776258154,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" directory","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258154-NNWHgVGytwMikj8hg6zv","object":"chat.completion.chunk","created":1776258154,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" contains","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258154-NNWHgVGytwMikj8hg6zv","object":"chat.completion.chunk","created":1776258154,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" `","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258154-NNWHgVGytwMikj8hg6zv","object":"chat.completion.chunk","created":1776258154,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"go","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258154-NNWHgVGytwMikj8hg6zv","object":"chat.completion.chunk","created":1776258154,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":".mod","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258154-NNWHgVGytwMikj8hg6zv","object":"chat.completion.chunk","created":1776258154,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"`","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258154-NNWHgVGytwMikj8hg6zv","object":"chat.completion.chunk","created":1776258154,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" and","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258154-NNWHgVGytwMikj8hg6zv","object":"chat.completion.chunk","created":1776258154,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" `","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258154-NNWHgVGytwMikj8hg6zv","object":"chat.completion.chunk","created":1776258154,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"main","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258154-NNWHgVGytwMikj8hg6zv","object":"chat.completion.chunk","created":1776258154,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":".go","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258154-NNWHgVGytwMikj8hg6zv","object":"chat.completion.chunk","created":1776258154,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"`.","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258154-NNWHgVGytwMikj8hg6zv","object":"chat.completion.chunk","created":1776258154,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}]}
-
- data: {"id":"gen-1776258154-NNWHgVGytwMikj8hg6zv","object":"chat.completion.chunk","created":1776258154,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}],"usage":{"prompt_tokens":6615,"completion_tokens":26,"total_tokens":6641,"cost":0.004034,"is_byok":false,"prompt_tokens_details":{"cached_tokens":6400,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.004034,"upstream_inference_prompt_cost":0.003969,"upstream_inference_completions_cost":0.000065},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 1.726075333s
diff --git a/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/read_a_file.yaml b/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/read_a_file.yaml
deleted file mode 100644
index 547c83868bea7f2c9a739bd1941545da3253a47f..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/read_a_file.yaml
+++ /dev/null
@@ -1,317 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 808
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nRead the go mod\n \n\n","role":"user"}],"model":"qwen/qwen3-next-80b-a3b-instruct","max_tokens":40,"stream_options":{"include_usage":true},"usage":{"include":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258048-mA5p6M64yZpcAEtaPS4m","object":"chat.completion.chunk","created":1776258048,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":"Read","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258048-mA5p6M64yZpcAEtaPS4m","object":"chat.completion.chunk","created":1776258048,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":" the go mod","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258048-mA5p6M64yZpcAEtaPS4m","object":"chat.completion.chunk","created":1776258048,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}]}
-
- data: {"id":"gen-1776258048-mA5p6M64yZpcAEtaPS4m","object":"chat.completion.chunk","created":1776258048,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}],"usage":{"prompt_tokens":151,"completion_tokens":4,"total_tokens":155,"cost":0.0000178425,"is_byok":false,"prompt_tokens_details":{"cached_tokens":0,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00002745,"upstream_inference_prompt_cost":0.00002265,"upstream_inference_completions_cost":0.0000048},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 2.776948s
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30775
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/read_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"Read the go mod\",\"role\":\"user\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258048-mcvz53EK72OW04pZLWp0","object":"chat.completion.chunk","created":1776258048,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"id":"ls:0","type":"function","function":{"name":"ls","arguments":""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258048-mcvz53EK72OW04pZLWp0","object":"chat.completion.chunk","created":1776258048,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258048-mcvz53EK72OW04pZLWp0","object":"chat.completion.chunk","created":1776258048,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"path"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258048-mcvz53EK72OW04pZLWp0","object":"chat.completion.chunk","created":1776258048,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258048-mcvz53EK72OW04pZLWp0","object":"chat.completion.chunk","created":1776258048,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \"/"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258048-mcvz53EK72OW04pZLWp0","object":"chat.completion.chunk","created":1776258048,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"tmp"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258048-mcvz53EK72OW04pZLWp0","object":"chat.completion.chunk","created":1776258048,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/c"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258048-mcvz53EK72OW04pZLWp0","object":"chat.completion.chunk","created":1776258048,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"rush"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258048-mcvz53EK72OW04pZLWp0","object":"chat.completion.chunk","created":1776258048,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-test"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258048-mcvz53EK72OW04pZLWp0","object":"chat.completion.chunk","created":1776258048,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/Test"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258048-mcvz53EK72OW04pZLWp0","object":"chat.completion.chunk","created":1776258048,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"Coder"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258048-mcvz53EK72OW04pZLWp0","object":"chat.completion.chunk","created":1776258048,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"Agent"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258048-mcvz53EK72OW04pZLWp0","object":"chat.completion.chunk","created":1776258048,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/open"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258048-mcvz53EK72OW04pZLWp0","object":"chat.completion.chunk","created":1776258048,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"router"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258048-mcvz53EK72OW04pZLWp0","object":"chat.completion.chunk","created":1776258048,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-k"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258048-mcvz53EK72OW04pZLWp0","object":"chat.completion.chunk","created":1776258048,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"imi"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258048-mcvz53EK72OW04pZLWp0","object":"chat.completion.chunk","created":1776258048,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-k"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258048-mcvz53EK72OW04pZLWp0","object":"chat.completion.chunk","created":1776258048,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"2"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258048-mcvz53EK72OW04pZLWp0","object":"chat.completion.chunk","created":1776258048,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/read"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258048-mcvz53EK72OW04pZLWp0","object":"chat.completion.chunk","created":1776258048,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"_a"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258048-mcvz53EK72OW04pZLWp0","object":"chat.completion.chunk","created":1776258048,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"_file"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258048-mcvz53EK72OW04pZLWp0","object":"chat.completion.chunk","created":1776258048,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\"}"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258048-mcvz53EK72OW04pZLWp0","object":"chat.completion.chunk","created":1776258048,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}]}
-
- data: {"id":"gen-1776258048-mcvz53EK72OW04pZLWp0","object":"chat.completion.chunk","created":1776258048,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}],"usage":{"prompt_tokens":6487,"completion_tokens":31,"total_tokens":6518,"cost":0.0039697,"is_byok":false,"prompt_tokens_details":{"cached_tokens":6144,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0039697,"upstream_inference_prompt_cost":0.0038922,"upstream_inference_completions_cost":0.0000775},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 3.174354708s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31067
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/read_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"Read the go mod\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"ls:0\",\"function\":{\"arguments\":\"{\\\"path\\\": \\\"/tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/read_a_file\\\"}\",\"name\":\"ls\"},\"type\":\"function\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"role\":\"assistant\"},{\"content\":\"\\n- /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/read_a_file/\\n - go.mod\\n - main.go\\n\",\"tool_call_id\":\"ls:0\",\"role\":\"tool\"}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258052-AMbH99qYTdGUvupipcGe","object":"chat.completion.chunk","created":1776258052,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"id":"view:1","type":"function","function":{"name":"view","arguments":""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258052-AMbH99qYTdGUvupipcGe","object":"chat.completion.chunk","created":1776258052,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258052-AMbH99qYTdGUvupipcGe","object":"chat.completion.chunk","created":1776258052,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"file"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258052-AMbH99qYTdGUvupipcGe","object":"chat.completion.chunk","created":1776258052,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"_path"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258052-AMbH99qYTdGUvupipcGe","object":"chat.completion.chunk","created":1776258052,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258052-AMbH99qYTdGUvupipcGe","object":"chat.completion.chunk","created":1776258052,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \"/"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258052-AMbH99qYTdGUvupipcGe","object":"chat.completion.chunk","created":1776258052,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"tmp"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258052-AMbH99qYTdGUvupipcGe","object":"chat.completion.chunk","created":1776258052,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/c"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258052-AMbH99qYTdGUvupipcGe","object":"chat.completion.chunk","created":1776258052,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"rush"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258052-AMbH99qYTdGUvupipcGe","object":"chat.completion.chunk","created":1776258052,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-test"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258052-AMbH99qYTdGUvupipcGe","object":"chat.completion.chunk","created":1776258052,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/Test"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258052-AMbH99qYTdGUvupipcGe","object":"chat.completion.chunk","created":1776258052,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"Coder"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258052-AMbH99qYTdGUvupipcGe","object":"chat.completion.chunk","created":1776258052,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"Agent"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258052-AMbH99qYTdGUvupipcGe","object":"chat.completion.chunk","created":1776258052,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/open"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258052-AMbH99qYTdGUvupipcGe","object":"chat.completion.chunk","created":1776258052,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"router"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258052-AMbH99qYTdGUvupipcGe","object":"chat.completion.chunk","created":1776258052,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-k"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258052-AMbH99qYTdGUvupipcGe","object":"chat.completion.chunk","created":1776258052,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"imi"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258052-AMbH99qYTdGUvupipcGe","object":"chat.completion.chunk","created":1776258052,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-k"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258052-AMbH99qYTdGUvupipcGe","object":"chat.completion.chunk","created":1776258052,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"2"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258052-AMbH99qYTdGUvupipcGe","object":"chat.completion.chunk","created":1776258052,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/read"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258052-AMbH99qYTdGUvupipcGe","object":"chat.completion.chunk","created":1776258052,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"_a"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258052-AMbH99qYTdGUvupipcGe","object":"chat.completion.chunk","created":1776258052,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"_file"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258052-AMbH99qYTdGUvupipcGe","object":"chat.completion.chunk","created":1776258052,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/go"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258052-AMbH99qYTdGUvupipcGe","object":"chat.completion.chunk","created":1776258052,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":".mod"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258052-AMbH99qYTdGUvupipcGe","object":"chat.completion.chunk","created":1776258052,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\"}"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258052-AMbH99qYTdGUvupipcGe","object":"chat.completion.chunk","created":1776258052,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}]}
-
- data: {"id":"gen-1776258052-AMbH99qYTdGUvupipcGe","object":"chat.completion.chunk","created":1776258052,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}],"usage":{"prompt_tokens":6564,"completion_tokens":33,"total_tokens":6597,"cost":0.0040209,"is_byok":false,"prompt_tokens_details":{"cached_tokens":6400,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0040209,"upstream_inference_prompt_cost":0.0039384,"upstream_inference_completions_cost":0.0000825},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 3.082959917s
-- id: 3
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31402
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/read_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"Read the go mod\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"ls:0\",\"function\":{\"arguments\":\"{\\\"path\\\": \\\"/tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/read_a_file\\\"}\",\"name\":\"ls\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n- /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/read_a_file/\\n - go.mod\\n - main.go\\n\",\"tool_call_id\":\"ls:0\",\"role\":\"tool\"},{\"tool_calls\":[{\"id\":\"view:1\",\"function\":{\"arguments\":\"{\\\"file_path\\\": \\\"/tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/read_a_file/go.mod\\\"}\",\"name\":\"view\"},\"type\":\"function\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"role\":\"assistant\"},{\"content\":\"\\n 1|module example.com/testproject\\n 2|\\n 3|go 1.23\\n\\n\",\"tool_call_id\":\"view:1\",\"role\":\"tool\"}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258055-9N8aQCdkF4dT6b1KLPjk","object":"chat.completion.chunk","created":1776258055,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"```\n","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258055-9N8aQCdkF4dT6b1KLPjk","object":"chat.completion.chunk","created":1776258055,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"module","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258055-9N8aQCdkF4dT6b1KLPjk","object":"chat.completion.chunk","created":1776258055,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" example","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258055-9N8aQCdkF4dT6b1KLPjk","object":"chat.completion.chunk","created":1776258055,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":".com","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258055-9N8aQCdkF4dT6b1KLPjk","object":"chat.completion.chunk","created":1776258055,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"/test","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258055-9N8aQCdkF4dT6b1KLPjk","object":"chat.completion.chunk","created":1776258055,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"project","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258055-9N8aQCdkF4dT6b1KLPjk","object":"chat.completion.chunk","created":1776258055,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"\n\n","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258055-9N8aQCdkF4dT6b1KLPjk","object":"chat.completion.chunk","created":1776258055,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"go","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258055-9N8aQCdkF4dT6b1KLPjk","object":"chat.completion.chunk","created":1776258055,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" ","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258055-9N8aQCdkF4dT6b1KLPjk","object":"chat.completion.chunk","created":1776258055,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"1","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258055-9N8aQCdkF4dT6b1KLPjk","object":"chat.completion.chunk","created":1776258055,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":".","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258055-9N8aQCdkF4dT6b1KLPjk","object":"chat.completion.chunk","created":1776258055,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"23","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258055-9N8aQCdkF4dT6b1KLPjk","object":"chat.completion.chunk","created":1776258055,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"\n","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258055-9N8aQCdkF4dT6b1KLPjk","object":"chat.completion.chunk","created":1776258055,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"```","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258055-9N8aQCdkF4dT6b1KLPjk","object":"chat.completion.chunk","created":1776258055,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}]}
-
- data: {"id":"gen-1776258055-9N8aQCdkF4dT6b1KLPjk","object":"chat.completion.chunk","created":1776258055,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}],"usage":{"prompt_tokens":6642,"completion_tokens":15,"total_tokens":6657,"cost":0.0040227,"is_byok":false,"prompt_tokens_details":{"cached_tokens":6400,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0040227,"upstream_inference_prompt_cost":0.0039852,"upstream_inference_completions_cost":0.0000375},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 2.032411875s
diff --git a/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/sourcegraph_tool.yaml b/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/sourcegraph_tool.yaml
deleted file mode 100644
index e56d59f4664fa6628f0f3a9d0afdab6eeaff8cf5..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/sourcegraph_tool.yaml
+++ /dev/null
@@ -1,451 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 853
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse sourcegraph to search for ''func main'' in Go repositories\n \n\n","role":"user"}],"model":"qwen/qwen3-next-80b-a3b-instruct","max_tokens":40,"stream_options":{"include_usage":true},"usage":{"include":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258132-DCbWlrI5SsLNlrlQyV8Z","object":"chat.completion.chunk","created":1776258132,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":"Search","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258132-DCbWlrI5SsLNlrlQyV8Z","object":"chat.completion.chunk","created":1776258132,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":" for func","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258132-DCbWlrI5SsLNlrlQyV8Z","object":"chat.completion.chunk","created":1776258132,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":" main in Go repositories","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258132-DCbWlrI5SsLNlrlQyV8Z","object":"chat.completion.chunk","created":1776258132,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":" using Sourcegraph","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258132-DCbWlrI5SsLNlrlQyV8Z","object":"chat.completion.chunk","created":1776258132,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}]}
-
- data: {"id":"gen-1776258132-DCbWlrI5SsLNlrlQyV8Z","object":"chat.completion.chunk","created":1776258132,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Alibaba","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}],"usage":{"prompt_tokens":160,"completion_tokens":10,"total_tokens":170,"cost":0.0000234,"is_byok":false,"prompt_tokens_details":{"cached_tokens":0,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.000036,"upstream_inference_prompt_cost":0.000024,"upstream_inference_completions_cost":0.000012},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 572.347583ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30825
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/sourcegraph_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"use sourcegraph to search for 'func main' in Go repositories\",\"role\":\"user\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258132-1skNjypiOWdLtuXZpbuE","object":"chat.completion.chunk","created":1776258132,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"id":"sourcegraph:0","type":"function","function":{"name":"sourcegraph","arguments":""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258132-1skNjypiOWdLtuXZpbuE","object":"chat.completion.chunk","created":1776258132,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258132-1skNjypiOWdLtuXZpbuE","object":"chat.completion.chunk","created":1776258132,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"query"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258132-1skNjypiOWdLtuXZpbuE","object":"chat.completion.chunk","created":1776258132,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258132-1skNjypiOWdLtuXZpbuE","object":"chat.completion.chunk","created":1776258132,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258132-1skNjypiOWdLtuXZpbuE","object":"chat.completion.chunk","created":1776258132,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"func"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258132-1skNjypiOWdLtuXZpbuE","object":"chat.completion.chunk","created":1776258132,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" main"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258132-1skNjypiOWdLtuXZpbuE","object":"chat.completion.chunk","created":1776258132,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\","}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258132-1skNjypiOWdLtuXZpbuE","object":"chat.completion.chunk","created":1776258132,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258132-1skNjypiOWdLtuXZpbuE","object":"chat.completion.chunk","created":1776258132,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"count"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258132-1skNjypiOWdLtuXZpbuE","object":"chat.completion.chunk","created":1776258132,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258132-1skNjypiOWdLtuXZpbuE","object":"chat.completion.chunk","created":1776258132,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" "}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258132-1skNjypiOWdLtuXZpbuE","object":"chat.completion.chunk","created":1776258132,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"10"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258132-1skNjypiOWdLtuXZpbuE","object":"chat.completion.chunk","created":1776258132,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"}"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258132-1skNjypiOWdLtuXZpbuE","object":"chat.completion.chunk","created":1776258132,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}]}
-
- data: {"id":"gen-1776258132-1skNjypiOWdLtuXZpbuE","object":"chat.completion.chunk","created":1776258132,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}],"usage":{"prompt_tokens":6496,"completion_tokens":23,"total_tokens":6519,"cost":0.0039551,"is_byok":false,"prompt_tokens_details":{"cached_tokens":6144,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0039551,"upstream_inference_prompt_cost":0.0038976,"upstream_inference_completions_cost":0.0000575},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 2.345788458s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 416
- host: sourcegraph.com
- body: '{"query":"query Search($query: String!) { search(query: $query, version: V2, patternType: keyword ) { results { matchCount, limitHit, resultCount, approximateResultCount, missing { name }, timedout { name }, indexUnavailable, results { __typename, ... on FileMatch { repository { name }, file { path, url, content }, lineMatches { preview, lineNumber, offsetAndLengths } } } } } }","variables":{"query":"func main"}}'
- headers:
- Content-Type:
- - application/json
- User-Agent:
- - crush/1.0
- url: https://sourcegraph.com/.api/graphql
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: '{"data":{"search":{"results":{"matchCount":30,"limitHit":true,"resultCount":30,"approximateResultCount":"30+","missing":[],"timedout":[],"indexUnavailable":false,"results":[{"__typename":"Repository"},{"__typename":"Repository"},{"__typename":"Repository"},{"__typename":"FileMatch","repository":{"name":"github.com/swiftlang/swift"},"file":{"path":"test/embedded/linkage-mergeable4.swift","url":"/r/github.com/swiftlang/swift/-/blob/test/embedded/linkage-mergeable4.swift","content":"// RUN: %empty-directory(%t)\n// RUN: %{python} %utils/split_file.py -o %t %s\n\n// RUN: %target-swift-frontend -num-threads 2 -O -c -emit-module -o %t/MyModule.o %t/MyModule.swift -enable-experimental-feature Embedded -parse-as-library\n// RUN: %target-swift-frontend -num-threads 2 -O -c -o %t/MainA.o -o %t/MainB.o %t/MainA.swift %t/MainB.swift -I %t -enable-experimental-feature Embedded -parse-as-library\n// RUN: %target-embedded-link %target-clang-resource-dir-opt %t/MainA.o %t/MainB.o %t/MyModule.o %target-embedded-posix-shim -o %t/a.out\n// RUN: %target-run %t/a.out | %FileCheck %s\n\n// REQUIRES: swift_in_compiler\n// REQUIRES: executable_test\n// REQUIRES: swift_feature_Embedded\n\n// BEGIN MyModule.swift\n\nvar dict: [Int:Int] = [1: 2, 3: 4]\n\npublic func module_func() {\n print(\"module_func: \\(dict[1]!)\")\n}\n\n// BEGIN MainA.swift\n\nimport MyModule\n\nfileprivate var dict: [Int:Int] = [5: 6, 7: 8]\n\nfunc mainA_func() {\n print(\"main_func: \\(dict[5]!)\")\n}\n\n// BEGIN MainB.swift\n\nimport MyModule\n\nfileprivate var dict: [Int:Int] = [5: 6, 7: 8]\n\nfunc mainB_func() {\n print(\"main_func: \\(dict[5]!)\")\n}\n\n@main\nstruct Main {\n static func main() {\n module_func()\n mainA_func()\n mainB_func()\n }\n}\n\n// CHECK: module_func: 2\n// CHECK: main_func: 6\n"},"lineMatches":[{"preview":"func mainA_func() {","lineNumber":26,"offsetAndLengths":[[0,9],[11,4]]},{"preview":" print(\"main_func: \\(dict[5]!)\")","lineNumber":27,"offsetAndLengths":[[9,4],[14,4]]},{"preview":"func mainB_func() {","lineNumber":36,"offsetAndLengths":[[0,9],[11,4]]},{"preview":" print(\"main_func: \\(dict[5]!)\")","lineNumber":37,"offsetAndLengths":[[9,4],[14,4]]},{"preview":"@main","lineNumber":40,"offsetAndLengths":[[1,4]]},{"preview":"struct Main {","lineNumber":41,"offsetAndLengths":[[7,4]]},{"preview":" static func main() {","lineNumber":42,"offsetAndLengths":[[9,9]]},{"preview":" module_func()","lineNumber":43,"offsetAndLengths":[[11,4]]},{"preview":" mainA_func()","lineNumber":44,"offsetAndLengths":[[4,4],[10,4]]},{"preview":" mainB_func()","lineNumber":45,"offsetAndLengths":[[4,4],[10,4]]},{"preview":"// CHECK: module_func: 2","lineNumber":49,"offsetAndLengths":[[17,4]]},{"preview":"// CHECK: main_func: 6","lineNumber":50,"offsetAndLengths":[[10,4],[15,4]]},{"preview":"public func module_func() {","lineNumber":16,"offsetAndLengths":[[7,4],[19,4]]},{"preview":" print(\"module_func: \\(dict[1]!)\")","lineNumber":17,"offsetAndLengths":[[16,4]]},{"preview":"// BEGIN MainA.swift","lineNumber":20,"offsetAndLengths":[[9,4]]},{"preview":"// BEGIN MainB.swift","lineNumber":30,"offsetAndLengths":[[9,4]]},{"preview":"// RUN: %target-swift-frontend -num-threads 2 -O -c -o %t/MainA.o -o %t/MainB.o %t/MainA.swift %t/MainB.swift -I %t -enable-experimental-feature Embedded -parse-as-library","lineNumber":4,"offsetAndLengths":[[71,4],[85,4],[97,4]]}]}]}}}}'
- headers:
- Content-Type:
- - application/json
- status: 200 OK
- code: 200
- duration: 1.305185583s
-- id: 3
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 38801
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/sourcegraph_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use sourcegraph to search for 'func main' in Go repositories\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"sourcegraph:0\",\"function\":{\"arguments\":\"{\\\"query\\\": \\\"func main\\\", \\\"count\\\": 10}\",\"name\":\"sourcegraph\"},\"type\":\"function\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"role\":\"assistant\"},{\"content\":\"# Sourcegraph Search Results\\n\\nFound 30 matches across 30 results\\n(Result limit reached, try a more specific query)\\n\\n## Result 4: github.com/swiftlang/swift/test/embedded/linkage-mergeable4.swift\\n\\nURL: /r/github.com/swiftlang/swift/-/blob/test/embedded/linkage-mergeable4.swift\\n\\n```\\n16| \\n17| public func module_func() {\\n18| print(\\\"module_func: \\\\(dict[1]!)\\\")\\n19| }\\n20| \\n21| // BEGIN MainA.swift\\n22| \\n23| import MyModule\\n24| \\n25| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n26| func mainA_func() {\\n27| func mainA_func() {\\n28| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n29| }\\n30| \\n31| // BEGIN MainB.swift\\n32| \\n33| import MyModule\\n34| \\n35| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n36| \\n```\\n\\n```\\n17| public func module_func() {\\n18| print(\\\"module_func: \\\\(dict[1]!)\\\")\\n19| }\\n20| \\n21| // BEGIN MainA.swift\\n22| \\n23| import MyModule\\n24| \\n25| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n26| \\n27| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n28| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n29| }\\n30| \\n31| // BEGIN MainB.swift\\n32| \\n33| import MyModule\\n34| \\n35| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n36| \\n37| func mainB_func() {\\n```\\n\\n```\\n26| \\n27| func mainA_func() {\\n28| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n29| }\\n30| \\n31| // BEGIN MainB.swift\\n32| \\n33| import MyModule\\n34| \\n35| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n36| func mainB_func() {\\n37| func mainB_func() {\\n38| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n39| }\\n40| \\n41| @main\\n42| struct Main {\\n43| static func main() {\\n44| module_func()\\n45| mainA_func()\\n46| mainB_func()\\n```\\n\\n```\\n27| func mainA_func() {\\n28| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n29| }\\n30| \\n31| // BEGIN MainB.swift\\n32| \\n33| import MyModule\\n34| \\n35| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n36| \\n37| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n38| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n39| }\\n40| \\n41| @main\\n42| struct Main {\\n43| static func main() {\\n44| module_func()\\n45| mainA_func()\\n46| mainB_func()\\n47| }\\n```\\n\\n```\\n30| \\n31| // BEGIN MainB.swift\\n32| \\n33| import MyModule\\n34| \\n35| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n36| \\n37| func mainB_func() {\\n38| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n39| }\\n40| @main\\n41| @main\\n42| struct Main {\\n43| static func main() {\\n44| module_func()\\n45| mainA_func()\\n46| mainB_func()\\n47| }\\n48| }\\n49| \\n50| // CHECK: module_func: 2\\n```\\n\\n```\\n31| // BEGIN MainB.swift\\n32| \\n33| import MyModule\\n34| \\n35| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n36| \\n37| func mainB_func() {\\n38| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n39| }\\n40| \\n41| struct Main {\\n42| struct Main {\\n43| static func main() {\\n44| module_func()\\n45| mainA_func()\\n46| mainB_func()\\n47| }\\n48| }\\n49| \\n50| // CHECK: module_func: 2\\n51| // CHECK: main_func: 6\\n```\\n\\n```\\n32| \\n33| import MyModule\\n34| \\n35| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n36| \\n37| func mainB_func() {\\n38| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n39| }\\n40| \\n41| @main\\n42| static func main() {\\n43| static func main() {\\n44| module_func()\\n45| mainA_func()\\n46| mainB_func()\\n47| }\\n48| }\\n49| \\n50| // CHECK: module_func: 2\\n51| // CHECK: main_func: 6\\n52| \\n```\\n\\n```\\n33| import MyModule\\n34| \\n35| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n36| \\n37| func mainB_func() {\\n38| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n39| }\\n40| \\n41| @main\\n42| struct Main {\\n43| module_func()\\n44| module_func()\\n45| mainA_func()\\n46| mainB_func()\\n47| }\\n48| }\\n49| \\n50| // CHECK: module_func: 2\\n51| // CHECK: main_func: 6\\n52| \\n```\\n\\n```\\n34| \\n35| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n36| \\n37| func mainB_func() {\\n38| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n39| }\\n40| \\n41| @main\\n42| struct Main {\\n43| static func main() {\\n44| mainA_func()\\n45| mainA_func()\\n46| mainB_func()\\n47| }\\n48| }\\n49| \\n50| // CHECK: module_func: 2\\n51| // CHECK: main_func: 6\\n52| \\n```\\n\\n```\\n35| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n36| \\n37| func mainB_func() {\\n38| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n39| }\\n40| \\n41| @main\\n42| struct Main {\\n43| static func main() {\\n44| module_func()\\n45| mainB_func()\\n46| mainB_func()\\n47| }\\n48| }\\n49| \\n50| // CHECK: module_func: 2\\n51| // CHECK: main_func: 6\\n52| \\n```\\n\\n```\\n39| }\\n40| \\n41| @main\\n42| struct Main {\\n43| static func main() {\\n44| module_func()\\n45| mainA_func()\\n46| mainB_func()\\n47| }\\n48| }\\n49| // CHECK: module_func: 2\\n50| // CHECK: module_func: 2\\n51| // CHECK: main_func: 6\\n52| \\n```\\n\\n```\\n40| \\n41| @main\\n42| struct Main {\\n43| static func main() {\\n44| module_func()\\n45| mainA_func()\\n46| mainB_func()\\n47| }\\n48| }\\n49| \\n50| // CHECK: main_func: 6\\n51| // CHECK: main_func: 6\\n52| \\n```\\n\\n```\\n6| // RUN: %target-embedded-link %target-clang-resource-dir-opt %t/MainA.o %t/MainB.o %t/MyModule.o %target-embedded-posix-shim -o %t/a.out\\n7| // RUN: %target-run %t/a.out | %FileCheck %s\\n8| \\n9| // REQUIRES: swift_in_compiler\\n10| // REQUIRES: executable_test\\n11| // REQUIRES: swift_feature_Embedded\\n12| \\n13| // BEGIN MyModule.swift\\n14| \\n15| var dict: [Int:Int] = [1: 2, 3: 4]\\n16| public func module_func() {\\n17| public func module_func() {\\n18| print(\\\"module_func: \\\\(dict[1]!)\\\")\\n19| }\\n20| \\n21| // BEGIN MainA.swift\\n22| \\n23| import MyModule\\n24| \\n25| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n26| \\n```\\n\\n```\\n7| // RUN: %target-run %t/a.out | %FileCheck %s\\n8| \\n9| // REQUIRES: swift_in_compiler\\n10| // REQUIRES: executable_test\\n11| // REQUIRES: swift_feature_Embedded\\n12| \\n13| // BEGIN MyModule.swift\\n14| \\n15| var dict: [Int:Int] = [1: 2, 3: 4]\\n16| \\n17| print(\\\"module_func: \\\\(dict[1]!)\\\")\\n18| print(\\\"module_func: \\\\(dict[1]!)\\\")\\n19| }\\n20| \\n21| // BEGIN MainA.swift\\n22| \\n23| import MyModule\\n24| \\n25| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n26| \\n27| func mainA_func() {\\n```\\n\\n```\\n10| // REQUIRES: executable_test\\n11| // REQUIRES: swift_feature_Embedded\\n12| \\n13| // BEGIN MyModule.swift\\n14| \\n15| var dict: [Int:Int] = [1: 2, 3: 4]\\n16| \\n17| public func module_func() {\\n18| print(\\\"module_func: \\\\(dict[1]!)\\\")\\n19| }\\n20| // BEGIN MainA.swift\\n21| // BEGIN MainA.swift\\n22| \\n23| import MyModule\\n24| \\n25| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n26| \\n27| func mainA_func() {\\n28| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n29| }\\n30| \\n```\\n\\n```\\n20| \\n21| // BEGIN MainA.swift\\n22| \\n23| import MyModule\\n24| \\n25| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n26| \\n27| func mainA_func() {\\n28| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n29| }\\n30| // BEGIN MainB.swift\\n31| // BEGIN MainB.swift\\n32| \\n33| import MyModule\\n34| \\n35| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n36| \\n37| func mainB_func() {\\n38| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n39| }\\n40| \\n```\\n\\n```\\n1| // RUN: %empty-directory(%t)\\n2| // RUN: %{python} %utils/split_file.py -o %t %s\\n3| \\n4| // RUN: %target-swift-frontend -num-threads 2 -O -c -o %t/MainA.o -o %t/MainB.o %t/MainA.swift %t/MainB.swift -I %t -enable-experimental-feature Embedded -parse-as-library\\n5| // RUN: %target-swift-frontend -num-threads 2 -O -c -o %t/MainA.o -o %t/MainB.o %t/MainA.swift %t/MainB.swift -I %t -enable-experimental-feature Embedded -parse-as-library\\n6| // RUN: %target-embedded-link %target-clang-resource-dir-opt %t/MainA.o %t/MainB.o %t/MyModule.o %target-embedded-posix-shim -o %t/a.out\\n7| // RUN: %target-run %t/a.out | %FileCheck %s\\n8| \\n9| // REQUIRES: swift_in_compiler\\n10| // REQUIRES: executable_test\\n11| // REQUIRES: swift_feature_Embedded\\n12| \\n13| // BEGIN MyModule.swift\\n14| \\n```\\n\\n\",\"tool_call_id\":\"sourcegraph:0\",\"role\":\"tool\"}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"The","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" search","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" returned","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" mixed","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" results","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" including","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" Swift","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" files","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":".","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" Let","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" me","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" narrow","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" the","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" search","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" specifically","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" to","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" Go","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" repositories","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":":","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"id":"sourcegraph:1","type":"function","function":{"name":"sourcegraph","arguments":""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"query"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"func"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" main"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" lang"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"go"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\","}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"count"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" "}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"10"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"}"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}]}
-
- data: {"id":"gen-1776258136-NagAcYUREftMbvvszkl2","object":"chat.completion.chunk","created":1776258136,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}],"usage":{"prompt_tokens":9324,"completion_tokens":45,"total_tokens":9369,"cost":0.0057069,"is_byok":false,"prompt_tokens_details":{"cached_tokens":6400,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0057069,"upstream_inference_prompt_cost":0.0055944,"upstream_inference_completions_cost":0.0001125},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 1.924628625s
-- id: 4
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 424
- host: sourcegraph.com
- body: '{"query":"query Search($query: String!) { search(query: $query, version: V2, patternType: keyword ) { results { matchCount, limitHit, resultCount, approximateResultCount, missing { name }, timedout { name }, indexUnavailable, results { __typename, ... on FileMatch { repository { name }, file { path, url, content }, lineMatches { preview, lineNumber, offsetAndLengths } } } } } }","variables":{"query":"func main lang:go"}}'
- headers:
- Content-Type:
- - application/json
- User-Agent:
- - crush/1.0
- url: https://sourcegraph.com/.api/graphql
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: '{"data":{"search":{"results":{"matchCount":30,"limitHit":true,"resultCount":30,"approximateResultCount":"30+","missing":[],"timedout":[],"indexUnavailable":false,"results":[{"__typename":"FileMatch","repository":{"name":"github.com/GoesToEleven/GolangTraining"},"file":{"path":"14_functions/10_func-expression/01_before-func-expression/main.go","url":"/r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/01_before-func-expression/main.go","content":"package main\n\nimport \"fmt\"\n\nfunc greeting() {\n\tfmt.Println(\"Hello world!\")\n}\n\nfunc main() {\n\tgreeting()\n}\n"},"lineMatches":[{"preview":"func main() {","lineNumber":8,"offsetAndLengths":[[0,9]]},{"preview":"package main","lineNumber":0,"offsetAndLengths":[[8,4]]},{"preview":"func greeting() {","lineNumber":4,"offsetAndLengths":[[0,4]]}]},{"__typename":"FileMatch","repository":{"name":"github.com/GoesToEleven/GolangTraining"},"file":{"path":"14_functions/10_func-expression/02_func-expression/main.go","url":"/r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/02_func-expression/main.go","content":"package main\n\nimport \"fmt\"\n\nfunc main() {\n\n\tgreeting := func() {\n\t\tfmt.Println(\"Hello world!\")\n\t}\n\n\tgreeting()\n}\n"},"lineMatches":[{"preview":"func main() {","lineNumber":4,"offsetAndLengths":[[0,9]]},{"preview":"\tgreeting := func() {","lineNumber":6,"offsetAndLengths":[[13,4]]},{"preview":"package main","lineNumber":0,"offsetAndLengths":[[8,4]]}]},{"__typename":"FileMatch","repository":{"name":"github.com/GoesToEleven/GolangTraining"},"file":{"path":"14_functions/10_func-expression/03_func-expression_shows-type/main.go","url":"/r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/03_func-expression_shows-type/main.go","content":"package main\n\nimport \"fmt\"\n\nfunc main() {\n\n\tgreeting := func() {\n\t\tfmt.Println(\"Hello world!\")\n\t}\n\n\tgreeting()\n\tfmt.Printf(\"%T\\n\", greeting)\n}\n"},"lineMatches":[{"preview":"func main() {","lineNumber":4,"offsetAndLengths":[[0,9]]},{"preview":"\tgreeting := func() {","lineNumber":6,"offsetAndLengths":[[13,4]]},{"preview":"package main","lineNumber":0,"offsetAndLengths":[[8,4]]}]},{"__typename":"FileMatch","repository":{"name":"github.com/GoesToEleven/GolangTraining"},"file":{"path":"14_functions/10_func-expression/04_another-way_func-expression/main.go","url":"/r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/04_another-way_func-expression/main.go","content":"package main\n\nimport \"fmt\"\n\nfunc makeGreeter() func() string {\n\treturn func() string {\n\t\treturn \"Hello world!\"\n\t}\n}\n\nfunc main() {\n\tgreet := makeGreeter()\n\tfmt.Println(greet())\n}\n"},"lineMatches":[{"preview":"func main() {","lineNumber":10,"offsetAndLengths":[[0,9]]},{"preview":"package main","lineNumber":0,"offsetAndLengths":[[8,4]]},{"preview":"func makeGreeter() func() string {","lineNumber":4,"offsetAndLengths":[[0,4],[19,4]]},{"preview":"\treturn func() string {","lineNumber":5,"offsetAndLengths":[[8,4]]}]},{"__typename":"FileMatch","repository":{"name":"github.com/jmbharathram/executeoncommand"},"file":{"path":"go/channel/main.go","url":"/r/github.com/jmbharathram/executeoncommand/-/blob/go/channel/main.go","content":"package main\n\nimport \"fmt\"\n\n// Example 0\nfunc dummy_func() {\n\tfmt.Println(\"dummy_func\")\n}\nfunc main() {\n\tgo dummy_func()\n fmt.Println(\"end of main\")\n}\n\n\n// Example 1\n\npackage main\n\nimport \"fmt\"\n\nfunc dummy_func(ch chan int) {\n // receive a value from channel\n val := \u003c-ch\n\tfmt.Println(val)\n}\nfunc main() {\n // create a channel\n\tch := make(chan int)\n // trigger a parallel go routine\n\tgo dummy_func(ch)\n // send a value into the channel\n\tch \u003c- 10\n}\n\n// Example 2\n\npackage main\n\nimport (\n \"fmt\"\n //\"flag\"\n)\n\nfunc dummy_func(ch chan int) {\n // receive a value from channel\n val := \u003c-ch\n\tfmt.Println(val)\n\n //send a value again\n ch \u003c- val*2\n}\nfunc main() {\n fmt.Println(\"Welcome to main function\")\n \n //flag statements for user input\n var input_value int\n input_value = 10\n //flag.IntVar(\u0026input_value, \"v\", 10, \"Input value to send/receive via channel.\" )\n //flag.Parse()\n\n // create a channel\n\tch := make(chan int)\n\n // trigger a parallel go routine\n\tgo dummy_func(ch)\n \n // send a value into the channel\n\tch \u003c- input_value\n\n // receive a value from channel\n response := \u003c- ch\n fmt.Println(response)\n\n}\n\n// example 3\n\n\npackage main\n\nimport(\n\"fmt\"\n\"time\"\n\"math/rand\"\n)\n\nfunc routine1(ch chan int, min int, max int) {\n\n rand.Seed(time.Now().UnixNano())\n t := rand.Intn(max - min) + min\n fmt.Println(\"routine 1 is waiting: \", t, \"secs\")\n time.Sleep(time.Duration(t)*time.Second)\n\n ch \u003c- t\n}\n\nfunc routine2(ch chan int, min int, max int) {\n\n rand.Seed(time.Now().UnixNano())\n t := rand.Intn(max - min) + min\n fmt.Println(\"routine 2 is waiting: \", t, \"secs\")\n time.Sleep(time.Duration(t)*time.Second)\n\n ch \u003c- t\n}\n\nfunc main(){\n\nmin := 1\nmax := 10\n\nch1:= make(chan int)\nch2:= make(chan int)\n\ngo routine1(ch1, min, max)\ngo routine2(ch2, min, max)\n\nselect{\n\ncase \u003c- ch1:\n fmt.Println(\"routine 1 finished\")\n\ncase \u003c- ch2:\n fmt.Println(\"routine 2 finished\")\n}\n\nfmt.Println(\"end of main function\")\n\n}\n"},"lineMatches":[{"preview":"func main(){","lineNumber":106,"offsetAndLengths":[[0,9]]},{"preview":"func dummy_func() {","lineNumber":5,"offsetAndLengths":[[0,4],[11,4]]},{"preview":"\tfmt.Println(\"dummy_func\")","lineNumber":6,"offsetAndLengths":[[20,4]]},{"preview":"func main() {","lineNumber":8,"offsetAndLengths":[[0,9]]},{"preview":"\tgo dummy_func()","lineNumber":9,"offsetAndLengths":[[10,4]]},{"preview":" fmt.Println(\"end of main\")","lineNumber":10,"offsetAndLengths":[[24,4]]},{"preview":"func main() {","lineNumber":25,"offsetAndLengths":[[0,9]]},{"preview":"func main() {","lineNumber":51,"offsetAndLengths":[[0,9]]},{"preview":" fmt.Println(\"Welcome to main function\")","lineNumber":52,"offsetAndLengths":[[28,4],[33,4]]},{"preview":"fmt.Println(\"end of main function\")","lineNumber":126,"offsetAndLengths":[[20,4],[25,4]]},{"preview":"package main","lineNumber":78,"offsetAndLengths":[[8,4]]},{"preview":"package main","lineNumber":16,"offsetAndLengths":[[8,4]]},{"preview":"package main","lineNumber":0,"offsetAndLengths":[[8,4]]}]}]}}}}'
- headers:
- Content-Type:
- - application/json
- status: 200 OK
- code: 200
- duration: 1.317874417s
-- id: 5
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 48209
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/sourcegraph_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use sourcegraph to search for 'func main' in Go repositories\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"sourcegraph:0\",\"function\":{\"arguments\":\"{\\\"query\\\": \\\"func main\\\", \\\"count\\\": 10}\",\"name\":\"sourcegraph\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"# Sourcegraph Search Results\\n\\nFound 30 matches across 30 results\\n(Result limit reached, try a more specific query)\\n\\n## Result 4: github.com/swiftlang/swift/test/embedded/linkage-mergeable4.swift\\n\\nURL: /r/github.com/swiftlang/swift/-/blob/test/embedded/linkage-mergeable4.swift\\n\\n```\\n16| \\n17| public func module_func() {\\n18| print(\\\"module_func: \\\\(dict[1]!)\\\")\\n19| }\\n20| \\n21| // BEGIN MainA.swift\\n22| \\n23| import MyModule\\n24| \\n25| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n26| func mainA_func() {\\n27| func mainA_func() {\\n28| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n29| }\\n30| \\n31| // BEGIN MainB.swift\\n32| \\n33| import MyModule\\n34| \\n35| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n36| \\n```\\n\\n```\\n17| public func module_func() {\\n18| print(\\\"module_func: \\\\(dict[1]!)\\\")\\n19| }\\n20| \\n21| // BEGIN MainA.swift\\n22| \\n23| import MyModule\\n24| \\n25| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n26| \\n27| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n28| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n29| }\\n30| \\n31| // BEGIN MainB.swift\\n32| \\n33| import MyModule\\n34| \\n35| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n36| \\n37| func mainB_func() {\\n```\\n\\n```\\n26| \\n27| func mainA_func() {\\n28| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n29| }\\n30| \\n31| // BEGIN MainB.swift\\n32| \\n33| import MyModule\\n34| \\n35| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n36| func mainB_func() {\\n37| func mainB_func() {\\n38| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n39| }\\n40| \\n41| @main\\n42| struct Main {\\n43| static func main() {\\n44| module_func()\\n45| mainA_func()\\n46| mainB_func()\\n```\\n\\n```\\n27| func mainA_func() {\\n28| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n29| }\\n30| \\n31| // BEGIN MainB.swift\\n32| \\n33| import MyModule\\n34| \\n35| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n36| \\n37| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n38| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n39| }\\n40| \\n41| @main\\n42| struct Main {\\n43| static func main() {\\n44| module_func()\\n45| mainA_func()\\n46| mainB_func()\\n47| }\\n```\\n\\n```\\n30| \\n31| // BEGIN MainB.swift\\n32| \\n33| import MyModule\\n34| \\n35| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n36| \\n37| func mainB_func() {\\n38| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n39| }\\n40| @main\\n41| @main\\n42| struct Main {\\n43| static func main() {\\n44| module_func()\\n45| mainA_func()\\n46| mainB_func()\\n47| }\\n48| }\\n49| \\n50| // CHECK: module_func: 2\\n```\\n\\n```\\n31| // BEGIN MainB.swift\\n32| \\n33| import MyModule\\n34| \\n35| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n36| \\n37| func mainB_func() {\\n38| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n39| }\\n40| \\n41| struct Main {\\n42| struct Main {\\n43| static func main() {\\n44| module_func()\\n45| mainA_func()\\n46| mainB_func()\\n47| }\\n48| }\\n49| \\n50| // CHECK: module_func: 2\\n51| // CHECK: main_func: 6\\n```\\n\\n```\\n32| \\n33| import MyModule\\n34| \\n35| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n36| \\n37| func mainB_func() {\\n38| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n39| }\\n40| \\n41| @main\\n42| static func main() {\\n43| static func main() {\\n44| module_func()\\n45| mainA_func()\\n46| mainB_func()\\n47| }\\n48| }\\n49| \\n50| // CHECK: module_func: 2\\n51| // CHECK: main_func: 6\\n52| \\n```\\n\\n```\\n33| import MyModule\\n34| \\n35| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n36| \\n37| func mainB_func() {\\n38| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n39| }\\n40| \\n41| @main\\n42| struct Main {\\n43| module_func()\\n44| module_func()\\n45| mainA_func()\\n46| mainB_func()\\n47| }\\n48| }\\n49| \\n50| // CHECK: module_func: 2\\n51| // CHECK: main_func: 6\\n52| \\n```\\n\\n```\\n34| \\n35| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n36| \\n37| func mainB_func() {\\n38| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n39| }\\n40| \\n41| @main\\n42| struct Main {\\n43| static func main() {\\n44| mainA_func()\\n45| mainA_func()\\n46| mainB_func()\\n47| }\\n48| }\\n49| \\n50| // CHECK: module_func: 2\\n51| // CHECK: main_func: 6\\n52| \\n```\\n\\n```\\n35| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n36| \\n37| func mainB_func() {\\n38| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n39| }\\n40| \\n41| @main\\n42| struct Main {\\n43| static func main() {\\n44| module_func()\\n45| mainB_func()\\n46| mainB_func()\\n47| }\\n48| }\\n49| \\n50| // CHECK: module_func: 2\\n51| // CHECK: main_func: 6\\n52| \\n```\\n\\n```\\n39| }\\n40| \\n41| @main\\n42| struct Main {\\n43| static func main() {\\n44| module_func()\\n45| mainA_func()\\n46| mainB_func()\\n47| }\\n48| }\\n49| // CHECK: module_func: 2\\n50| // CHECK: module_func: 2\\n51| // CHECK: main_func: 6\\n52| \\n```\\n\\n```\\n40| \\n41| @main\\n42| struct Main {\\n43| static func main() {\\n44| module_func()\\n45| mainA_func()\\n46| mainB_func()\\n47| }\\n48| }\\n49| \\n50| // CHECK: main_func: 6\\n51| // CHECK: main_func: 6\\n52| \\n```\\n\\n```\\n6| // RUN: %target-embedded-link %target-clang-resource-dir-opt %t/MainA.o %t/MainB.o %t/MyModule.o %target-embedded-posix-shim -o %t/a.out\\n7| // RUN: %target-run %t/a.out | %FileCheck %s\\n8| \\n9| // REQUIRES: swift_in_compiler\\n10| // REQUIRES: executable_test\\n11| // REQUIRES: swift_feature_Embedded\\n12| \\n13| // BEGIN MyModule.swift\\n14| \\n15| var dict: [Int:Int] = [1: 2, 3: 4]\\n16| public func module_func() {\\n17| public func module_func() {\\n18| print(\\\"module_func: \\\\(dict[1]!)\\\")\\n19| }\\n20| \\n21| // BEGIN MainA.swift\\n22| \\n23| import MyModule\\n24| \\n25| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n26| \\n```\\n\\n```\\n7| // RUN: %target-run %t/a.out | %FileCheck %s\\n8| \\n9| // REQUIRES: swift_in_compiler\\n10| // REQUIRES: executable_test\\n11| // REQUIRES: swift_feature_Embedded\\n12| \\n13| // BEGIN MyModule.swift\\n14| \\n15| var dict: [Int:Int] = [1: 2, 3: 4]\\n16| \\n17| print(\\\"module_func: \\\\(dict[1]!)\\\")\\n18| print(\\\"module_func: \\\\(dict[1]!)\\\")\\n19| }\\n20| \\n21| // BEGIN MainA.swift\\n22| \\n23| import MyModule\\n24| \\n25| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n26| \\n27| func mainA_func() {\\n```\\n\\n```\\n10| // REQUIRES: executable_test\\n11| // REQUIRES: swift_feature_Embedded\\n12| \\n13| // BEGIN MyModule.swift\\n14| \\n15| var dict: [Int:Int] = [1: 2, 3: 4]\\n16| \\n17| public func module_func() {\\n18| print(\\\"module_func: \\\\(dict[1]!)\\\")\\n19| }\\n20| // BEGIN MainA.swift\\n21| // BEGIN MainA.swift\\n22| \\n23| import MyModule\\n24| \\n25| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n26| \\n27| func mainA_func() {\\n28| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n29| }\\n30| \\n```\\n\\n```\\n20| \\n21| // BEGIN MainA.swift\\n22| \\n23| import MyModule\\n24| \\n25| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n26| \\n27| func mainA_func() {\\n28| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n29| }\\n30| // BEGIN MainB.swift\\n31| // BEGIN MainB.swift\\n32| \\n33| import MyModule\\n34| \\n35| fileprivate var dict: [Int:Int] = [5: 6, 7: 8]\\n36| \\n37| func mainB_func() {\\n38| print(\\\"main_func: \\\\(dict[5]!)\\\")\\n39| }\\n40| \\n```\\n\\n```\\n1| // RUN: %empty-directory(%t)\\n2| // RUN: %{python} %utils/split_file.py -o %t %s\\n3| \\n4| // RUN: %target-swift-frontend -num-threads 2 -O -c -o %t/MainA.o -o %t/MainB.o %t/MainA.swift %t/MainB.swift -I %t -enable-experimental-feature Embedded -parse-as-library\\n5| // RUN: %target-swift-frontend -num-threads 2 -O -c -o %t/MainA.o -o %t/MainB.o %t/MainA.swift %t/MainB.swift -I %t -enable-experimental-feature Embedded -parse-as-library\\n6| // RUN: %target-embedded-link %target-clang-resource-dir-opt %t/MainA.o %t/MainB.o %t/MyModule.o %target-embedded-posix-shim -o %t/a.out\\n7| // RUN: %target-run %t/a.out | %FileCheck %s\\n8| \\n9| // REQUIRES: swift_in_compiler\\n10| // REQUIRES: executable_test\\n11| // REQUIRES: swift_feature_Embedded\\n12| \\n13| // BEGIN MyModule.swift\\n14| \\n```\\n\\n\",\"tool_call_id\":\"sourcegraph:0\",\"role\":\"tool\"},{\"content\":\"The search returned mixed results including Swift files. Let me narrow the search specifically to Go repositories:\",\"tool_calls\":[{\"id\":\"sourcegraph:1\",\"function\":{\"arguments\":\"{\\\"query\\\": \\\"func main lang:go\\\", \\\"count\\\": 10}\",\"name\":\"sourcegraph\"},\"type\":\"function\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"role\":\"assistant\"},{\"content\":\"# Sourcegraph Search Results\\n\\nFound 30 matches across 30 results\\n(Result limit reached, try a more specific query)\\n\\n## Result 1: github.com/GoesToEleven/GolangTraining/14_functions/10_func-expression/01_before-func-expression/main.go\\n\\nURL: /r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/01_before-func-expression/main.go\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func greeting() {\\n6| \\tfmt.Println(\\\"Hello world!\\\")\\n7| }\\n8| func main() {\\n9| func main() {\\n10| \\tgreeting()\\n11| }\\n12| \\n```\\n\\n```\\n0| package main\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func greeting() {\\n6| \\tfmt.Println(\\\"Hello world!\\\")\\n7| }\\n8| \\n9| func main() {\\n10| \\tgreeting()\\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| func greeting() {\\n5| func greeting() {\\n6| \\tfmt.Println(\\\"Hello world!\\\")\\n7| }\\n8| \\n9| func main() {\\n10| \\tgreeting()\\n11| }\\n12| \\n```\\n\\n## Result 2: github.com/GoesToEleven/GolangTraining/14_functions/10_func-expression/02_func-expression/main.go\\n\\nURL: /r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/02_func-expression/main.go\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| func main() {\\n5| func main() {\\n6| \\n7| \\tgreeting := func() {\\n8| \\t\\tfmt.Println(\\\"Hello world!\\\")\\n9| \\t}\\n10| \\n11| \\tgreeting()\\n12| }\\n13| \\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func main() {\\n6| \\tgreeting := func() {\\n7| \\tgreeting := func() {\\n8| \\t\\tfmt.Println(\\\"Hello world!\\\")\\n9| \\t}\\n10| \\n11| \\tgreeting()\\n12| }\\n13| \\n```\\n\\n```\\n0| package main\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func main() {\\n6| \\n7| \\tgreeting := func() {\\n8| \\t\\tfmt.Println(\\\"Hello world!\\\")\\n9| \\t}\\n10| \\n```\\n\\n## Result 3: github.com/GoesToEleven/GolangTraining/14_functions/10_func-expression/03_func-expression_shows-type/main.go\\n\\nURL: /r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/03_func-expression_shows-type/main.go\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| func main() {\\n5| func main() {\\n6| \\n7| \\tgreeting := func() {\\n8| \\t\\tfmt.Println(\\\"Hello world!\\\")\\n9| \\t}\\n10| \\n11| \\tgreeting()\\n12| \\tfmt.Printf(\\\"%T\\\\n\\\", greeting)\\n13| }\\n14| \\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func main() {\\n6| \\tgreeting := func() {\\n7| \\tgreeting := func() {\\n8| \\t\\tfmt.Println(\\\"Hello world!\\\")\\n9| \\t}\\n10| \\n11| \\tgreeting()\\n12| \\tfmt.Printf(\\\"%T\\\\n\\\", greeting)\\n13| }\\n14| \\n```\\n\\n```\\n0| package main\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func main() {\\n6| \\n7| \\tgreeting := func() {\\n8| \\t\\tfmt.Println(\\\"Hello world!\\\")\\n9| \\t}\\n10| \\n```\\n\\n## Result 4: github.com/GoesToEleven/GolangTraining/14_functions/10_func-expression/04_another-way_func-expression/main.go\\n\\nURL: /r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/04_another-way_func-expression/main.go\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func makeGreeter() func() string {\\n6| \\treturn func() string {\\n7| \\t\\treturn \\\"Hello world!\\\"\\n8| \\t}\\n9| }\\n10| func main() {\\n11| func main() {\\n12| \\tgreet := makeGreeter()\\n13| \\tfmt.Println(greet())\\n14| }\\n15| \\n```\\n\\n```\\n0| package main\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func makeGreeter() func() string {\\n6| \\treturn func() string {\\n7| \\t\\treturn \\\"Hello world!\\\"\\n8| \\t}\\n9| }\\n10| \\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| func makeGreeter() func() string {\\n5| func makeGreeter() func() string {\\n6| \\treturn func() string {\\n7| \\t\\treturn \\\"Hello world!\\\"\\n8| \\t}\\n9| }\\n10| \\n11| func main() {\\n12| \\tgreet := makeGreeter()\\n13| \\tfmt.Println(greet())\\n14| }\\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| \\treturn func() string {\\n6| \\treturn func() string {\\n7| \\t\\treturn \\\"Hello world!\\\"\\n8| \\t}\\n9| }\\n10| \\n11| func main() {\\n12| \\tgreet := makeGreeter()\\n13| \\tfmt.Println(greet())\\n14| }\\n15| \\n```\\n\\n## Result 5: github.com/jmbharathram/executeoncommand/go/channel/main.go\\n\\nURL: /r/github.com/jmbharathram/executeoncommand/-/blob/go/channel/main.go\\n\\n```\\n96| \\n97| func routine2(ch chan int, min int, max int) {\\n98| \\n99| rand.Seed(time.Now().UnixNano())\\n100| t := rand.Intn(max - min) + min\\n101| fmt.Println(\\\"routine 2 is waiting: \\\", t, \\\"secs\\\")\\n102| time.Sleep(time.Duration(t)*time.Second)\\n103| \\n104| ch <- t\\n105| }\\n106| func main(){\\n107| func main(){\\n108| \\n109| min := 1\\n110| max := 10\\n111| \\n112| ch1:= make(chan int)\\n113| ch2:= make(chan int)\\n114| \\n115| go routine1(ch1, min, max)\\n116| go routine2(ch2, min, max)\\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func dummy_func() {\\n6| func dummy_func() {\\n7| \\tfmt.Println(\\\"dummy_func\\\")\\n8| }\\n9| func main() {\\n10| \\tgo dummy_func()\\n11| fmt.Println(\\\"end of main\\\")\\n12| }\\n13| \\n14| \\n15| // Example 1\\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| // Example 0\\n6| \\tfmt.Println(\\\"dummy_func\\\")\\n7| \\tfmt.Println(\\\"dummy_func\\\")\\n8| }\\n9| func main() {\\n10| \\tgo dummy_func()\\n11| fmt.Println(\\\"end of main\\\")\\n12| }\\n13| \\n14| \\n15| // Example 1\\n16| \\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| // Example 0\\n6| func dummy_func() {\\n7| \\tfmt.Println(\\\"dummy_func\\\")\\n8| func main() {\\n9| func main() {\\n10| \\tgo dummy_func()\\n11| fmt.Println(\\\"end of main\\\")\\n12| }\\n13| \\n14| \\n15| // Example 1\\n16| \\n17| package main\\n18| \\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| // Example 0\\n6| func dummy_func() {\\n7| \\tfmt.Println(\\\"dummy_func\\\")\\n8| }\\n9| \\tgo dummy_func()\\n10| \\tgo dummy_func()\\n11| fmt.Println(\\\"end of main\\\")\\n12| }\\n13| \\n14| \\n15| // Example 1\\n16| \\n17| package main\\n18| \\n19| import \\\"fmt\\\"\\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| // Example 0\\n6| func dummy_func() {\\n7| \\tfmt.Println(\\\"dummy_func\\\")\\n8| }\\n9| func main() {\\n10| fmt.Println(\\\"end of main\\\")\\n11| fmt.Println(\\\"end of main\\\")\\n12| }\\n13| \\n14| \\n15| // Example 1\\n16| \\n17| package main\\n18| \\n19| import \\\"fmt\\\"\\n20| \\n```\\n\\n```\\n15| // Example 1\\n16| \\n17| package main\\n18| \\n19| import \\\"fmt\\\"\\n20| \\n21| func dummy_func(ch chan int) {\\n22| // receive a value from channel\\n23| val := <-ch\\n24| \\tfmt.Println(val)\\n25| func main() {\\n26| func main() {\\n27| // create a channel\\n28| \\tch := make(chan int)\\n29| // trigger a parallel go routine\\n30| \\tgo dummy_func(ch)\\n31| // send a value into the channel\\n32| \\tch <- 10\\n33| }\\n34| \\n35| // Example 2\\n```\\n\\n```\\n41| //\\\"flag\\\"\\n42| )\\n43| \\n44| func dummy_func(ch chan int) {\\n45| // receive a value from channel\\n46| val := <-ch\\n47| \\tfmt.Println(val)\\n48| \\n49| //send a value again\\n50| ch <- val*2\\n51| func main() {\\n52| func main() {\\n53| fmt.Println(\\\"Welcome to main function\\\")\\n54| \\n55| //flag statements for user input\\n56| var input_value int\\n57| input_value = 10\\n58| //flag.IntVar(&input_value, \\\"v\\\", 10, \\\"Input value to send/receive via channel.\\\" )\\n59| //flag.Parse()\\n60| \\n61| // create a channel\\n```\\n\\n```\\n42| )\\n43| \\n44| func dummy_func(ch chan int) {\\n45| // receive a value from channel\\n46| val := <-ch\\n47| \\tfmt.Println(val)\\n48| \\n49| //send a value again\\n50| ch <- val*2\\n51| }\\n52| fmt.Println(\\\"Welcome to main function\\\")\\n53| fmt.Println(\\\"Welcome to main function\\\")\\n54| \\n55| //flag statements for user input\\n56| var input_value int\\n57| input_value = 10\\n58| //flag.IntVar(&input_value, \\\"v\\\", 10, \\\"Input value to send/receive via channel.\\\" )\\n59| //flag.Parse()\\n60| \\n61| // create a channel\\n62| \\tch := make(chan int)\\n```\\n\\n```\\n116| go routine2(ch2, min, max)\\n117| \\n118| select{\\n119| \\n120| case <- ch1:\\n121| fmt.Println(\\\"routine 1 finished\\\")\\n122| \\n123| case <- ch2:\\n124| fmt.Println(\\\"routine 2 finished\\\")\\n125| }\\n126| fmt.Println(\\\"end of main function\\\")\\n127| fmt.Println(\\\"end of main function\\\")\\n128| \\n129| }\\n130| \\n```\\n\\n```\\n68| \\tch <- input_value\\n69| \\n70| // receive a value from channel\\n71| response := <- ch\\n72| fmt.Println(response)\\n73| \\n74| }\\n75| \\n76| // example 3\\n77| \\n78| package main\\n79| package main\\n80| \\n81| import(\\n82| \\\"fmt\\\"\\n83| \\\"time\\\"\\n84| \\\"math/rand\\\"\\n85| )\\n86| \\n87| func routine1(ch chan int, min int, max int) {\\n88| \\n```\\n\\n```\\n6| func dummy_func() {\\n7| \\tfmt.Println(\\\"dummy_func\\\")\\n8| }\\n9| func main() {\\n10| \\tgo dummy_func()\\n11| fmt.Println(\\\"end of main\\\")\\n12| }\\n13| \\n14| \\n15| // Example 1\\n16| package main\\n17| package main\\n18| \\n19| import \\\"fmt\\\"\\n20| \\n21| func dummy_func(ch chan int) {\\n22| // receive a value from channel\\n23| val := <-ch\\n24| \\tfmt.Println(val)\\n25| }\\n26| func main() {\\n```\\n\\n```\\n0| package main\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| // Example 0\\n6| func dummy_func() {\\n7| \\tfmt.Println(\\\"dummy_func\\\")\\n8| }\\n9| func main() {\\n10| \\tgo dummy_func()\\n```\\n\\n\",\"tool_call_id\":\"sourcegraph:1\",\"role\":\"tool\"}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"I","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" found","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" Go","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" repositories","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" with","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" `","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"func","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" main","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"`","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" declarations","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":".","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" The","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" search","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" returned","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" training","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" materials","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" and","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" examples","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" from","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" repositories","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" like","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" \"","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"G","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"olang","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"Training","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"\"","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" and","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" various","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" Go","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" code","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" samples","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" showing","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" different","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" main","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" function","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" implementations","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":",","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" including","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" basic","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" examples","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":",","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" function","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" expressions","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":",","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" and","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" channel","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"/g","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"or","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"outine","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" examples","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":".","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}]}
-
- data: {"id":"gen-1776258141-EocMmnknJyAvfRST0HjL","object":"chat.completion.chunk","created":1776258141,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}],"usage":{"prompt_tokens":12190,"completion_tokens":52,"total_tokens":12242,"cost":0.007444,"is_byok":false,"prompt_tokens_details":{"cached_tokens":9216,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.007444,"upstream_inference_prompt_cost":0.007314,"upstream_inference_completions_cost":0.00013},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 1.980978916s
diff --git a/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/update_a_file.yaml b/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/update_a_file.yaml
deleted file mode 100644
index 32950858b362ade748b93b572395ca7c674ed35b..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/openrouter-kimi-k2/update_a_file.yaml
+++ /dev/null
@@ -1,479 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 862
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nupdate the main.go file by changing the print to say hello from crush\n \n\n","role":"user"}],"model":"qwen/qwen3-next-80b-a3b-instruct","max_tokens":40,"stream_options":{"include_usage":true},"usage":{"include":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258057-xbejiwF312OZqf1rmqEG","object":"chat.completion.chunk","created":1776258057,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Parasail","choices":[{"index":0,"delta":{"content":"Update","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-xbejiwF312OZqf1rmqEG","object":"chat.completion.chunk","created":1776258057,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Parasail","choices":[{"index":0,"delta":{"content":" main","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-xbejiwF312OZqf1rmqEG","object":"chat.completion.chunk","created":1776258057,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Parasail","choices":[{"index":0,"delta":{"content":".go","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-xbejiwF312OZqf1rmqEG","object":"chat.completion.chunk","created":1776258057,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Parasail","choices":[{"index":0,"delta":{"content":" to","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-xbejiwF312OZqf1rmqEG","object":"chat.completion.chunk","created":1776258057,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Parasail","choices":[{"index":0,"delta":{"content":" print","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-xbejiwF312OZqf1rmqEG","object":"chat.completion.chunk","created":1776258057,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Parasail","choices":[{"index":0,"delta":{"content":" hello","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-xbejiwF312OZqf1rmqEG","object":"chat.completion.chunk","created":1776258057,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Parasail","choices":[{"index":0,"delta":{"content":" from","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-xbejiwF312OZqf1rmqEG","object":"chat.completion.chunk","created":1776258057,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Parasail","choices":[{"index":0,"delta":{"content":" crush","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-xbejiwF312OZqf1rmqEG","object":"chat.completion.chunk","created":1776258057,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Parasail","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}]}
-
- data: {"id":"gen-1776258057-xbejiwF312OZqf1rmqEG","object":"chat.completion.chunk","created":1776258057,"model":"qwen/qwen3-next-80b-a3b-instruct-2509","provider":"Parasail","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}],"usage":{"prompt_tokens":157,"completion_tokens":9,"total_tokens":166,"cost":0.0000256,"is_byok":false,"prompt_tokens_details":{"cached_tokens":0,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0000256,"upstream_inference_prompt_cost":0.0000157,"upstream_inference_completions_cost":0.0000099},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 875.837708ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30831
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/update_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"update the main.go file by changing the print to say hello from crush\",\"role\":\"user\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"I","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" need","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" to","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" first","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" check","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" what","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" the","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" current","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" main","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":".go","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" file","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" contains","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":",","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" then","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" update","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" the","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" print","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" statement","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":".","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"id":"view:0","type":"function","function":{"name":"view","arguments":""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"file"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"_path"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \"/"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"tmp"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/c"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"rush"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-test"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/Test"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"Coder"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"Agent"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/open"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"router"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-k"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"imi"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-k"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"2"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/update"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"_a"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"_file"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/main"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":".go"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\"}"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}]}
-
- data: {"id":"gen-1776258057-Pl1paaQhLzfUT5pVBCdg","object":"chat.completion.chunk","created":1776258057,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}],"usage":{"prompt_tokens":6497,"completion_tokens":52,"total_tokens":6549,"cost":0.0040282,"is_byok":false,"prompt_tokens_details":{"cached_tokens":6144,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0040282,"upstream_inference_prompt_cost":0.0038982,"upstream_inference_completions_cost":0.00013},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 2.0418605s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31310
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/update_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"update the main.go file by changing the print to say hello from crush\",\"role\":\"user\"},{\"content\":\"I need to first check what the current main.go file contains, then update the print statement.\",\"tool_calls\":[{\"id\":\"view:0\",\"function\":{\"arguments\":\"{\\\"file_path\\\": \\\"/tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/update_a_file/main.go\\\"}\",\"name\":\"view\"},\"type\":\"function\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"role\":\"assistant\"},{\"content\":\"\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\tfmt.Println(\\\"Hello, World!\\\")\\n 7|}\\n\\n\",\"tool_call_id\":\"view:0\",\"role\":\"tool\"}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"Now","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" I'll","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" update","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" the","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" print","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" statement","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" to","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" say","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" \"","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"Hello","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" from","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" Crush","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"\":","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"id":"edit:1","type":"function","function":{"name":"edit","arguments":""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"{\""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"file"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"_path"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \"/"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"tmp"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/c"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"rush"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-test"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/Test"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"Coder"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"Agent"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/open"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"router"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-k"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"imi"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"-k"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"2"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/update"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"_a"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"_file"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"/main"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":".go"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\","}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"old"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"_string"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \"\\"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"tf"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"mt"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":".Println"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"(\\\""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"Hello"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":","}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" World"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"!\\"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\")"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\","}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"new"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"_string"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\":"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" \"\\"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"tf"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"mt"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":".Println"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"(\\\""}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"Hello"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" from"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":" Crush"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\\\")"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":null,"role":"assistant","tool_calls":[{"index":0,"function":{"arguments":"\"}"}}]},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}]}
-
- data: {"id":"gen-1776258062-URKJLTo0LSNoM6klFUgA","object":"chat.completion.chunk","created":1776258062,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"tool_calls","native_finish_reason":"tool_calls"}],"usage":{"prompt_tokens":6617,"completion_tokens":75,"total_tokens":6692,"cost":0.0041577,"is_byok":false,"prompt_tokens_details":{"cached_tokens":6400,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0041577,"upstream_inference_prompt_cost":0.0039702,"upstream_inference_completions_cost":0.0001875},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 1.760861208s
-- id: 3
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31885
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/update_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"update the main.go file by changing the print to say hello from crush\",\"role\":\"user\"},{\"content\":\"I need to first check what the current main.go file contains, then update the print statement.\",\"tool_calls\":[{\"id\":\"view:0\",\"function\":{\"arguments\":\"{\\\"file_path\\\": \\\"/tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/update_a_file/main.go\\\"}\",\"name\":\"view\"},\"type\":\"function\"}],\"role\":\"assistant\"},{\"content\":\"\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\tfmt.Println(\\\"Hello, World!\\\")\\n 7|}\\n\\n\",\"tool_call_id\":\"view:0\",\"role\":\"tool\"},{\"content\":\"Now I'll update the print statement to say \\\"Hello from Crush\\\":\",\"tool_calls\":[{\"id\":\"edit:1\",\"function\":{\"arguments\":\"{\\\"file_path\\\": \\\"/tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/update_a_file/main.go\\\", \\\"old_string\\\": \\\"\\\\tfmt.Println(\\\\\\\"Hello, World!\\\\\\\")\\\", \\\"new_string\\\": \\\"\\\\tfmt.Println(\\\\\\\"Hello from Crush\\\\\\\")\\\"}\",\"name\":\"edit\"},\"type\":\"function\",\"cache_control\":{\"type\":\"ephemeral\"}}],\"role\":\"assistant\"},{\"content\":\"\\nContent replaced in file: /tmp/crush-test/TestCoderAgent/openrouter-kimi-k2/update_a_file/main.go\\n\\n\",\"tool_call_id\":\"edit:1\",\"role\":\"tool\"}],\"model\":\"moonshotai/kimi-k2-0905\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"usage\":{\"include\":true},\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://openrouter.ai/api/v1/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258068-VcRY5UPDJxPgYOdtVmU4","object":"chat.completion.chunk","created":1776258068,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"Done","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258068-VcRY5UPDJxPgYOdtVmU4","object":"chat.completion.chunk","created":1776258068,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":".","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258068-VcRY5UPDJxPgYOdtVmU4","object":"chat.completion.chunk","created":1776258068,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" The","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258068-VcRY5UPDJxPgYOdtVmU4","object":"chat.completion.chunk","created":1776258068,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" main","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258068-VcRY5UPDJxPgYOdtVmU4","object":"chat.completion.chunk","created":1776258068,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":".go","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258068-VcRY5UPDJxPgYOdtVmU4","object":"chat.completion.chunk","created":1776258068,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" file","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258068-VcRY5UPDJxPgYOdtVmU4","object":"chat.completion.chunk","created":1776258068,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" has","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258068-VcRY5UPDJxPgYOdtVmU4","object":"chat.completion.chunk","created":1776258068,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" been","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258068-VcRY5UPDJxPgYOdtVmU4","object":"chat.completion.chunk","created":1776258068,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" updated","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- : OPENROUTER PROCESSING
-
- data: {"id":"gen-1776258068-VcRY5UPDJxPgYOdtVmU4","object":"chat.completion.chunk","created":1776258068,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" to","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258068-VcRY5UPDJxPgYOdtVmU4","object":"chat.completion.chunk","created":1776258068,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" print","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258068-VcRY5UPDJxPgYOdtVmU4","object":"chat.completion.chunk","created":1776258068,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" \"","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258068-VcRY5UPDJxPgYOdtVmU4","object":"chat.completion.chunk","created":1776258068,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"Hello","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258068-VcRY5UPDJxPgYOdtVmU4","object":"chat.completion.chunk","created":1776258068,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" from","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258068-VcRY5UPDJxPgYOdtVmU4","object":"chat.completion.chunk","created":1776258068,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" Crush","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258068-VcRY5UPDJxPgYOdtVmU4","object":"chat.completion.chunk","created":1776258068,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"\"","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258068-VcRY5UPDJxPgYOdtVmU4","object":"chat.completion.chunk","created":1776258068,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" instead","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258068-VcRY5UPDJxPgYOdtVmU4","object":"chat.completion.chunk","created":1776258068,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" of","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258068-VcRY5UPDJxPgYOdtVmU4","object":"chat.completion.chunk","created":1776258068,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" \"","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258068-VcRY5UPDJxPgYOdtVmU4","object":"chat.completion.chunk","created":1776258068,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"Hello","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258068-VcRY5UPDJxPgYOdtVmU4","object":"chat.completion.chunk","created":1776258068,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":",","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258068-VcRY5UPDJxPgYOdtVmU4","object":"chat.completion.chunk","created":1776258068,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":" World","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258068-VcRY5UPDJxPgYOdtVmU4","object":"chat.completion.chunk","created":1776258068,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"!\"","role":"assistant"},"finish_reason":null,"native_finish_reason":null}]}
-
- data: {"id":"gen-1776258068-VcRY5UPDJxPgYOdtVmU4","object":"chat.completion.chunk","created":1776258068,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}]}
-
- data: {"id":"gen-1776258068-VcRY5UPDJxPgYOdtVmU4","object":"chat.completion.chunk","created":1776258068,"model":"moonshotai/kimi-k2-0905","provider":"Novita","system_fingerprint":"fpv0_e76aa377","choices":[{"index":0,"delta":{"content":"","role":"assistant"},"finish_reason":"stop","native_finish_reason":"stop"}],"usage":{"prompt_tokens":6739,"completion_tokens":24,"total_tokens":6763,"cost":0.0041034,"is_byok":false,"prompt_tokens_details":{"cached_tokens":6400,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0041034,"upstream_inference_prompt_cost":0.0040434,"upstream_inference_completions_cost":0.00006},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream
- status: 200 OK
- code: 200
- duration: 1.819550291s
diff --git a/internal/agent/testdata/TestCoderAgent/zai-glm4.6/bash_tool.yaml b/internal/agent/testdata/TestCoderAgent/zai-glm4.6/bash_tool.yaml
deleted file mode 100644
index 36d5ba7eec592b6f1b470edb2383b44dd2e47dde..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/zai-glm4.6/bash_tool.yaml
+++ /dev/null
@@ -1,413 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 841
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse bash to create a file named test.txt with content ''hello bash''. do not print its timestamp\n \n\n","role":"user"}],"model":"glm-4.5-air","max_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"20260415210258f0eca00bc5504951","created":1776258178,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n"}}]}
-
- data: {"id":"20260415210258f0eca00bc5504951","created":1776258178,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":"Create"}}]}
-
- data: {"id":"20260415210258f0eca00bc5504951","created":1776258178,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" file"}}]}
-
- data: {"id":"20260415210258f0eca00bc5504951","created":1776258178,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" with"}}]}
-
- data: {"id":"20260415210258f0eca00bc5504951","created":1776258178,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" bash"}}]}
-
- data: {"id":"20260415210258f0eca00bc5504951","created":1776258178,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" content"}}]}
-
- data: {"id":"20260415210258f0eca00bc5504951","created":1776258178,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"finish_reason":"stop","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":156,"completion_tokens":9,"total_tokens":165,"prompt_tokens_details":{"cached_tokens":130}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 660.719708ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30692
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/bash_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use bash to create a file named test.txt with content 'hello bash'. do not print its timestamp\",\"role\":\"user\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"The"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" user"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" wants"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" me"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" use"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" bash"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" create"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" a"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" file"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" named"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" test"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".txt"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" with"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" content"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" '"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"hello"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" bash"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"'."}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" They"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" also"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" specifically"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" mention"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" not"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" print"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" its"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" timestamp"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".\n\n"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"I"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"'ll"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" use"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" bash"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" tool"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" create"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" this"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" file"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" I"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" can"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" use"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" echo"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" write"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" content"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" file"}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_0942c571299f440a9c3cc635","index":0,"type":"function","function":{"name":"bash","arguments":"{\"command\":\"echo 'hello bash' > test.txt\",\"description\":\"Create test.txt with content\"}"}}]}}]}
-
- data: {"id":"202604152102582954017eb9914e9a","created":1776258179,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"tool_calls","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":7135,"completion_tokens":81,"total_tokens":7216,"prompt_tokens_details":{"cached_tokens":7003},"completion_tokens_details":{"reasoning_tokens":53}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 980.608083ms
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31273
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/bash_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use bash to create a file named test.txt with content 'hello bash'. do not print its timestamp\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_0942c571299f440a9c3cc635\",\"function\":{\"arguments\":\"{\\\"command\\\":\\\"echo 'hello bash' > test.txt\\\",\\\"description\\\":\\\"Create test.txt with content\\\"}\",\"name\":\"bash\"},\"type\":\"function\"}],\"role\":\"assistant\",\"reasoning_content\":\"The user wants me to use bash to create a file named test.txt with content 'hello bash'. They also specifically mention not to print its timestamp.\\n\\nI'll use the bash tool to create this file. I can use echo to write the content to the file.\"},{\"content\":\"no output\",\"tool_call_id\":\"call_0942c571299f440a9c3cc635\",\"role\":\"tool\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"The"}}]}
-
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" file"}}]}
-
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" has"}}]}
-
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" been"}}]}
-
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" created"}}]}
-
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" successfully"}}]}
-
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Let"}}]}
-
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" me"}}]}
-
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" verify"}}]}
-
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" it"}}]}
-
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" was"}}]}
-
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" created"}}]}
-
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" correctly"}}]}
-
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" by"}}]}
-
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" checking"}}]}
-
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" file"}}]}
-
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" exists"}}]}
-
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" and"}}]}
-
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" has"}}]}
-
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" right"}}]}
-
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" content"}}]}
-
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":","}}]}
-
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" without"}}]}
-
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" showing"}}]}
-
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" timestamps"}}]}
-
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_d135bd54908144eab3f80625","index":0,"type":"function","function":{"name":"bash","arguments":"{\"command\":\"cat test.txt\",\"description\":\"Verify file contents\"}"}}]}}]}
-
- data: {"id":"2026041521030108b184f114484763","created":1776258181,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"tool_calls","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":7222,"completion_tokens":50,"total_tokens":7272,"prompt_tokens_details":{"cached_tokens":7214},"completion_tokens_details":{"reasoning_tokens":29}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 1.190932041s
-- id: 3
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31817
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/bash_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use bash to create a file named test.txt with content 'hello bash'. do not print its timestamp\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_0942c571299f440a9c3cc635\",\"function\":{\"arguments\":\"{\\\"command\\\":\\\"echo 'hello bash' > test.txt\\\",\\\"description\\\":\\\"Create test.txt with content\\\"}\",\"name\":\"bash\"},\"type\":\"function\"}],\"role\":\"assistant\",\"reasoning_content\":\"The user wants me to use bash to create a file named test.txt with content 'hello bash'. They also specifically mention not to print its timestamp.\\n\\nI'll use the bash tool to create this file. I can use echo to write the content to the file.\"},{\"content\":\"no output\",\"tool_call_id\":\"call_0942c571299f440a9c3cc635\",\"role\":\"tool\"},{\"tool_calls\":[{\"id\":\"call_d135bd54908144eab3f80625\",\"function\":{\"arguments\":\"{\\\"command\\\":\\\"cat test.txt\\\",\\\"description\\\":\\\"Verify file contents\\\"}\",\"name\":\"bash\"},\"type\":\"function\"}],\"role\":\"assistant\",\"reasoning_content\":\"The file has been created successfully. Let me verify it was created correctly by checking the file exists and has the right content, without showing timestamps.\"},{\"content\":\"hello bash\\n\\n\\n/tmp/crush-test/TestCoderAgent/zai-glm4.6/bash_tool\",\"tool_call_id\":\"call_d135bd54908144eab3f80625\",\"role\":\"tool\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"The"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" file"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" has"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" been"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" created"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" with"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" correct"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" content"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" The"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" user"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" asked"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" not"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" print"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" its"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" timestamp"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":","}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" which"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" I"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"'ve"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" honored"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" by"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" only"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" showing"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" file"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" content"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" with"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" cat"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":","}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" not"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" using"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" ls"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" with"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" timestamp"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" options"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"Done"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"."}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" File"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" created"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" with"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" content"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" '"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"hello"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" bash"}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"'."}}]}
-
- data: {"id":"20260415210304554f9c881a254046","created":1776258184,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"stop","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":7301,"completion_tokens":51,"total_tokens":7352,"prompt_tokens_details":{"cached_tokens":7270},"completion_tokens_details":{"reasoning_tokens":39}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 1.25688675s
diff --git a/internal/agent/testdata/TestCoderAgent/zai-glm4.6/download_tool.yaml b/internal/agent/testdata/TestCoderAgent/zai-glm4.6/download_tool.yaml
deleted file mode 100644
index 7bbf10d5c0256c4a89fdba72bb0e526dad4f3d5d..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/zai-glm4.6/download_tool.yaml
+++ /dev/null
@@ -1,502 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30717
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/download_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"download the file from https://example-files.online-convert.com/document/txt/example.txt and save it as example.txt\",\"role\":\"user\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"The"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" user"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" wants"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" me"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" download"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" a"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" file"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" from"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" a"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" URL"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" and"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" save"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" it"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" locally"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" I"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" need"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" use"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" `"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"download"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"`"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" tool"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" for"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" this"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" task"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".\n\n"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"The"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" URL"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" is"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":":"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" https"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"://"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"example"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"-files"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".online"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"-"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"convert"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".com"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"/document"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"/t"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"xt"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"/example"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".txt"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"The"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" save"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" path"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" should"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" be"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":":"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" example"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".txt"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" ("}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"relative"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" current"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" working"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" directory"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":")\n\n"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"According"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" my"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" instructions"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":","}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" I"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" need"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" use"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" download"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" tool"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" with"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" `"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"url"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"`"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" and"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" `"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"file"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"_path"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"`"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" parameters"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" The"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" file"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"_path"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" should"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" be"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" an"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" absolute"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" path"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":","}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" so"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" I"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"'ll"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" use"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" working"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" directory"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" /"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"tmp"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"/c"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"rush"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"-test"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"/Test"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Coder"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Agent"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"/z"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"ai"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"-g"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"lm"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"4"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"6"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"/download"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"_tool"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" combined"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" with"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" example"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".txt"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".\n\n"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Let"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" me"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" make"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" download"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" call"}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_d34c6bf5bbfc4c0aa073170d","index":0,"type":"function","function":{"name":"download","arguments":"{\"file_path\":\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/download_tool/example.txt\",\"url\":\"https://example-files.online-convert.com/document/txt/example.txt\"}"}}]}}]}
-
- data: {"id":"2026041521030733940ac728164b92","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"tool_calls","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":7138,"completion_tokens":179,"total_tokens":7317,"prompt_tokens_details":{"cached_tokens":7003},"completion_tokens_details":{"reasoning_tokens":132}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 1.004082084s
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 0
- host: example-files.online-convert.com
- headers:
- User-Agent:
- - crush/1.0
- url: https://example-files.online-convert.com/document/txt/example.txt
- method: GET
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |-
- TXT test file
- Purpose: Provide example of this file type
- Document file type: TXT
- Version: 1.0
- Remark:
-
- Example content:
- The names "John Doe" for males, "Jane Doe" or "Jane Roe" for females, or "Jonnie Doe" and "Janie Doe" for children, or just "Doe" non-gender-specifically are used as placeholder names for a party whose true identity is unknown or must be withheld in a legal action, case, or discussion. The names are also used to refer to acorpse or hospital patient whose identity is unknown. This practice is widely used in the United States and Canada, but is rarely used in other English-speaking countries including the United Kingdom itself, from where the use of "John Doe" in a legal context originates. The names Joe Bloggs or John Smith are used in the UK instead, as well as in Australia and New Zealand.
-
- John Doe is sometimes used to refer to a typical male in other contexts as well, in a similar manner to John Q. Public, known in Great Britain as Joe Public, John Smith or Joe Bloggs. For example, the first name listed on a form is often John Doe, along with a fictional address or other fictional information to provide an example of how to fill in the form. The name is also used frequently in popular culture, for example in the Frank Capra film Meet John Doe. John Doe was also the name of a 2002 American television series.
-
- Similarly, a child or baby whose identity is unknown may be referred to as Baby Doe. A notorious murder case in Kansas City, Missouri, referred to the baby victim as Precious Doe. Other unidentified female murder victims are Cali Doe and Princess Doe. Additional persons may be called James Doe, Judy Doe, etc. However, to avoid possible confusion, if two anonymous or unknown parties are cited in a specific case or action, the surnames Doe and Roe may be used simultaneously; for example, "John Doe v. Jane Roe". If several anonymous parties are referenced, they may simply be labelled John Doe #1, John Doe #2, etc. (the U.S. Operation Delego cited 21 (numbered) "John Doe"s) or labelled with other variants of Doe / Roe / Poe / etc. Other early alternatives such as John Stiles and Richard Miles are now rarely used, and Mary Major has been used in some American federal cases.
-
-
-
- File created by https://www.online-convert.com
- More example files: https://www.online-convert.com/file-type
- Text of Example content: Wikipedia (https://en.wikipedia.org/wiki/John_Doe)
- License: Attribution-ShareAlike 4.0 (https://creativecommons.org/licenses/by-sa/4.0/)
-
- Feel free to use and share the file according to the license above.
- headers:
- Content-Type:
- - text/plain; charset=UTF-8
- status: 200 OK
- code: 200
- duration: 726.913ms
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31792
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/download_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"download the file from https://example-files.online-convert.com/document/txt/example.txt and save it as example.txt\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_d34c6bf5bbfc4c0aa073170d\",\"function\":{\"arguments\":\"{\\\"file_path\\\":\\\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/download_tool/example.txt\\\",\\\"url\\\":\\\"https://example-files.online-convert.com/document/txt/example.txt\\\"}\",\"name\":\"download\"},\"type\":\"function\"}],\"role\":\"assistant\",\"reasoning_content\":\"The user wants me to download a file from a URL and save it locally. I need to use the `download` tool for this task.\\n\\nThe URL is: https://example-files.online-convert.com/document/txt/example.txt\\nThe save path should be: example.txt (relative to the current working directory)\\n\\nAccording to my instructions, I need to use the download tool with the `url` and `file_path` parameters. The file_path should be an absolute path, so I'll use the working directory /tmp/crush-test/TestCoderAgent/zai-glm4.6/download_tool combined with example.txt.\\n\\nLet me make the download call.\"},{\"content\":\"Successfully downloaded 2574 bytes to example.txt (Content-Type: text/plain; charset=UTF-8)\",\"tool_call_id\":\"call_d34c6bf5bbfc4c0aa073170d\",\"role\":\"tool\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"The"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" download"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" was"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" successful"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" The"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" file"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" has"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" been"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" downloaded"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" as"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" example"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".txt"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" with"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" "}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"25"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"7"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"4"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" bytes"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" The"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" task"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" is"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" complete"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"Done"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"."}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" Download"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"ed"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" "}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"25"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"7"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"4"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" bytes"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" to"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" example"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":".txt"}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"."}}]}
-
- data: {"id":"202604152103115ed866e3676a49ca","created":1776258191,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"stop","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":7343,"completion_tokens":40,"total_tokens":7383,"prompt_tokens_details":{"cached_tokens":7315},"completion_tokens_details":{"reasoning_tokens":25}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 888.450208ms
-- id: 3
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 862
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\ndownload the file from https://example-files.online-convert.com/document/txt/example.txt and save it as example.txt\n \n\n","role":"user"}],"model":"glm-4.5-air","max_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"20260415210307b4395ecb42984557","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n"}}]}
-
- data: {"id":"20260415210307b4395ecb42984557","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":"Download"}}]}
-
- data: {"id":"20260415210307b4395ecb42984557","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" and"}}]}
-
- data: {"id":"20260415210307b4395ecb42984557","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" save"}}]}
-
- data: {"id":"20260415210307b4395ecb42984557","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" text"}}]}
-
- data: {"id":"20260415210307b4395ecb42984557","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" file"}}]}
-
- data: {"id":"20260415210307b4395ecb42984557","created":1776258187,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"finish_reason":"stop","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":159,"completion_tokens":9,"total_tokens":168,"prompt_tokens_details":{"cached_tokens":130}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 9.810928792s
diff --git a/internal/agent/testdata/TestCoderAgent/zai-glm4.6/fetch_tool.yaml b/internal/agent/testdata/TestCoderAgent/zai-glm4.6/fetch_tool.yaml
deleted file mode 100644
index 02d8bf7edb9e5d7d33cd7d664febc498f90c2cb4..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/zai-glm4.6/fetch_tool.yaml
+++ /dev/null
@@ -1,631 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30735
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/fetch_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"fetch the content from https://example-files.online-convert.com/website/html/example.html and tell me if it contains the word 'John Doe'\",\"role\":\"user\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"The"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" user"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" wants"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" me"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" fetch"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" content"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" from"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" a"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" URL"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" and"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" check"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" if"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" it"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" contains"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" a"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" specific"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" word"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" '"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"John"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Doe"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"'."}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" I"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" should"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" use"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" fetch"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" tool"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" get"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" content"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" from"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" URL"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":","}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" then"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" analyze"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" it"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" see"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" if"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" it"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" contains"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" that"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" word"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".\n\n"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Let"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" me"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" use"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" fetch"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" tool"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" with"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" URL"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" provided"}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_4cede8dc358a4fffb4c5b018","index":0,"type":"function","function":{"name":"fetch","arguments":"{\"format\":\"text\",\"url\":\"https://example-files.online-convert.com/website/html/example.html\"}"}}]}}]}
-
- data: {"id":"202604152103176b7373a38f794a5f","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"tool_calls","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":7145,"completion_tokens":87,"total_tokens":7232,"prompt_tokens_details":{"cached_tokens":7003},"completion_tokens_details":{"reasoning_tokens":58}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 891.405ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 0
- host: example-files.online-convert.com
- headers:
- User-Agent:
- - crush/1.0
- url: https://example-files.online-convert.com/website/html/example.html
- method: GET
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: |-
-
-
-
-
-
- HTML test file
-
- Purpose: Provide example of this file type
- Document file type: HTML
- Version: 1.0
- Remark:
-
-
-
- Example content:
- The names "John Doe" for males, "Jane Doe" or "Jane Roe" for females, or "Jonnie Doe" and "Janie Doe" for children, or just "Doe" non-gender-specifically are used as placeholder names for a party whose true identity is unknown or must be withheld in a legal action, case, or discussion. The names are also used to refer to acorpse or hospital patient whose identity is unknown. This practice is widely used in the United States and Canada, but is rarely used in other English-speaking countries including the United Kingdom itself, from where the use of "John Doe" in a legal context originates. The names Joe Bloggs or John Smith are used in the UK instead, as well as in Australia and New Zealand.
-
- John Doe is sometimes used to refer to a typical male in other contexts as well, in a similar manner to John Q. Public, known in Great Britain as Joe Public, John Smith or Joe Bloggs. For example, the first name listed on a form is often John Doe, along with a fictional address or other fictional information to provide an example of how to fill in the form. The name is also used frequently in popular culture, for example in the Frank Capra film Meet John Doe. John Doe was also the name of a 2002 American television series.
-
- Similarly, a child or baby whose identity is unknown may be referred to as Baby Doe. A notorious murder case in Kansas City, Missouri, referred to the baby victim as Precious Doe. Other unidentified female murder victims are Cali Doe and Princess Doe. Additional persons may be called James Doe, Judy Doe, etc. However, to avoid possible confusion, if two anonymous or unknown parties are cited in a specific case or action, the surnames Doe and Roe may be used simultaneously; for example, "John Doe v. Jane Roe". If several anonymous parties are referenced, they may simply be labelled John Doe #1, John Doe #2, etc. (the U.S. Operation Delego cited 21 (numbered) "John Doe"s) or labelled with other variants of Doe / Roe / Poe / etc. Other early alternatives such as John Stiles and Richard Miles are now rarely used, and Mary Major has been used in some American federal cases.
-
-
-
- File created by https://www.online-convert.com
- More example files: https://www.online-convert.com/file-type
- Text of "Example content": Wikipedia
- License: Attribution-ShareAlike 3.0 Unported
-
-
-
- Feel free to use and share the file according to license above.
-
-
-
- headers:
- Content-Type:
- - text/html; charset=UTF-8
- status: 200 OK
- code: 200
- duration: 399.92475ms
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 883
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nfetch the content from https://example-files.online-convert.com/website/html/example.html and tell me if it contains the word ''John Doe''\n \n\n","role":"user"}],"model":"glm-4.5-air","max_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"202604152103178d96496161a54297","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n"}}]}
-
- data: {"id":"202604152103178d96496161a54297","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":"Check"}}]}
-
- data: {"id":"202604152103178d96496161a54297","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" if"}}]}
-
- data: {"id":"202604152103178d96496161a54297","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" '"}}]}
-
- data: {"id":"202604152103178d96496161a54297","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":"John"}}]}
-
- data: {"id":"202604152103178d96496161a54297","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" Doe"}}]}
-
- data: {"id":"202604152103178d96496161a54297","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":"'"}}]}
-
- data: {"id":"202604152103178d96496161a54297","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" in"}}]}
-
- data: {"id":"202604152103178d96496161a54297","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" HTML"}}]}
-
- data: {"id":"202604152103178d96496161a54297","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" content"}}]}
-
- data: {"id":"202604152103178d96496161a54297","created":1776258197,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"finish_reason":"stop","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":164,"completion_tokens":13,"total_tokens":177,"prompt_tokens_details":{"cached_tokens":130}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 3.154837667s
-- id: 3
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 33835
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/fetch_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"fetch the content from https://example-files.online-convert.com/website/html/example.html and tell me if it contains the word 'John Doe'\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_4cede8dc358a4fffb4c5b018\",\"function\":{\"arguments\":\"{\\\"format\\\":\\\"text\\\",\\\"url\\\":\\\"https://example-files.online-convert.com/website/html/example.html\\\"}\",\"name\":\"fetch\"},\"type\":\"function\"}],\"role\":\"assistant\",\"reasoning_content\":\"The user wants me to fetch content from a URL and check if it contains a specific word 'John Doe'. I should use the fetch tool to get the content from the URL, then analyze it to see if it contains that word.\\n\\nLet me use the fetch tool with the URL provided.\"},{\"content\":\"HTML test file Purpose: Provide example of this file type Document file type: HTML Version: 1.0 Remark: Example content: The names \\\"John Doe\\\" for males, \\\"Jane Doe\\\" or \\\"Jane Roe\\\" for females, or \\\"Jonnie Doe\\\" and \\\"Janie Doe\\\" for children, or just \\\"Doe\\\" non-gender-specifically are used as placeholder names for a party whose true identity is unknown or must be withheld in a legal action, case, or discussion. The names are also used to refer to acorpse or hospital patient whose identity is unknown. This practice is widely used in the United States and Canada, but is rarely used in other English-speaking countries including the United Kingdom itself, from where the use of \\\"John Doe\\\" in a legal context originates. The names Joe Bloggs or John Smith are used in the UK instead, as well as in Australia and New Zealand. John Doe is sometimes used to refer to a typical male in other contexts as well, in a similar manner to John Q. Public, known in Great Britain as Joe Public, John Smith or Joe Bloggs. For example, the first name listed on a form is often John Doe, along with a fictional address or other fictional information to provide an example of how to fill in the form. The name is also used frequently in popular culture, for example in the Frank Capra film Meet John Doe. John Doe was also the name of a 2002 American television series. Similarly, a child or baby whose identity is unknown may be referred to as Baby Doe. A notorious murder case in Kansas City, Missouri, referred to the baby victim as Precious Doe. Other unidentified female murder victims are Cali Doe and Princess Doe. Additional persons may be called James Doe, Judy Doe, etc. However, to avoid possible confusion, if two anonymous or unknown parties are cited in a specific case or action, the surnames Doe and Roe may be used simultaneously; for example, \\\"John Doe v. Jane Roe\\\". If several anonymous parties are referenced, they may simply be labelled John Doe #1, John Doe #2, etc. (the U.S. Operation Delego cited 21 (numbered) \\\"John Doe\\\"s) or labelled with other variants of Doe / Roe / Poe / etc. Other early alternatives such as John Stiles and Richard Miles are now rarely used, and Mary Major has been used in some American federal cases. File created by https://www.online-convert.com More example files: https://www.online-convert.com/file-type Text of \\\"Example content\\\": Wikipedia License: Attribution-ShareAlike 3.0 Unported Feel free to use and share the file according to license above.\",\"tool_call_id\":\"call_4cede8dc358a4fffb4c5b018\",\"role\":\"tool\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Looking"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" at"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" fetched"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" content"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":","}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" I"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" can"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" clearly"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" see"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" that"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" phrase"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" '"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"John"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Doe"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"'"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" appears"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" multiple"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" times"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" in"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" text"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":":\n"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"1"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" \""}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"The"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" names"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" '"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"John"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Doe"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"'"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" for"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" males"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\"\n"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"2"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" \""}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"the"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" use"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" of"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" '"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"John"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Doe"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"'"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" in"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" a"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" legal"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" context"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" originates"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\"\n"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"3"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" \""}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"John"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Doe"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" is"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" sometimes"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" used"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\"\n"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"4"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" \""}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"For"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" example"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":","}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" first"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" name"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" listed"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" on"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" a"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" form"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" is"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" often"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" John"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Doe"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\"\n"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"5"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" \""}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"in"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Frank"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Cap"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"ra"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" film"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Meet"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" John"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Doe"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\"\n"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"6"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" \""}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"John"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Doe"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" was"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" also"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" name"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" of"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" a"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" "}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"200"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"2"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" American"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" television"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" series"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\"\n"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"7"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" \""}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"they"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" may"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" simply"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" be"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" labelled"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" John"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Doe"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" #"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"1"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":","}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" John"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Doe"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" #"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"2"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":","}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" etc"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".\"\n\n"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"So"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" yes"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":","}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" content"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" definitely"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" contains"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" word"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" '"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"John"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Doe"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"'."}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"Yes"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":","}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" the"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" content"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" contains"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" the"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" phrase"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" '"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"John"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" Doe"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"'"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" multiple"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" times"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" throughout"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" the"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" text"}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"."}}]}
-
- data: {"id":"20260415210320abc6be116675416c","created":1776258200,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"stop","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":7788,"completion_tokens":162,"total_tokens":7950,"prompt_tokens_details":{"cached_tokens":7230},"completion_tokens_details":{"reasoning_tokens":143}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 856.122083ms
diff --git a/internal/agent/testdata/TestCoderAgent/zai-glm4.6/glob_tool.yaml b/internal/agent/testdata/TestCoderAgent/zai-glm4.6/glob_tool.yaml
deleted file mode 100644
index f83890ddb2a0b0eef50398cc377bf47c4ec3801e..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/zai-glm4.6/glob_tool.yaml
+++ /dev/null
@@ -1,282 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 802
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse glob to find all .go files in the current directory\n \n\n","role":"user"}],"model":"glm-4.5-air","max_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"20260415210323904ea969925e4e79","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n"}}]}
-
- data: {"id":"20260415210323904ea969925e4e79","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":"Find"}}]}
-
- data: {"id":"20260415210323904ea969925e4e79","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" Go"}}]}
-
- data: {"id":"20260415210323904ea969925e4e79","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" files"}}]}
-
- data: {"id":"20260415210323904ea969925e4e79","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" with"}}]}
-
- data: {"id":"20260415210323904ea969925e4e79","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" glob"}}]}
-
- data: {"id":"20260415210323904ea969925e4e79","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"finish_reason":"stop","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":148,"completion_tokens":9,"total_tokens":157,"prompt_tokens_details":{"cached_tokens":130}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 782.422041ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30653
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/glob_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use glob to find all .go files in the current directory\",\"role\":\"user\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"The"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" user"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" wants"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" use"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" glob"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" find"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" all"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" ."}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"go"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" files"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" in"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" current"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" directory"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" I"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" should"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" use"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" glob"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" tool"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" with"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" pattern"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" \"*."}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"go"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\""}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" in"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" current"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" working"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" directory"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" (/"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"tmp"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"/c"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"rush"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"-test"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"/Test"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Coder"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Agent"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"/z"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"ai"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"-g"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"lm"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"4"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"6"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"/g"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"lob"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"_tool"}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":")."}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_5dbb430cc9be41729bbad356","index":0,"type":"function","function":{"name":"glob","arguments":"{\"path\":\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/glob_tool\",\"pattern\":\"*.go\"}"}}]}}]}
-
- data: {"id":"20260415210323721bdb91d7ce4098","created":1776258203,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"tool_calls","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":7128,"completion_tokens":87,"total_tokens":7215,"prompt_tokens_details":{"cached_tokens":7003},"completion_tokens_details":{"reasoning_tokens":53}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 866.243959ms
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31248
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/glob_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use glob to find all .go files in the current directory\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_5dbb430cc9be41729bbad356\",\"function\":{\"arguments\":\"{\\\"path\\\":\\\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/glob_tool\\\",\\\"pattern\\\":\\\"*.go\\\"}\",\"name\":\"glob\"},\"type\":\"function\"}],\"role\":\"assistant\",\"reasoning_content\":\"The user wants to use glob to find all .go files in the current directory. I should use the glob tool with the pattern \\\"*.go\\\" in the current working directory (/tmp/crush-test/TestCoderAgent/zai-glm4.6/glob_tool).\"},{\"content\":\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/glob_tool/main.go\",\"tool_call_id\":\"call_5dbb430cc9be41729bbad356\",\"role\":\"tool\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"The"}}]}
-
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" glob"}}]}
-
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" tool"}}]}
-
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" found"}}]}
-
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" one"}}]}
-
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" ."}}]}
-
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"go"}}]}
-
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" file"}}]}
-
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" in"}}]}
-
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" current"}}]}
-
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" directory"}}]}
-
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":":"}}]}
-
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" main"}}]}
-
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".go"}}]}
-
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" I"}}]}
-
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" should"}}]}
-
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" present"}}]}
-
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" this"}}]}
-
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" result"}}]}
-
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" user"}}]}
-
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" conc"}}]}
-
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"is"}}]}
-
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"ely"}}]}
-
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"main"}}]}
-
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":".go"}}]}
-
- data: {"id":"2026041521032649ee2a7354694c69","created":1776258206,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"stop","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":7238,"completion_tokens":32,"total_tokens":7270,"prompt_tokens_details":{"cached_tokens":7213},"completion_tokens_details":{"reasoning_tokens":28}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 4.424121167s
diff --git a/internal/agent/testdata/TestCoderAgent/zai-glm4.6/grep_tool.yaml b/internal/agent/testdata/TestCoderAgent/zai-glm4.6/grep_tool.yaml
deleted file mode 100644
index a1341c1bcc4e1fcc626770d52f8a01e2d778f577..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/zai-glm4.6/grep_tool.yaml
+++ /dev/null
@@ -1,384 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 800
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse grep to search for the word ''package'' in go files\n \n\n","role":"user"}],"model":"glm-4.5-air","max_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"20260415210332d05b163091ee424f","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n"}}]}
-
- data: {"id":"20260415210332d05b163091ee424f","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":"Search"}}]}
-
- data: {"id":"20260415210332d05b163091ee424f","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" '"}}]}
-
- data: {"id":"20260415210332d05b163091ee424f","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":"package"}}]}
-
- data: {"id":"20260415210332d05b163091ee424f","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":"'"}}]}
-
- data: {"id":"20260415210332d05b163091ee424f","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" in"}}]}
-
- data: {"id":"20260415210332d05b163091ee424f","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" Go"}}]}
-
- data: {"id":"20260415210332d05b163091ee424f","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" files"}}]}
-
- data: {"id":"20260415210332d05b163091ee424f","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" with"}}]}
-
- data: {"id":"20260415210332d05b163091ee424f","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" grep"}}]}
-
- data: {"id":"20260415210332d05b163091ee424f","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"finish_reason":"stop","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":149,"completion_tokens":13,"total_tokens":162,"prompt_tokens_details":{"cached_tokens":131}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 692.243375ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30651
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/grep_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use grep to search for the word 'package' in go files\",\"role\":\"user\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"The"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" user"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" wants"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" me"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" use"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" grep"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" search"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" for"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" word"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" '"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"package"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"'"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" in"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Go"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" files"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" I"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" should"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" use"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" grep"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" tool"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" with"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":":\n"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"-"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" pattern"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":":"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" '"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"package"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"'"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" ("}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"literal"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" text"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":")\n"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"-"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" include"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":":"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" '*."}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"go"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"'"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" ("}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"to"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" filter"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Go"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" files"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":")\n\n"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Let"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" me"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" do"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" this"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" search"}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_714eb2f1e3bc4c18baa59b96","index":0,"type":"function","function":{"name":"grep","arguments":"{\"include\":\"*.go\",\"literal_text\":true,\"pattern\":\"package\"}"}}]}}]}
-
- data: {"id":"2026041521033279890e9c7f884654","created":1776258212,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"tool_calls","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":7129,"completion_tokens":80,"total_tokens":7209,"prompt_tokens_details":{"cached_tokens":7004},"completion_tokens_details":{"reasoning_tokens":55}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 931.857958ms
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31275
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/grep_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use grep to search for the word 'package' in go files\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_714eb2f1e3bc4c18baa59b96\",\"function\":{\"arguments\":\"{\\\"include\\\":\\\"*.go\\\",\\\"literal_text\\\":true,\\\"pattern\\\":\\\"package\\\"}\",\"name\":\"grep\"},\"type\":\"function\"}],\"role\":\"assistant\",\"reasoning_content\":\"The user wants me to use grep to search for the word 'package' in Go files. I should use the grep tool with:\\n- pattern: 'package' (literal text)\\n- include: '*.go' (to filter Go files)\\n\\nLet me do this search.\"},{\"content\":\"Found 1 matches\\n/tmp/crush-test/TestCoderAgent/zai-glm4.6/grep_tool/main.go:\\n Line 1, Char 1: package main\\n\",\"tool_call_id\":\"call_714eb2f1e3bc4c18baa59b96\",\"role\":\"tool\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"The"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" grep"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" search"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" found"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" one"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" match"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" in"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" file"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" main"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".go"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" at"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" line"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" "}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"1"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" The"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" word"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" '"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"package"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"'"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" appears"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" in"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" \""}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"package"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" main"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\"."}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" This"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" is"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" expected"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" output"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" -"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" it"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" found"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" package"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" declaration"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" at"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" beginning"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" of"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" a"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Go"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" file"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".\n\n"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"I"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" should"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" provide"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" a"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" concise"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" response"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" showing"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" what"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" was"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" found"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"Found"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" "}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"1"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" match"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" in"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" `"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"main"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":".go"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":":"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"1"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"`"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":":\n"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"``"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"`\n"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"package"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" main"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"\n"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"```"}}]}
-
- data: {"id":"20260415210335da3e399920b34252","created":1776258215,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"stop","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":7250,"completion_tokens":77,"total_tokens":7327,"prompt_tokens_details":{"cached_tokens":7207},"completion_tokens_details":{"reasoning_tokens":57}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 1.188370167s
diff --git a/internal/agent/testdata/TestCoderAgent/zai-glm4.6/ls_tool.yaml b/internal/agent/testdata/TestCoderAgent/zai-glm4.6/ls_tool.yaml
deleted file mode 100644
index 741905f77fdce914687efbdabd831719b96a1b61..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/zai-glm4.6/ls_tool.yaml
+++ /dev/null
@@ -1,286 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30645
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/ls_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use ls to list the files in the current directory\",\"role\":\"user\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"The"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" user"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" wants"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" me"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" list"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" files"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" in"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" current"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" directory"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" using"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" ls"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" tool"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" I"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" should"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" use"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" ls"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" tool"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" with"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" no"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" parameters"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" list"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" files"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" in"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" current"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" working"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" directory"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" (/"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"tmp"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"/c"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"rush"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"-test"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"/Test"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Coder"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Agent"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"/z"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"ai"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"-g"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"lm"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"4"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"6"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"/"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"ls"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"_tool"}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":")."}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_8314e65ae63c4e8e89f02ed4","index":0,"type":"function","function":{"name":"ls","arguments":"{\"path\":\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/ls_tool\"}"}}]}}]}
-
- data: {"id":"20260415210337622558783e864b3a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"tool_calls","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":7126,"completion_tokens":79,"total_tokens":7205,"prompt_tokens_details":{"cached_tokens":7003},"completion_tokens_details":{"reasoning_tokens":52}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 955.599417ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31241
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/ls_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use ls to list the files in the current directory\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_8314e65ae63c4e8e89f02ed4\",\"function\":{\"arguments\":\"{\\\"path\\\":\\\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/ls_tool\\\"}\",\"name\":\"ls\"},\"type\":\"function\"}],\"role\":\"assistant\",\"reasoning_content\":\"The user wants me to list files in the current directory using the ls tool. I should use the ls tool with no parameters to list files in the current working directory (/tmp/crush-test/TestCoderAgent/zai-glm4.6/ls_tool).\"},{\"content\":\"\\n- /tmp/crush-test/TestCoderAgent/zai-glm4.6/ls_tool/\\n - go.mod\\n - main.go\\n\",\"tool_call_id\":\"call_8314e65ae63c4e8e89f02ed4\",\"role\":\"tool\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"The"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" ls"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" tool"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" returned"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" two"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" files"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" in"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" current"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" directory"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":":\n"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"-"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" go"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".mod"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"-"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" main"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".go"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n\n"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"I"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" should"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" provide"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" a"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" concise"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" answer"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" user"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"go"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":".mod"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":","}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" main"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":".go"}}]}
-
- data: {"id":"20260415210339051a50fc7e94468e","created":1776258219,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"stop","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":7240,"completion_tokens":36,"total_tokens":7276,"prompt_tokens_details":{"cached_tokens":7203},"completion_tokens_details":{"reasoning_tokens":29}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 880.156458ms
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 796
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse ls to list the files in the current directory\n \n\n","role":"user"}],"model":"glm-4.5-air","max_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"20260415210337b7dc84b8bda5470a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n"}}]}
-
- data: {"id":"20260415210337b7dc84b8bda5470a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":"List"}}]}
-
- data: {"id":"20260415210337b7dc84b8bda5470a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" files"}}]}
-
- data: {"id":"20260415210337b7dc84b8bda5470a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" in"}}]}
-
- data: {"id":"20260415210337b7dc84b8bda5470a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" directory"}}]}
-
- data: {"id":"20260415210337b7dc84b8bda5470a","created":1776258217,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"finish_reason":"stop","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":146,"completion_tokens":8,"total_tokens":154,"prompt_tokens_details":{"cached_tokens":131}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 10.769523041s
diff --git a/internal/agent/testdata/TestCoderAgent/zai-glm4.6/multiedit_tool.yaml b/internal/agent/testdata/TestCoderAgent/zai-glm4.6/multiedit_tool.yaml
deleted file mode 100644
index b9e443b7c5a56e92f79d1c011f4d40e3f609136c..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/zai-glm4.6/multiedit_tool.yaml
+++ /dev/null
@@ -1,602 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 875
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse multiedit to change ''Hello, World!'' to ''Hello, Crush!'' and add a comment ''// Greeting'' above the fmt.Println line in main.go\n \n\n","role":"user"}],"model":"glm-4.5-air","max_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"202604152103499ea9502e7e524c5d","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n"}}]}
-
- data: {"id":"202604152103499ea9502e7e524c5d","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":"Mult"}}]}
-
- data: {"id":"202604152103499ea9502e7e524c5d","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":"ied"}}]}
-
- data: {"id":"202604152103499ea9502e7e524c5d","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":"it"}}]}
-
- data: {"id":"202604152103499ea9502e7e524c5d","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":":"}}]}
-
- data: {"id":"202604152103499ea9502e7e524c5d","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" G"}}]}
-
- data: {"id":"202604152103499ea9502e7e524c5d","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":"reeting"}}]}
-
- data: {"id":"202604152103499ea9502e7e524c5d","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" Code"}}]}
-
- data: {"id":"202604152103499ea9502e7e524c5d","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" Modification"}}]}
-
- data: {"id":"202604152103499ea9502e7e524c5d","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"finish_reason":"stop","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":169,"completion_tokens":12,"total_tokens":181,"prompt_tokens_details":{"cached_tokens":131}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 849.242583ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30731
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/multiedit_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use multiedit to change 'Hello, World!' to 'Hello, Crush!' and add a comment '// Greeting' above the fmt.Println line in main.go\",\"role\":\"user\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"The"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" user"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" wants"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" me"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":":\n"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"1"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Change"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" '"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Hello"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":","}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" World"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"!'"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" '"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Hello"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":","}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Crush"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"!'"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" \n"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"2"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Add"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" a"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" comment"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" '//"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" G"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"reeting"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"'"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" above"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" fmt"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".Println"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" line"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"3"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Use"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" mult"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"ied"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"it"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" tool"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"4"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Do"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" this"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" in"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" main"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".go"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n\n"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"First"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":","}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" I"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" need"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" read"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" main"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".go"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" file"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" see"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" its"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" current"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" state"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":","}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" then"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" use"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" mult"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"ied"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"it"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" make"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" changes"}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_f8bd4bb20bae4a51a1c0a858","index":0,"type":"function","function":{"name":"view","arguments":"{\"file_path\":\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/multiedit_tool/main.go\"}"}}]}}]}
-
- data: {"id":"20260415210349421ab2784b784a7e","created":1776258229,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"tool_calls","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":7151,"completion_tokens":111,"total_tokens":7262,"prompt_tokens_details":{"cached_tokens":7003},"completion_tokens_details":{"reasoning_tokens":79}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 1.160542625s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31486
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/multiedit_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use multiedit to change 'Hello, World!' to 'Hello, Crush!' and add a comment '// Greeting' above the fmt.Println line in main.go\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_f8bd4bb20bae4a51a1c0a858\",\"function\":{\"arguments\":\"{\\\"file_path\\\":\\\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/multiedit_tool/main.go\\\"}\",\"name\":\"view\"},\"type\":\"function\"}],\"role\":\"assistant\",\"reasoning_content\":\"The user wants me to:\\n1. Change 'Hello, World!' to 'Hello, Crush!' \\n2. Add a comment '// Greeting' above the fmt.Println line\\n3. Use the multiedit tool\\n4. Do this in main.go\\n\\nFirst, I need to read the main.go file to see its current state, then use multiedit to make the changes.\"},{\"content\":\"\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\tfmt.Println(\\\"Hello, World!\\\")\\n 7|}\\n\\n\",\"tool_call_id\":\"call_f8bd4bb20bae4a51a1c0a858\",\"role\":\"tool\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Now"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" I"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" can"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" see"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" file"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" I"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" need"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":":\n"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"1"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Change"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" \""}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Hello"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":","}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" World"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"!\""}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" \""}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Hello"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":","}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Crush"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"!\"\n"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"2"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Add"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" \"//"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" G"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"reeting"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\""}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" comment"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" above"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" fmt"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".Println"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" line"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n\n"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"I"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"'ll"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" use"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" mult"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"ied"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"it"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" make"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" both"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" changes"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" in"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" one"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" operation"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" The"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" edits"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" need"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" be"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" done"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" sequentially"}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_1589765f2857482ba0cdd7d3","index":0,"type":"function","function":{"name":"multiedit","arguments":"{\"edits\":[{\"new_string\": \"\\t// Greeting\\n\\tfmt.Println(\\\"Hello, Crush!\\\")\", \"old_string\": \"\\tfmt.Println(\\\"Hello, World!\\\")\"}],\"file_path\":\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/multiedit_tool/main.go\"}"}}]}}]}
-
- data: {"id":"2026041521035260de6e795c104233","created":1776258232,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"tool_calls","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":7318,"completion_tokens":138,"total_tokens":7456,"prompt_tokens_details":{"cached_tokens":7260},"completion_tokens_details":{"reasoning_tokens":61}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 962.306125ms
-- id: 3
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 32313
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/multiedit_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use multiedit to change 'Hello, World!' to 'Hello, Crush!' and add a comment '// Greeting' above the fmt.Println line in main.go\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_f8bd4bb20bae4a51a1c0a858\",\"function\":{\"arguments\":\"{\\\"file_path\\\":\\\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/multiedit_tool/main.go\\\"}\",\"name\":\"view\"},\"type\":\"function\"}],\"role\":\"assistant\",\"reasoning_content\":\"The user wants me to:\\n1. Change 'Hello, World!' to 'Hello, Crush!' \\n2. Add a comment '// Greeting' above the fmt.Println line\\n3. Use the multiedit tool\\n4. Do this in main.go\\n\\nFirst, I need to read the main.go file to see its current state, then use multiedit to make the changes.\"},{\"content\":\"\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\tfmt.Println(\\\"Hello, World!\\\")\\n 7|}\\n\\n\",\"tool_call_id\":\"call_f8bd4bb20bae4a51a1c0a858\",\"role\":\"tool\"},{\"tool_calls\":[{\"id\":\"call_1589765f2857482ba0cdd7d3\",\"function\":{\"arguments\":\"{\\\"edits\\\":[{\\\"new_string\\\": \\\"\\\\t// Greeting\\\\n\\\\tfmt.Println(\\\\\\\"Hello, Crush!\\\\\\\")\\\", \\\"old_string\\\": \\\"\\\\tfmt.Println(\\\\\\\"Hello, World!\\\\\\\")\\\"}],\\\"file_path\\\":\\\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/multiedit_tool/main.go\\\"}\",\"name\":\"multiedit\"},\"type\":\"function\"}],\"role\":\"assistant\",\"reasoning_content\":\"Now I can see the file. I need to:\\n1. Change \\\"Hello, World!\\\" to \\\"Hello, Crush!\\\"\\n2. Add \\\"// Greeting\\\" comment above the fmt.Println line\\n\\nI'll use multiedit to make both changes in one operation. The edits need to be done sequentially.\"},{\"content\":\"\\nApplied 1 edits to file: /tmp/crush-test/TestCoderAgent/zai-glm4.6/multiedit_tool/main.go\\n\\n\",\"tool_call_id\":\"call_1589765f2857482ba0cdd7d3\",\"role\":\"tool\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"20260415210354d634eaa0bf4b4759","created":1776258234,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"The"}}]}
-
- data: {"id":"20260415210354d634eaa0bf4b4759","created":1776258234,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" mult"}}]}
-
- data: {"id":"20260415210354d634eaa0bf4b4759","created":1776258234,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"ied"}}]}
-
- data: {"id":"20260415210354d634eaa0bf4b4759","created":1776258234,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"it"}}]}
-
- data: {"id":"20260415210354d634eaa0bf4b4759","created":1776258234,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" was"}}]}
-
- data: {"id":"20260415210354d634eaa0bf4b4759","created":1776258234,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" successful"}}]}
-
- data: {"id":"20260415210354d634eaa0bf4b4759","created":1776258234,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210354d634eaa0bf4b4759","created":1776258234,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Let"}}]}
-
- data: {"id":"20260415210354d634eaa0bf4b4759","created":1776258234,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" me"}}]}
-
- data: {"id":"20260415210354d634eaa0bf4b4759","created":1776258234,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" verify"}}]}
-
- data: {"id":"20260415210354d634eaa0bf4b4759","created":1776258234,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210354d634eaa0bf4b4759","created":1776258234,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" changes"}}]}
-
- data: {"id":"20260415210354d634eaa0bf4b4759","created":1776258234,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" by"}}]}
-
- data: {"id":"20260415210354d634eaa0bf4b4759","created":1776258234,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" viewing"}}]}
-
- data: {"id":"20260415210354d634eaa0bf4b4759","created":1776258234,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210354d634eaa0bf4b4759","created":1776258234,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" file"}}]}
-
- data: {"id":"20260415210354d634eaa0bf4b4759","created":1776258234,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" again"}}]}
-
- data: {"id":"20260415210354d634eaa0bf4b4759","created":1776258234,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210354d634eaa0bf4b4759","created":1776258234,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_0411cca90ba04c4381f224c2","index":0,"type":"function","function":{"name":"view","arguments":"{\"file_path\":\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/multiedit_tool/main.go\"}"}}]}}]}
-
- data: {"id":"20260415210354d634eaa0bf4b4759","created":1776258234,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"tool_calls","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":7496,"completion_tokens":50,"total_tokens":7546,"prompt_tokens_details":{"cached_tokens":7454},"completion_tokens_details":{"reasoning_tokens":18}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 879.770625ms
-- id: 4
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 32887
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/multiedit_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use multiedit to change 'Hello, World!' to 'Hello, Crush!' and add a comment '// Greeting' above the fmt.Println line in main.go\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_f8bd4bb20bae4a51a1c0a858\",\"function\":{\"arguments\":\"{\\\"file_path\\\":\\\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/multiedit_tool/main.go\\\"}\",\"name\":\"view\"},\"type\":\"function\"}],\"role\":\"assistant\",\"reasoning_content\":\"The user wants me to:\\n1. Change 'Hello, World!' to 'Hello, Crush!' \\n2. Add a comment '// Greeting' above the fmt.Println line\\n3. Use the multiedit tool\\n4. Do this in main.go\\n\\nFirst, I need to read the main.go file to see its current state, then use multiedit to make the changes.\"},{\"content\":\"\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\tfmt.Println(\\\"Hello, World!\\\")\\n 7|}\\n\\n\",\"tool_call_id\":\"call_f8bd4bb20bae4a51a1c0a858\",\"role\":\"tool\"},{\"tool_calls\":[{\"id\":\"call_1589765f2857482ba0cdd7d3\",\"function\":{\"arguments\":\"{\\\"edits\\\":[{\\\"new_string\\\": \\\"\\\\t// Greeting\\\\n\\\\tfmt.Println(\\\\\\\"Hello, Crush!\\\\\\\")\\\", \\\"old_string\\\": \\\"\\\\tfmt.Println(\\\\\\\"Hello, World!\\\\\\\")\\\"}],\\\"file_path\\\":\\\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/multiedit_tool/main.go\\\"}\",\"name\":\"multiedit\"},\"type\":\"function\"}],\"role\":\"assistant\",\"reasoning_content\":\"Now I can see the file. I need to:\\n1. Change \\\"Hello, World!\\\" to \\\"Hello, Crush!\\\"\\n2. Add \\\"// Greeting\\\" comment above the fmt.Println line\\n\\nI'll use multiedit to make both changes in one operation. The edits need to be done sequentially.\"},{\"content\":\"\\nApplied 1 edits to file: /tmp/crush-test/TestCoderAgent/zai-glm4.6/multiedit_tool/main.go\\n\\n\",\"tool_call_id\":\"call_1589765f2857482ba0cdd7d3\",\"role\":\"tool\"},{\"tool_calls\":[{\"id\":\"call_0411cca90ba04c4381f224c2\",\"function\":{\"arguments\":\"{\\\"file_path\\\":\\\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/multiedit_tool/main.go\\\"}\",\"name\":\"view\"},\"type\":\"function\"}],\"role\":\"assistant\",\"reasoning_content\":\"The multiedit was successful. Let me verify the changes by viewing the file again.\"},{\"content\":\"\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\t// Greeting\\n 7|\\tfmt.Println(\\\"Hello, Crush!\\\")\\n 8|}\\n\\n\",\"tool_call_id\":\"call_0411cca90ba04c4381f224c2\",\"role\":\"tool\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Perfect"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"!"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" The"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" changes"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" have"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" been"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" applied"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" successfully"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":":\n"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"1"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" \""}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Hello"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":","}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" World"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"!\""}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" has"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" been"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" changed"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" \""}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Hello"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":","}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Crush"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"!\"\n"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"2"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" \"//"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" G"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"reeting"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\""}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" comment"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" has"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" been"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" added"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" above"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" fmt"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".Println"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" line"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n\n"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Both"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" tasks"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" are"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" complete"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"Done"}}]}
-
- data: {"id":"2026041521035726496bade0354601","created":1776258237,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"stop","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":7611,"completion_tokens":49,"total_tokens":7660,"prompt_tokens_details":{"cached_tokens":7544},"completion_tokens_details":{"reasoning_tokens":46}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 1.125342667s
diff --git a/internal/agent/testdata/TestCoderAgent/zai-glm4.6/parallel_tool_calls.yaml b/internal/agent/testdata/TestCoderAgent/zai-glm4.6/parallel_tool_calls.yaml
deleted file mode 100644
index 4e29b8154560249867787af04f84f2d5103ca7db..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/zai-glm4.6/parallel_tool_calls.yaml
+++ /dev/null
@@ -1,402 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 881
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse glob to find all .go files and use ls to list the current directory, it is very important that you run both tool calls in parallel\n \n\n","role":"user"}],"model":"glm-4.5-air","max_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"2026041521042468d4466b9b45454f","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n"}}]}
-
- data: {"id":"2026041521042468d4466b9b45454f","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":"Find"}}]}
-
- data: {"id":"2026041521042468d4466b9b45454f","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" and"}}]}
-
- data: {"id":"2026041521042468d4466b9b45454f","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" list"}}]}
-
- data: {"id":"2026041521042468d4466b9b45454f","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" Go"}}]}
-
- data: {"id":"2026041521042468d4466b9b45454f","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" files"}}]}
-
- data: {"id":"2026041521042468d4466b9b45454f","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" concurrently"}}]}
-
- data: {"id":"2026041521042468d4466b9b45454f","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"finish_reason":"stop","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":165,"completion_tokens":10,"total_tokens":175,"prompt_tokens_details":{"cached_tokens":131}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 671.357792ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30742
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/parallel_tool_calls\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use glob to find all .go files and use ls to list the current directory, it is very important that you run both tool calls in parallel\",\"role\":\"user\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"The"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" user"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" wants"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" me"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":":\n"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"1"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Use"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" glob"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" find"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" all"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" ."}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"go"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" files"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"2"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Use"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" ls"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" list"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" current"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" directory"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"3"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" They"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" emphasize"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" that"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" it"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"'s"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" very"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" important"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" run"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" both"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" tool"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" calls"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" in"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" parallel"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n\n"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Since"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" these"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" are"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" independent"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" operations"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":","}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" I"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" can"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" run"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" them"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" in"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" parallel"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" by"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" making"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" both"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" tool"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" calls"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" in"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" a"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" single"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" message"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" I"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" don"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"'t"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" need"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" a"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" description"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" parameter"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" for"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" glob"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" or"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" ls"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" as"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" they"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"'re"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" just"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" listing"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" operations"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".\n\n"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Let"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" me"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" make"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" both"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" calls"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" in"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" parallel"}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_eed4dcc84484436dabda6a0d","index":0,"type":"function","function":{"name":"glob","arguments":"{\"pattern\":\"**/*.go\"}"}}]}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_f2fe9a4ba66845568e4cbd34","index":1,"type":"function","function":{"name":"ls","arguments":"{\"path\":\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/parallel_tool_calls\"}"}}]}}]}
-
- data: {"id":"2026041521042463324cc282e74c01","created":1776258264,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"tool_calls","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":7146,"completion_tokens":131,"total_tokens":7277,"prompt_tokens_details":{"cached_tokens":7003},"completion_tokens_details":{"reasoning_tokens":92}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 1.81098325s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31832
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/parallel_tool_calls\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use glob to find all .go files and use ls to list the current directory, it is very important that you run both tool calls in parallel\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_eed4dcc84484436dabda6a0d\",\"function\":{\"arguments\":\"{\\\"pattern\\\":\\\"**/*.go\\\"}\",\"name\":\"glob\"},\"type\":\"function\"},{\"id\":\"call_f2fe9a4ba66845568e4cbd34\",\"function\":{\"arguments\":\"{\\\"path\\\":\\\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/parallel_tool_calls\\\"}\",\"name\":\"ls\"},\"type\":\"function\"}],\"role\":\"assistant\",\"reasoning_content\":\"The user wants me to:\\n1. Use glob to find all .go files\\n2. Use ls to list the current directory\\n3. They emphasize that it's very important to run both tool calls in parallel\\n\\nSince these are independent operations, I can run them in parallel by making both tool calls in a single message. I don't need a description parameter for glob or ls as they're just listing operations.\\n\\nLet me make both calls in parallel.\"},{\"content\":\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/parallel_tool_calls/main.go\",\"tool_call_id\":\"call_eed4dcc84484436dabda6a0d\",\"role\":\"tool\"},{\"content\":\"\\n- /tmp/crush-test/TestCoderAgent/zai-glm4.6/parallel_tool_calls/\\n - go.mod\\n - main.go\\n\",\"tool_call_id\":\"call_f2fe9a4ba66845568e4cbd34\",\"role\":\"tool\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Both"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" tool"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" calls"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" completed"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" successfully"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" in"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" parallel"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" I"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" found"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":":\n"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"-"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" "}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"1"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" ."}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"go"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" file"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":":"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" main"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".go"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"-"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Current"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" directory"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" contains"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":":"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" go"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".mod"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" and"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" main"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".go"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"Found"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" "}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"1"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" ."}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"go"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" file"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":":"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" main"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":".go"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"\n"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"Current"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" directory"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":":"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" go"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":".mod"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":","}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" main"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":".go"}}]}
-
- data: {"id":"2026041521042983107c7ff99b43d7","created":1776258269,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"stop","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":7335,"completion_tokens":51,"total_tokens":7386,"prompt_tokens_details":{"cached_tokens":7275},"completion_tokens_details":{"reasoning_tokens":31}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 794.090041ms
diff --git a/internal/agent/testdata/TestCoderAgent/zai-glm4.6/read_a_file.yaml b/internal/agent/testdata/TestCoderAgent/zai-glm4.6/read_a_file.yaml
deleted file mode 100644
index 7ae474158053a90a8cf7eb5a9e63784c0c44edc2..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/zai-glm4.6/read_a_file.yaml
+++ /dev/null
@@ -1,283 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 762
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nRead the go mod\n \n\n","role":"user"}],"model":"glm-4.5-air","max_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"20260415210242aaced0dd2f274718","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n"}}]}
-
- data: {"id":"20260415210242aaced0dd2f274718","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":"Understanding"}}]}
-
- data: {"id":"20260415210242aaced0dd2f274718","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" Go"}}]}
-
- data: {"id":"20260415210242aaced0dd2f274718","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" Module"}}]}
-
- data: {"id":"20260415210242aaced0dd2f274718","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"finish_reason":"stop","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":140,"completion_tokens":7,"total_tokens":147,"prompt_tokens_details":{"cached_tokens":130}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 770.922042ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30615
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/read_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"Read the go mod\",\"role\":\"user\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"202604152102425b9a6fe9af5f4fcd","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"The"}}]}
-
- data: {"id":"202604152102425b9a6fe9af5f4fcd","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" user"}}]}
-
- data: {"id":"202604152102425b9a6fe9af5f4fcd","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" wants"}}]}
-
- data: {"id":"202604152102425b9a6fe9af5f4fcd","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" me"}}]}
-
- data: {"id":"202604152102425b9a6fe9af5f4fcd","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"202604152102425b9a6fe9af5f4fcd","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" read"}}]}
-
- data: {"id":"202604152102425b9a6fe9af5f4fcd","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"202604152102425b9a6fe9af5f4fcd","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" go"}}]}
-
- data: {"id":"202604152102425b9a6fe9af5f4fcd","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".mod"}}]}
-
- data: {"id":"202604152102425b9a6fe9af5f4fcd","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" file"}}]}
-
- data: {"id":"202604152102425b9a6fe9af5f4fcd","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"202604152102425b9a6fe9af5f4fcd","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" I"}}]}
-
- data: {"id":"202604152102425b9a6fe9af5f4fcd","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" should"}}]}
-
- data: {"id":"202604152102425b9a6fe9af5f4fcd","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" look"}}]}
-
- data: {"id":"202604152102425b9a6fe9af5f4fcd","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" for"}}]}
-
- data: {"id":"202604152102425b9a6fe9af5f4fcd","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" it"}}]}
-
- data: {"id":"202604152102425b9a6fe9af5f4fcd","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" in"}}]}
-
- data: {"id":"202604152102425b9a6fe9af5f4fcd","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"202604152102425b9a6fe9af5f4fcd","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" current"}}]}
-
- data: {"id":"202604152102425b9a6fe9af5f4fcd","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" directory"}}]}
-
- data: {"id":"202604152102425b9a6fe9af5f4fcd","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" and"}}]}
-
- data: {"id":"202604152102425b9a6fe9af5f4fcd","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" read"}}]}
-
- data: {"id":"202604152102425b9a6fe9af5f4fcd","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" its"}}]}
-
- data: {"id":"202604152102425b9a6fe9af5f4fcd","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" contents"}}]}
-
- data: {"id":"202604152102425b9a6fe9af5f4fcd","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"202604152102425b9a6fe9af5f4fcd","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_30d1149ae2744dd489953d90","index":0,"type":"function","function":{"name":"glob","arguments":"{\"pattern\":\"go.mod\"}"}}]}}]}
-
- data: {"id":"202604152102425b9a6fe9af5f4fcd","created":1776258162,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"tool_calls","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":7120,"completion_tokens":37,"total_tokens":7157,"prompt_tokens_details":{"cached_tokens":7003},"completion_tokens_details":{"reasoning_tokens":25}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 1.112772375s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31044
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/read_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"Read the go mod\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_30d1149ae2744dd489953d90\",\"function\":{\"arguments\":\"{\\\"pattern\\\":\\\"go.mod\\\"}\",\"name\":\"glob\"},\"type\":\"function\"}],\"role\":\"assistant\",\"reasoning_content\":\"The user wants me to read the go.mod file. I should look for it in the current directory and read its contents.\"},{\"content\":\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/read_a_file/go.mod\",\"tool_call_id\":\"call_30d1149ae2744dd489953d90\",\"role\":\"tool\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"20260415210244fe5ae7969f4741fc","created":1776258164,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Found"}}]}
-
- data: {"id":"20260415210244fe5ae7969f4741fc","created":1776258164,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210244fe5ae7969f4741fc","created":1776258164,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" go"}}]}
-
- data: {"id":"20260415210244fe5ae7969f4741fc","created":1776258164,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".mod"}}]}
-
- data: {"id":"20260415210244fe5ae7969f4741fc","created":1776258164,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" file"}}]}
-
- data: {"id":"20260415210244fe5ae7969f4741fc","created":1776258164,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210244fe5ae7969f4741fc","created":1776258164,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Now"}}]}
-
- data: {"id":"20260415210244fe5ae7969f4741fc","created":1776258164,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" I"}}]}
-
- data: {"id":"20260415210244fe5ae7969f4741fc","created":1776258164,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"'ll"}}]}
-
- data: {"id":"20260415210244fe5ae7969f4741fc","created":1776258164,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" read"}}]}
-
- data: {"id":"20260415210244fe5ae7969f4741fc","created":1776258164,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" its"}}]}
-
- data: {"id":"20260415210244fe5ae7969f4741fc","created":1776258164,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" contents"}}]}
-
- data: {"id":"20260415210244fe5ae7969f4741fc","created":1776258164,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210244fe5ae7969f4741fc","created":1776258164,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_08cad177ee164623842d1095","index":0,"type":"function","function":{"name":"view","arguments":"{\"file_path\":\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/read_a_file/go.mod\"}"}}]}}]}
-
- data: {"id":"20260415210244fe5ae7969f4741fc","created":1776258164,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"tool_calls","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":7180,"completion_tokens":43,"total_tokens":7223,"prompt_tokens_details":{"cached_tokens":7155},"completion_tokens_details":{"reasoning_tokens":13}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 1.215563667s
-- id: 3
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31489
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/read_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"Read the go mod\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_30d1149ae2744dd489953d90\",\"function\":{\"arguments\":\"{\\\"pattern\\\":\\\"go.mod\\\"}\",\"name\":\"glob\"},\"type\":\"function\"}],\"role\":\"assistant\",\"reasoning_content\":\"The user wants me to read the go.mod file. I should look for it in the current directory and read its contents.\"},{\"content\":\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/read_a_file/go.mod\",\"tool_call_id\":\"call_30d1149ae2744dd489953d90\",\"role\":\"tool\"},{\"tool_calls\":[{\"id\":\"call_08cad177ee164623842d1095\",\"function\":{\"arguments\":\"{\\\"file_path\\\":\\\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/read_a_file/go.mod\\\"}\",\"name\":\"view\"},\"type\":\"function\"}],\"role\":\"assistant\",\"reasoning_content\":\"Found the go.mod file. Now I'll read its contents.\"},{\"content\":\"\\n 1|module example.com/testproject\\n 2|\\n 3|go 1.23\\n\\n\",\"tool_call_id\":\"call_08cad177ee164623842d1095\",\"role\":\"tool\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"The"}}]}
-
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" go"}}]}
-
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".mod"}}]}
-
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" file"}}]}
-
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" is"}}]}
-
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" simple"}}]}
-
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" with"}}]}
-
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" just"}}]}
-
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" module"}}]}
-
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" name"}}]}
-
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" and"}}]}
-
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Go"}}]}
-
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" version"}}]}
-
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"``"}}]}
-
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"`\n"}}]}
-
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"module"}}]}
-
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" example"}}]}
-
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":".com"}}]}
-
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"/test"}}]}
-
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"project"}}]}
-
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"\n\n"}}]}
-
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"go"}}]}
-
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" "}}]}
-
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"1"}}]}
-
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"."}}]}
-
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"23"}}]}
-
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"\n"}}]}
-
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"```"}}]}
-
- data: {"id":"202604152102464a151d45a9c2456a","created":1776258166,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"stop","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":7257,"completion_tokens":32,"total_tokens":7289,"prompt_tokens_details":{"cached_tokens":7221},"completion_tokens_details":{"reasoning_tokens":15}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 1.11567975s
diff --git a/internal/agent/testdata/TestCoderAgent/zai-glm4.6/simple_test.yaml b/internal/agent/testdata/TestCoderAgent/zai-glm4.6/simple_test.yaml
deleted file mode 100644
index 89e0b54a5285a8d7df0c7143330f821ea8fe8a1e..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/zai-glm4.6/simple_test.yaml
+++ /dev/null
@@ -1,227 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 752
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nHello\n \n\n","role":"user"}],"model":"glm-4.5-air","max_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"202604152102399ca97490d8064fea","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n"}}]}
-
- data: {"id":"202604152102399ca97490d8064fea","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":"Hello"}}]}
-
- data: {"id":"202604152102399ca97490d8064fea","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"finish_reason":"stop","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":137,"completion_tokens":5,"total_tokens":142,"prompt_tokens_details":{"cached_tokens":130}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 1.940147958s
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30605
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/simple_test\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"Hello\",\"role\":\"user\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"The"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" user"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" just"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" said"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" \""}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Hello"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\"."}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" This"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" is"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" a"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" simple"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" greeting"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" According"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" communication"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" style"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" guidelines"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":","}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" I"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" should"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" keep"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" responses"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" minimal"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" and"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" use"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" one"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"-word"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" answers"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" when"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" possible"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" I"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" should"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" not"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" use"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" pre"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"amb"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"les"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" like"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" \""}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Hello"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"!\""}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" or"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" \""}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Hi"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" there"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"!\""}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" -"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" just"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" a"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" direct"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" acknowledgment"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".\n\n"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"I"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"'ll"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" respond"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" with"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" a"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" simple"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" \""}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Hello"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\""}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" acknowledge"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" them"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"Hello"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"!"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" How"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" can"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" I"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" help"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" you"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" today"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"?"}}]}
-
- data: {"id":"202604152102395f3e6a2e5516487a","created":1776258159,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"stop","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":7116,"completion_tokens":79,"total_tokens":7195,"prompt_tokens_details":{"cached_tokens":43},"completion_tokens_details":{"reasoning_tokens":68}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 2.719987708s
diff --git a/internal/agent/testdata/TestCoderAgent/zai-glm4.6/sourcegraph_tool.yaml b/internal/agent/testdata/TestCoderAgent/zai-glm4.6/sourcegraph_tool.yaml
deleted file mode 100644
index d0c439268d60b058463a538f79b4f26c94d50d03..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/zai-glm4.6/sourcegraph_tool.yaml
+++ /dev/null
@@ -1,730 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30665
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/sourcegraph_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use sourcegraph to search for 'func main' in Go repositories\",\"role\":\"user\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"The"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" user"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" wants"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" me"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" use"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Source"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"graph"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" tool"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" search"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" for"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" '"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"func"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" main"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"'"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" in"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Go"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" repositories"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" I"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"'ll"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" use"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" source"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"graph"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" tool"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" with"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" query"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" '"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"func"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" main"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"'"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" and"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" specify"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Go"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" repositories"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" as"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" a"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" filter"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".\n\n"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Let"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" me"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" construct"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" search"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" query"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":":\n"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"-"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Query"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":":"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" `"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"func"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" main"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"`\n"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"-"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Language"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" filter"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":":"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" `"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"lang"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":":"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Go"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"`\n"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"-"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" This"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" will"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" search"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" for"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" main"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" function"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" in"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Go"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" repositories"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n\n"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"I"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"'ll"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" use"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" source"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"graph"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" tool"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" perform"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" this"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" search"}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_ad8b8092390d4cd9a01a865c","index":0,"type":"function","function":{"name":"sourcegraph","arguments":"{\"count\":10,\"query\":\"func main lang:Go\"}"}}]}}]}
-
- data: {"id":"202604152103593a9fa27ae4314d66","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"tool_calls","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":7129,"completion_tokens":112,"total_tokens":7241,"prompt_tokens_details":{"cached_tokens":7003},"completion_tokens_details":{"reasoning_tokens":90}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 1.016685708s
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 424
- host: sourcegraph.com
- body: '{"query":"query Search($query: String!) { search(query: $query, version: V2, patternType: keyword ) { results { matchCount, limitHit, resultCount, approximateResultCount, missing { name }, timedout { name }, indexUnavailable, results { __typename, ... on FileMatch { repository { name }, file { path, url, content }, lineMatches { preview, lineNumber, offsetAndLengths } } } } } }","variables":{"query":"func main lang:Go"}}'
- headers:
- Content-Type:
- - application/json
- User-Agent:
- - crush/1.0
- url: https://sourcegraph.com/.api/graphql
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- uncompressed: true
- body: '{"data":{"search":{"results":{"matchCount":30,"limitHit":true,"resultCount":30,"approximateResultCount":"30+","missing":[],"timedout":[],"indexUnavailable":false,"results":[{"__typename":"FileMatch","repository":{"name":"github.com/GoesToEleven/GolangTraining"},"file":{"path":"14_functions/10_func-expression/01_before-func-expression/main.go","url":"/r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/01_before-func-expression/main.go","content":"package main\n\nimport \"fmt\"\n\nfunc greeting() {\n\tfmt.Println(\"Hello world!\")\n}\n\nfunc main() {\n\tgreeting()\n}\n"},"lineMatches":[{"preview":"func main() {","lineNumber":8,"offsetAndLengths":[[0,9]]},{"preview":"package main","lineNumber":0,"offsetAndLengths":[[8,4]]},{"preview":"func greeting() {","lineNumber":4,"offsetAndLengths":[[0,4]]}]},{"__typename":"FileMatch","repository":{"name":"github.com/GoesToEleven/GolangTraining"},"file":{"path":"14_functions/10_func-expression/02_func-expression/main.go","url":"/r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/02_func-expression/main.go","content":"package main\n\nimport \"fmt\"\n\nfunc main() {\n\n\tgreeting := func() {\n\t\tfmt.Println(\"Hello world!\")\n\t}\n\n\tgreeting()\n}\n"},"lineMatches":[{"preview":"func main() {","lineNumber":4,"offsetAndLengths":[[0,9]]},{"preview":"\tgreeting := func() {","lineNumber":6,"offsetAndLengths":[[13,4]]},{"preview":"package main","lineNumber":0,"offsetAndLengths":[[8,4]]}]},{"__typename":"FileMatch","repository":{"name":"github.com/GoesToEleven/GolangTraining"},"file":{"path":"14_functions/10_func-expression/03_func-expression_shows-type/main.go","url":"/r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/03_func-expression_shows-type/main.go","content":"package main\n\nimport \"fmt\"\n\nfunc main() {\n\n\tgreeting := func() {\n\t\tfmt.Println(\"Hello world!\")\n\t}\n\n\tgreeting()\n\tfmt.Printf(\"%T\\n\", greeting)\n}\n"},"lineMatches":[{"preview":"func main() {","lineNumber":4,"offsetAndLengths":[[0,9]]},{"preview":"\tgreeting := func() {","lineNumber":6,"offsetAndLengths":[[13,4]]},{"preview":"package main","lineNumber":0,"offsetAndLengths":[[8,4]]}]},{"__typename":"FileMatch","repository":{"name":"github.com/GoesToEleven/GolangTraining"},"file":{"path":"14_functions/10_func-expression/04_another-way_func-expression/main.go","url":"/r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/04_another-way_func-expression/main.go","content":"package main\n\nimport \"fmt\"\n\nfunc makeGreeter() func() string {\n\treturn func() string {\n\t\treturn \"Hello world!\"\n\t}\n}\n\nfunc main() {\n\tgreet := makeGreeter()\n\tfmt.Println(greet())\n}\n"},"lineMatches":[{"preview":"func main() {","lineNumber":10,"offsetAndLengths":[[0,9]]},{"preview":"package main","lineNumber":0,"offsetAndLengths":[[8,4]]},{"preview":"func makeGreeter() func() string {","lineNumber":4,"offsetAndLengths":[[0,4],[19,4]]},{"preview":"\treturn func() string {","lineNumber":5,"offsetAndLengths":[[8,4]]}]},{"__typename":"FileMatch","repository":{"name":"github.com/jmbharathram/executeoncommand"},"file":{"path":"go/channel/main.go","url":"/r/github.com/jmbharathram/executeoncommand/-/blob/go/channel/main.go","content":"package main\n\nimport \"fmt\"\n\n// Example 0\nfunc dummy_func() {\n\tfmt.Println(\"dummy_func\")\n}\nfunc main() {\n\tgo dummy_func()\n fmt.Println(\"end of main\")\n}\n\n\n// Example 1\n\npackage main\n\nimport \"fmt\"\n\nfunc dummy_func(ch chan int) {\n // receive a value from channel\n val := \u003c-ch\n\tfmt.Println(val)\n}\nfunc main() {\n // create a channel\n\tch := make(chan int)\n // trigger a parallel go routine\n\tgo dummy_func(ch)\n // send a value into the channel\n\tch \u003c- 10\n}\n\n// Example 2\n\npackage main\n\nimport (\n \"fmt\"\n //\"flag\"\n)\n\nfunc dummy_func(ch chan int) {\n // receive a value from channel\n val := \u003c-ch\n\tfmt.Println(val)\n\n //send a value again\n ch \u003c- val*2\n}\nfunc main() {\n fmt.Println(\"Welcome to main function\")\n \n //flag statements for user input\n var input_value int\n input_value = 10\n //flag.IntVar(\u0026input_value, \"v\", 10, \"Input value to send/receive via channel.\" )\n //flag.Parse()\n\n // create a channel\n\tch := make(chan int)\n\n // trigger a parallel go routine\n\tgo dummy_func(ch)\n \n // send a value into the channel\n\tch \u003c- input_value\n\n // receive a value from channel\n response := \u003c- ch\n fmt.Println(response)\n\n}\n\n// example 3\n\n\npackage main\n\nimport(\n\"fmt\"\n\"time\"\n\"math/rand\"\n)\n\nfunc routine1(ch chan int, min int, max int) {\n\n rand.Seed(time.Now().UnixNano())\n t := rand.Intn(max - min) + min\n fmt.Println(\"routine 1 is waiting: \", t, \"secs\")\n time.Sleep(time.Duration(t)*time.Second)\n\n ch \u003c- t\n}\n\nfunc routine2(ch chan int, min int, max int) {\n\n rand.Seed(time.Now().UnixNano())\n t := rand.Intn(max - min) + min\n fmt.Println(\"routine 2 is waiting: \", t, \"secs\")\n time.Sleep(time.Duration(t)*time.Second)\n\n ch \u003c- t\n}\n\nfunc main(){\n\nmin := 1\nmax := 10\n\nch1:= make(chan int)\nch2:= make(chan int)\n\ngo routine1(ch1, min, max)\ngo routine2(ch2, min, max)\n\nselect{\n\ncase \u003c- ch1:\n fmt.Println(\"routine 1 finished\")\n\ncase \u003c- ch2:\n fmt.Println(\"routine 2 finished\")\n}\n\nfmt.Println(\"end of main function\")\n\n}\n"},"lineMatches":[{"preview":"func main(){","lineNumber":106,"offsetAndLengths":[[0,9]]},{"preview":"func dummy_func() {","lineNumber":5,"offsetAndLengths":[[0,4],[11,4]]},{"preview":"\tfmt.Println(\"dummy_func\")","lineNumber":6,"offsetAndLengths":[[20,4]]},{"preview":"func main() {","lineNumber":8,"offsetAndLengths":[[0,9]]},{"preview":"\tgo dummy_func()","lineNumber":9,"offsetAndLengths":[[10,4]]},{"preview":" fmt.Println(\"end of main\")","lineNumber":10,"offsetAndLengths":[[24,4]]},{"preview":"func main() {","lineNumber":25,"offsetAndLengths":[[0,9]]},{"preview":"func main() {","lineNumber":51,"offsetAndLengths":[[0,9]]},{"preview":" fmt.Println(\"Welcome to main function\")","lineNumber":52,"offsetAndLengths":[[28,4],[33,4]]},{"preview":"fmt.Println(\"end of main function\")","lineNumber":126,"offsetAndLengths":[[20,4],[25,4]]},{"preview":"package main","lineNumber":78,"offsetAndLengths":[[8,4]]},{"preview":"package main","lineNumber":16,"offsetAndLengths":[[8,4]]},{"preview":"package main","lineNumber":0,"offsetAndLengths":[[8,4]]}]}]}}}}'
- headers:
- Content-Type:
- - application/json
- status: 200 OK
- code: 200
- duration: 1.307873833s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 807
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse sourcegraph to search for ''func main'' in Go repositories\n \n\n","role":"user"}],"model":"glm-4.5-air","max_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"2026041521035902acbcfa6bd94d50","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n"}}]}
-
- data: {"id":"2026041521035902acbcfa6bd94d50","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":"Searching"}}]}
-
- data: {"id":"2026041521035902acbcfa6bd94d50","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" func"}}]}
-
- data: {"id":"2026041521035902acbcfa6bd94d50","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" main"}}]}
-
- data: {"id":"2026041521035902acbcfa6bd94d50","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" in"}}]}
-
- data: {"id":"2026041521035902acbcfa6bd94d50","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" Go"}}]}
-
- data: {"id":"2026041521035902acbcfa6bd94d50","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" repos"}}]}
-
- data: {"id":"2026041521035902acbcfa6bd94d50","created":1776258239,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"finish_reason":"stop","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":149,"completion_tokens":10,"total_tokens":159,"prompt_tokens_details":{"cached_tokens":131}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 8.405971209s
-- id: 3
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 40395
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/sourcegraph_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use sourcegraph to search for 'func main' in Go repositories\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_ad8b8092390d4cd9a01a865c\",\"function\":{\"arguments\":\"{\\\"count\\\":10,\\\"query\\\":\\\"func main lang:Go\\\"}\",\"name\":\"sourcegraph\"},\"type\":\"function\"}],\"role\":\"assistant\",\"reasoning_content\":\"The user wants me to use the Sourcegraph tool to search for 'func main' in Go repositories. I'll use the sourcegraph tool with the query 'func main' and specify Go repositories as a filter.\\n\\nLet me construct the search query:\\n- Query: `func main`\\n- Language filter: `lang:Go`\\n- This will search for the main function in Go repositories\\n\\nI'll use the sourcegraph tool to perform this search.\"},{\"content\":\"# Sourcegraph Search Results\\n\\nFound 30 matches across 30 results\\n(Result limit reached, try a more specific query)\\n\\n## Result 1: github.com/GoesToEleven/GolangTraining/14_functions/10_func-expression/01_before-func-expression/main.go\\n\\nURL: /r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/01_before-func-expression/main.go\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func greeting() {\\n6| \\tfmt.Println(\\\"Hello world!\\\")\\n7| }\\n8| func main() {\\n9| func main() {\\n10| \\tgreeting()\\n11| }\\n12| \\n```\\n\\n```\\n0| package main\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func greeting() {\\n6| \\tfmt.Println(\\\"Hello world!\\\")\\n7| }\\n8| \\n9| func main() {\\n10| \\tgreeting()\\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| func greeting() {\\n5| func greeting() {\\n6| \\tfmt.Println(\\\"Hello world!\\\")\\n7| }\\n8| \\n9| func main() {\\n10| \\tgreeting()\\n11| }\\n12| \\n```\\n\\n## Result 2: github.com/GoesToEleven/GolangTraining/14_functions/10_func-expression/02_func-expression/main.go\\n\\nURL: /r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/02_func-expression/main.go\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| func main() {\\n5| func main() {\\n6| \\n7| \\tgreeting := func() {\\n8| \\t\\tfmt.Println(\\\"Hello world!\\\")\\n9| \\t}\\n10| \\n11| \\tgreeting()\\n12| }\\n13| \\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func main() {\\n6| \\tgreeting := func() {\\n7| \\tgreeting := func() {\\n8| \\t\\tfmt.Println(\\\"Hello world!\\\")\\n9| \\t}\\n10| \\n11| \\tgreeting()\\n12| }\\n13| \\n```\\n\\n```\\n0| package main\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func main() {\\n6| \\n7| \\tgreeting := func() {\\n8| \\t\\tfmt.Println(\\\"Hello world!\\\")\\n9| \\t}\\n10| \\n```\\n\\n## Result 3: github.com/GoesToEleven/GolangTraining/14_functions/10_func-expression/03_func-expression_shows-type/main.go\\n\\nURL: /r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/03_func-expression_shows-type/main.go\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| func main() {\\n5| func main() {\\n6| \\n7| \\tgreeting := func() {\\n8| \\t\\tfmt.Println(\\\"Hello world!\\\")\\n9| \\t}\\n10| \\n11| \\tgreeting()\\n12| \\tfmt.Printf(\\\"%T\\\\n\\\", greeting)\\n13| }\\n14| \\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func main() {\\n6| \\tgreeting := func() {\\n7| \\tgreeting := func() {\\n8| \\t\\tfmt.Println(\\\"Hello world!\\\")\\n9| \\t}\\n10| \\n11| \\tgreeting()\\n12| \\tfmt.Printf(\\\"%T\\\\n\\\", greeting)\\n13| }\\n14| \\n```\\n\\n```\\n0| package main\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func main() {\\n6| \\n7| \\tgreeting := func() {\\n8| \\t\\tfmt.Println(\\\"Hello world!\\\")\\n9| \\t}\\n10| \\n```\\n\\n## Result 4: github.com/GoesToEleven/GolangTraining/14_functions/10_func-expression/04_another-way_func-expression/main.go\\n\\nURL: /r/github.com/GoesToEleven/GolangTraining/-/blob/14_functions/10_func-expression/04_another-way_func-expression/main.go\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func makeGreeter() func() string {\\n6| \\treturn func() string {\\n7| \\t\\treturn \\\"Hello world!\\\"\\n8| \\t}\\n9| }\\n10| func main() {\\n11| func main() {\\n12| \\tgreet := makeGreeter()\\n13| \\tfmt.Println(greet())\\n14| }\\n15| \\n```\\n\\n```\\n0| package main\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func makeGreeter() func() string {\\n6| \\treturn func() string {\\n7| \\t\\treturn \\\"Hello world!\\\"\\n8| \\t}\\n9| }\\n10| \\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| func makeGreeter() func() string {\\n5| func makeGreeter() func() string {\\n6| \\treturn func() string {\\n7| \\t\\treturn \\\"Hello world!\\\"\\n8| \\t}\\n9| }\\n10| \\n11| func main() {\\n12| \\tgreet := makeGreeter()\\n13| \\tfmt.Println(greet())\\n14| }\\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| \\treturn func() string {\\n6| \\treturn func() string {\\n7| \\t\\treturn \\\"Hello world!\\\"\\n8| \\t}\\n9| }\\n10| \\n11| func main() {\\n12| \\tgreet := makeGreeter()\\n13| \\tfmt.Println(greet())\\n14| }\\n15| \\n```\\n\\n## Result 5: github.com/jmbharathram/executeoncommand/go/channel/main.go\\n\\nURL: /r/github.com/jmbharathram/executeoncommand/-/blob/go/channel/main.go\\n\\n```\\n96| \\n97| func routine2(ch chan int, min int, max int) {\\n98| \\n99| rand.Seed(time.Now().UnixNano())\\n100| t := rand.Intn(max - min) + min\\n101| fmt.Println(\\\"routine 2 is waiting: \\\", t, \\\"secs\\\")\\n102| time.Sleep(time.Duration(t)*time.Second)\\n103| \\n104| ch <- t\\n105| }\\n106| func main(){\\n107| func main(){\\n108| \\n109| min := 1\\n110| max := 10\\n111| \\n112| ch1:= make(chan int)\\n113| ch2:= make(chan int)\\n114| \\n115| go routine1(ch1, min, max)\\n116| go routine2(ch2, min, max)\\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| func dummy_func() {\\n6| func dummy_func() {\\n7| \\tfmt.Println(\\\"dummy_func\\\")\\n8| }\\n9| func main() {\\n10| \\tgo dummy_func()\\n11| fmt.Println(\\\"end of main\\\")\\n12| }\\n13| \\n14| \\n15| // Example 1\\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| // Example 0\\n6| \\tfmt.Println(\\\"dummy_func\\\")\\n7| \\tfmt.Println(\\\"dummy_func\\\")\\n8| }\\n9| func main() {\\n10| \\tgo dummy_func()\\n11| fmt.Println(\\\"end of main\\\")\\n12| }\\n13| \\n14| \\n15| // Example 1\\n16| \\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| // Example 0\\n6| func dummy_func() {\\n7| \\tfmt.Println(\\\"dummy_func\\\")\\n8| func main() {\\n9| func main() {\\n10| \\tgo dummy_func()\\n11| fmt.Println(\\\"end of main\\\")\\n12| }\\n13| \\n14| \\n15| // Example 1\\n16| \\n17| package main\\n18| \\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| // Example 0\\n6| func dummy_func() {\\n7| \\tfmt.Println(\\\"dummy_func\\\")\\n8| }\\n9| \\tgo dummy_func()\\n10| \\tgo dummy_func()\\n11| fmt.Println(\\\"end of main\\\")\\n12| }\\n13| \\n14| \\n15| // Example 1\\n16| \\n17| package main\\n18| \\n19| import \\\"fmt\\\"\\n```\\n\\n```\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| // Example 0\\n6| func dummy_func() {\\n7| \\tfmt.Println(\\\"dummy_func\\\")\\n8| }\\n9| func main() {\\n10| fmt.Println(\\\"end of main\\\")\\n11| fmt.Println(\\\"end of main\\\")\\n12| }\\n13| \\n14| \\n15| // Example 1\\n16| \\n17| package main\\n18| \\n19| import \\\"fmt\\\"\\n20| \\n```\\n\\n```\\n15| // Example 1\\n16| \\n17| package main\\n18| \\n19| import \\\"fmt\\\"\\n20| \\n21| func dummy_func(ch chan int) {\\n22| // receive a value from channel\\n23| val := <-ch\\n24| \\tfmt.Println(val)\\n25| func main() {\\n26| func main() {\\n27| // create a channel\\n28| \\tch := make(chan int)\\n29| // trigger a parallel go routine\\n30| \\tgo dummy_func(ch)\\n31| // send a value into the channel\\n32| \\tch <- 10\\n33| }\\n34| \\n35| // Example 2\\n```\\n\\n```\\n41| //\\\"flag\\\"\\n42| )\\n43| \\n44| func dummy_func(ch chan int) {\\n45| // receive a value from channel\\n46| val := <-ch\\n47| \\tfmt.Println(val)\\n48| \\n49| //send a value again\\n50| ch <- val*2\\n51| func main() {\\n52| func main() {\\n53| fmt.Println(\\\"Welcome to main function\\\")\\n54| \\n55| //flag statements for user input\\n56| var input_value int\\n57| input_value = 10\\n58| //flag.IntVar(&input_value, \\\"v\\\", 10, \\\"Input value to send/receive via channel.\\\" )\\n59| //flag.Parse()\\n60| \\n61| // create a channel\\n```\\n\\n```\\n42| )\\n43| \\n44| func dummy_func(ch chan int) {\\n45| // receive a value from channel\\n46| val := <-ch\\n47| \\tfmt.Println(val)\\n48| \\n49| //send a value again\\n50| ch <- val*2\\n51| }\\n52| fmt.Println(\\\"Welcome to main function\\\")\\n53| fmt.Println(\\\"Welcome to main function\\\")\\n54| \\n55| //flag statements for user input\\n56| var input_value int\\n57| input_value = 10\\n58| //flag.IntVar(&input_value, \\\"v\\\", 10, \\\"Input value to send/receive via channel.\\\" )\\n59| //flag.Parse()\\n60| \\n61| // create a channel\\n62| \\tch := make(chan int)\\n```\\n\\n```\\n116| go routine2(ch2, min, max)\\n117| \\n118| select{\\n119| \\n120| case <- ch1:\\n121| fmt.Println(\\\"routine 1 finished\\\")\\n122| \\n123| case <- ch2:\\n124| fmt.Println(\\\"routine 2 finished\\\")\\n125| }\\n126| fmt.Println(\\\"end of main function\\\")\\n127| fmt.Println(\\\"end of main function\\\")\\n128| \\n129| }\\n130| \\n```\\n\\n```\\n68| \\tch <- input_value\\n69| \\n70| // receive a value from channel\\n71| response := <- ch\\n72| fmt.Println(response)\\n73| \\n74| }\\n75| \\n76| // example 3\\n77| \\n78| package main\\n79| package main\\n80| \\n81| import(\\n82| \\\"fmt\\\"\\n83| \\\"time\\\"\\n84| \\\"math/rand\\\"\\n85| )\\n86| \\n87| func routine1(ch chan int, min int, max int) {\\n88| \\n```\\n\\n```\\n6| func dummy_func() {\\n7| \\tfmt.Println(\\\"dummy_func\\\")\\n8| }\\n9| func main() {\\n10| \\tgo dummy_func()\\n11| fmt.Println(\\\"end of main\\\")\\n12| }\\n13| \\n14| \\n15| // Example 1\\n16| package main\\n17| package main\\n18| \\n19| import \\\"fmt\\\"\\n20| \\n21| func dummy_func(ch chan int) {\\n22| // receive a value from channel\\n23| val := <-ch\\n24| \\tfmt.Println(val)\\n25| }\\n26| func main() {\\n```\\n\\n```\\n0| package main\\n1| package main\\n2| \\n3| import \\\"fmt\\\"\\n4| \\n5| // Example 0\\n6| func dummy_func() {\\n7| \\tfmt.Println(\\\"dummy_func\\\")\\n8| }\\n9| func main() {\\n10| \\tgo dummy_func()\\n```\\n\\n\",\"tool_call_id\":\"call_ad8b8092390d4cd9a01a865c\",\"role\":\"tool\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"The"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" search"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" returned"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" "}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"30"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" matches"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" for"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" '"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"func"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" main"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"'"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" in"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Go"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" repositories"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" The"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" results"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" show"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" various"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" examples"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" from"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" different"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" repositories"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" including"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":":\n\n"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"1"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Several"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" examples"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" from"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" G"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"olang"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Training"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" repository"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" showing"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" different"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" patterns"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" of"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" `"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"func"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" main"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"()`"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" usage"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"2"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Examples"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" from"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" execute"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"on"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"command"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" repository"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" demonstrating"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Go"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" channels"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" and"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" gor"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"outines"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" with"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" main"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" functions"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n\n"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"The"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" results"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" include"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" code"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" snippets"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" showing"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":":\n"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"-"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Basic"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" main"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" functions"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" that"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" call"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" other"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" functions"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"-"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Main"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" functions"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" using"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" function"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" expressions"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"-"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Main"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" functions"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" working"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" with"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" channels"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" and"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" gor"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"outines"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"-"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Various"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Go"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" programming"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" patterns"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n\n"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"The"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" search"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" was"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" successful"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" and"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" found"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" relevant"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" matches"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" I"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" should"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" present"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" a"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" concise"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" summary"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" user"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"Found"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" "}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"30"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" matches"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" for"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" '"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"func"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" main"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"'"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" in"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" Go"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" repositories"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":":\n\n"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"**"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"Key"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" results"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":":"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"**\n"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"-"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" **"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"G"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"olang"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"Training"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"**:"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" Multiple"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" examples"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" showing"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" basic"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" main"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" functions"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":","}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" function"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" expressions"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":","}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" and"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" closure"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" patterns"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"\n"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"-"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" **"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"execute"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"on"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"command"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"**:"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" Examples"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" demonstrating"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" channels"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" and"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" gor"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"outines"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" with"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" select"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" statements"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"\n\n"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"All"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" results"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" show"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" standard"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" Go"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" `"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"main"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"()`"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" function"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" patterns"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" including"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" simple"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" function"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" calls"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":","}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" anonymous"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" functions"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":","}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" channel"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" operations"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":","}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" and"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" concurrent"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" programming"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" examples"}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"."}}]}
-
- data: {"id":"20260415210404613b5b7eaf7c4943","created":1776258244,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"stop","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":10104,"completion_tokens":204,"total_tokens":10308,"prompt_tokens_details":{"cached_tokens":7239},"completion_tokens_details":{"reasoning_tokens":122}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 6.670605167s
diff --git a/internal/agent/testdata/TestCoderAgent/zai-glm4.6/update_a_file.yaml b/internal/agent/testdata/TestCoderAgent/zai-glm4.6/update_a_file.yaml
deleted file mode 100644
index 854458776b93c1f41d67dbc35a29ea538085c502..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/zai-glm4.6/update_a_file.yaml
+++ /dev/null
@@ -1,556 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 816
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nupdate the main.go file by changing the print to say hello from crush\n \n\n","role":"user"}],"model":"glm-4.5-air","max_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"202604152102480d420eea3b86429c","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n"}}]}
-
- data: {"id":"202604152102480d420eea3b86429c","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":"Update"}}]}
-
- data: {"id":"202604152102480d420eea3b86429c","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" main"}}]}
-
- data: {"id":"202604152102480d420eea3b86429c","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":".go"}}]}
-
- data: {"id":"202604152102480d420eea3b86429c","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" to"}}]}
-
- data: {"id":"202604152102480d420eea3b86429c","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" print"}}]}
-
- data: {"id":"202604152102480d420eea3b86429c","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" \""}}]}
-
- data: {"id":"202604152102480d420eea3b86429c","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":"hello"}}]}
-
- data: {"id":"202604152102480d420eea3b86429c","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" from"}}]}
-
- data: {"id":"202604152102480d420eea3b86429c","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" crush"}}]}
-
- data: {"id":"202604152102480d420eea3b86429c","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":"\""}}]}
-
- data: {"id":"202604152102480d420eea3b86429c","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"finish_reason":"stop","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":150,"completion_tokens":14,"total_tokens":164,"prompt_tokens_details":{"cached_tokens":130}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 751.116625ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30671
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/update_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"update the main.go file by changing the print to say hello from crush\",\"role\":\"user\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"The"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" user"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" wants"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" me"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" update"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" main"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".go"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" file"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" change"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" a"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" print"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" statement"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" say"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" \""}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"hello"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" from"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" crush"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\"."}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" \n\n"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"First"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":","}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" I"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" should"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":":\n"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"1"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Check"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" if"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" main"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".go"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" exists"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"2"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Read"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" its"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" contents"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" see"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" what"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" needs"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" be"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" changed"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"3"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Make"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" edit"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"4"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Verify"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" change"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n\n"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Let"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" me"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" start"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" by"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" looking"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" at"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" directory"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" structure"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" and"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" finding"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" main"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".go"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" file"}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_2aed5b1b5514488388bf2533","index":0,"type":"function","function":{"name":"ls","arguments":"{\"path\":\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/update_a_file\"}"}}]}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_faaf6945c0104e19b89a9663","index":1,"type":"function","function":{"name":"view","arguments":"{\"file_path\":\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/update_a_file/main.go\"}"}}]}}]}
-
- data: {"id":"20260415210248bf874b07adff4d5e","created":1776258168,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"tool_calls","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":7130,"completion_tokens":132,"total_tokens":7262,"prompt_tokens_details":{"cached_tokens":7003},"completion_tokens_details":{"reasoning_tokens":77}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 1.271525333s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31798
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/update_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"update the main.go file by changing the print to say hello from crush\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_2aed5b1b5514488388bf2533\",\"function\":{\"arguments\":\"{\\\"path\\\":\\\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/update_a_file\\\"}\",\"name\":\"ls\"},\"type\":\"function\"},{\"id\":\"call_faaf6945c0104e19b89a9663\",\"function\":{\"arguments\":\"{\\\"file_path\\\":\\\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/update_a_file/main.go\\\"}\",\"name\":\"view\"},\"type\":\"function\"}],\"role\":\"assistant\",\"reasoning_content\":\"The user wants me to update the main.go file to change a print statement to say \\\"hello from crush\\\". \\n\\nFirst, I should:\\n1. Check if main.go exists\\n2. Read its contents to see what needs to be changed\\n3. Make the edit\\n4. Verify the change\\n\\nLet me start by looking at the directory structure and finding the main.go file.\"},{\"content\":\"\\n- /tmp/crush-test/TestCoderAgent/zai-glm4.6/update_a_file/\\n - go.mod\\n - main.go\\n\",\"tool_call_id\":\"call_2aed5b1b5514488388bf2533\",\"role\":\"tool\"},{\"content\":\"\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\tfmt.Println(\\\"Hello, World!\\\")\\n 7|}\\n\\n\",\"tool_call_id\":\"call_faaf6945c0104e19b89a9663\",\"role\":\"tool\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"I"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" can"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" see"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" main"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".go"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" file"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" has"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" a"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" print"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" statement"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" that"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" says"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" \""}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Hello"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":","}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" World"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"!\""}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" on"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" line"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" "}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"6"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" I"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" need"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" change"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" it"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" \""}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"hello"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" from"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" crush"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\"."}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Let"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" me"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" make"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" that"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" edit"}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_af084abf43fa46639c6e4a78","index":0,"type":"function","function":{"name":"edit","arguments":"{\"file_path\":\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/update_a_file/main.go\",\"old_string\":\"\\tfmt.Println(\\\"Hello, World!\\\")\",\"new_string\":\"\\tfmt.Println(\\\"hello from crush\\\")\"}"}}]}}]}
-
- data: {"id":"202604152102528620385c28744349","created":1776258172,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"tool_calls","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":7351,"completion_tokens":96,"total_tokens":7447,"prompt_tokens_details":{"cached_tokens":7260},"completion_tokens_details":{"reasoning_tokens":40}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 1.061009958s
-- id: 3
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 32499
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/update_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"update the main.go file by changing the print to say hello from crush\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_2aed5b1b5514488388bf2533\",\"function\":{\"arguments\":\"{\\\"path\\\":\\\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/update_a_file\\\"}\",\"name\":\"ls\"},\"type\":\"function\"},{\"id\":\"call_faaf6945c0104e19b89a9663\",\"function\":{\"arguments\":\"{\\\"file_path\\\":\\\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/update_a_file/main.go\\\"}\",\"name\":\"view\"},\"type\":\"function\"}],\"role\":\"assistant\",\"reasoning_content\":\"The user wants me to update the main.go file to change a print statement to say \\\"hello from crush\\\". \\n\\nFirst, I should:\\n1. Check if main.go exists\\n2. Read its contents to see what needs to be changed\\n3. Make the edit\\n4. Verify the change\\n\\nLet me start by looking at the directory structure and finding the main.go file.\"},{\"content\":\"\\n- /tmp/crush-test/TestCoderAgent/zai-glm4.6/update_a_file/\\n - go.mod\\n - main.go\\n\",\"tool_call_id\":\"call_2aed5b1b5514488388bf2533\",\"role\":\"tool\"},{\"content\":\"\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\tfmt.Println(\\\"Hello, World!\\\")\\n 7|}\\n\\n\",\"tool_call_id\":\"call_faaf6945c0104e19b89a9663\",\"role\":\"tool\"},{\"tool_calls\":[{\"id\":\"call_af084abf43fa46639c6e4a78\",\"function\":{\"arguments\":\"{\\\"file_path\\\":\\\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/update_a_file/main.go\\\",\\\"old_string\\\":\\\"\\\\tfmt.Println(\\\\\\\"Hello, World!\\\\\\\")\\\",\\\"new_string\\\":\\\"\\\\tfmt.Println(\\\\\\\"hello from crush\\\\\\\")\\\"}\",\"name\":\"edit\"},\"type\":\"function\"}],\"role\":\"assistant\",\"reasoning_content\":\"I can see the main.go file has a print statement that says \\\"Hello, World!\\\" on line 6. I need to change it to \\\"hello from crush\\\". Let me make that edit.\"},{\"content\":\"\\nContent replaced in file: /tmp/crush-test/TestCoderAgent/zai-glm4.6/update_a_file/main.go\\n\\n\",\"tool_call_id\":\"call_af084abf43fa46639c6e4a78\",\"role\":\"tool\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"2026041521025448e32c0fe41e492c","created":1776258174,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"The"}}]}
-
- data: {"id":"2026041521025448e32c0fe41e492c","created":1776258174,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" edit"}}]}
-
- data: {"id":"2026041521025448e32c0fe41e492c","created":1776258174,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" was"}}]}
-
- data: {"id":"2026041521025448e32c0fe41e492c","created":1776258174,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" successful"}}]}
-
- data: {"id":"2026041521025448e32c0fe41e492c","created":1776258174,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521025448e32c0fe41e492c","created":1776258174,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" Let"}}]}
-
- data: {"id":"2026041521025448e32c0fe41e492c","created":1776258174,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" me"}}]}
-
- data: {"id":"2026041521025448e32c0fe41e492c","created":1776258174,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" verify"}}]}
-
- data: {"id":"2026041521025448e32c0fe41e492c","created":1776258174,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"2026041521025448e32c0fe41e492c","created":1776258174,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" change"}}]}
-
- data: {"id":"2026041521025448e32c0fe41e492c","created":1776258174,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" by"}}]}
-
- data: {"id":"2026041521025448e32c0fe41e492c","created":1776258174,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" reading"}}]}
-
- data: {"id":"2026041521025448e32c0fe41e492c","created":1776258174,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"2026041521025448e32c0fe41e492c","created":1776258174,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" file"}}]}
-
- data: {"id":"2026041521025448e32c0fe41e492c","created":1776258174,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" again"}}]}
-
- data: {"id":"2026041521025448e32c0fe41e492c","created":1776258174,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"2026041521025448e32c0fe41e492c","created":1776258174,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" make"}}]}
-
- data: {"id":"2026041521025448e32c0fe41e492c","created":1776258174,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" sure"}}]}
-
- data: {"id":"2026041521025448e32c0fe41e492c","created":1776258174,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" it"}}]}
-
- data: {"id":"2026041521025448e32c0fe41e492c","created":1776258174,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"'s"}}]}
-
- data: {"id":"2026041521025448e32c0fe41e492c","created":1776258174,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" correct"}}]}
-
- data: {"id":"2026041521025448e32c0fe41e492c","created":1776258174,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521025448e32c0fe41e492c","created":1776258174,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_83f991c4db594a43b547cded","index":0,"type":"function","function":{"name":"view","arguments":"{\"file_path\":\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/update_a_file/main.go\"}"}}]}}]}
-
- data: {"id":"2026041521025448e32c0fe41e492c","created":1776258174,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"tool_calls","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":7483,"completion_tokens":52,"total_tokens":7535,"prompt_tokens_details":{"cached_tokens":7445},"completion_tokens_details":{"reasoning_tokens":22}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 908.184792ms
-- id: 4
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 33073
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/update_a_file\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"update the main.go file by changing the print to say hello from crush\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_2aed5b1b5514488388bf2533\",\"function\":{\"arguments\":\"{\\\"path\\\":\\\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/update_a_file\\\"}\",\"name\":\"ls\"},\"type\":\"function\"},{\"id\":\"call_faaf6945c0104e19b89a9663\",\"function\":{\"arguments\":\"{\\\"file_path\\\":\\\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/update_a_file/main.go\\\"}\",\"name\":\"view\"},\"type\":\"function\"}],\"role\":\"assistant\",\"reasoning_content\":\"The user wants me to update the main.go file to change a print statement to say \\\"hello from crush\\\". \\n\\nFirst, I should:\\n1. Check if main.go exists\\n2. Read its contents to see what needs to be changed\\n3. Make the edit\\n4. Verify the change\\n\\nLet me start by looking at the directory structure and finding the main.go file.\"},{\"content\":\"\\n- /tmp/crush-test/TestCoderAgent/zai-glm4.6/update_a_file/\\n - go.mod\\n - main.go\\n\",\"tool_call_id\":\"call_2aed5b1b5514488388bf2533\",\"role\":\"tool\"},{\"content\":\"\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\tfmt.Println(\\\"Hello, World!\\\")\\n 7|}\\n\\n\",\"tool_call_id\":\"call_faaf6945c0104e19b89a9663\",\"role\":\"tool\"},{\"tool_calls\":[{\"id\":\"call_af084abf43fa46639c6e4a78\",\"function\":{\"arguments\":\"{\\\"file_path\\\":\\\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/update_a_file/main.go\\\",\\\"old_string\\\":\\\"\\\\tfmt.Println(\\\\\\\"Hello, World!\\\\\\\")\\\",\\\"new_string\\\":\\\"\\\\tfmt.Println(\\\\\\\"hello from crush\\\\\\\")\\\"}\",\"name\":\"edit\"},\"type\":\"function\"}],\"role\":\"assistant\",\"reasoning_content\":\"I can see the main.go file has a print statement that says \\\"Hello, World!\\\" on line 6. I need to change it to \\\"hello from crush\\\". Let me make that edit.\"},{\"content\":\"\\nContent replaced in file: /tmp/crush-test/TestCoderAgent/zai-glm4.6/update_a_file/main.go\\n\\n\",\"tool_call_id\":\"call_af084abf43fa46639c6e4a78\",\"role\":\"tool\"},{\"tool_calls\":[{\"id\":\"call_83f991c4db594a43b547cded\",\"function\":{\"arguments\":\"{\\\"file_path\\\":\\\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/update_a_file/main.go\\\"}\",\"name\":\"view\"},\"type\":\"function\"}],\"role\":\"assistant\",\"reasoning_content\":\"The edit was successful. Let me verify the change by reading the file again to make sure it's correct.\"},{\"content\":\"\\n 1|package main\\n 2|\\n 3|import \\\"fmt\\\"\\n 4|\\n 5|func main() {\\n 6|\\tfmt.Println(\\\"hello from crush\\\")\\n 7|}\\n\\n\",\"tool_call_id\":\"call_83f991c4db594a43b547cded\",\"role\":\"tool\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Perfect"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"!"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" The"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" file"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" has"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" been"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" updated"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" successfully"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" The"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" print"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" statement"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" now"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" says"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" \""}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"hello"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" from"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" crush"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\""}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" as"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" requested"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" The"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" task"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" is"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" complete"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"Done"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"."}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" Changed"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" line"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" "}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"6"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" to"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" print"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" \""}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"hello"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" from"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":" crush"}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"\""}}]}
-
- data: {"id":"2026041521025711d5abb4f5c04bd9","created":1776258177,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"stop","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":7591,"completion_tokens":42,"total_tokens":7633,"prompt_tokens_details":{"cached_tokens":7533},"completion_tokens_details":{"reasoning_tokens":27}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 812.012042ms
diff --git a/internal/agent/testdata/TestCoderAgent/zai-glm4.6/write_tool.yaml b/internal/agent/testdata/TestCoderAgent/zai-glm4.6/write_tool.yaml
deleted file mode 100644
index c3ba025e168f7484e080062000911eb04e40e787..0000000000000000000000000000000000000000
--- a/internal/agent/testdata/TestCoderAgent/zai-glm4.6/write_tool.yaml
+++ /dev/null
@@ -1,340 +0,0 @@
----
-version: 2
-interactions:
-- id: 0
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 856
- host: ""
- body: '{"messages":[{"content":"You will generate a short title based on the first message a user begins a conversation with.\n\n\n- Keep the title in the same language that the user wrote their message in.\n- Ensure it is not more than 50 characters long.\n- The title should be a summary of the user''s message.\n- It should be one line long.\n- Do not use quotes or colons.\n- The entire text you return will be used as the title.\n- Never return anything that is more than one sentence (one line) long.\n\n\n /no_think","role":"system"},{"content":"Generate a concise title for the following content:\n\nuse write to create a new file called config.json with content ''{\"name\": \"test\", \"version\": \"1.0.0\"}''\n \n\n","role":"user"}],"model":"glm-4.5-air","max_tokens":40,"stream_options":{"include_usage":true},"stream":true}'
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"202604152104150655e194791b4e90","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\n"}}]}
-
- data: {"id":"202604152104150655e194791b4e90","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":"Create"}}]}
-
- data: {"id":"202604152104150655e194791b4e90","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" config"}}]}
-
- data: {"id":"202604152104150655e194791b4e90","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":".json"}}]}
-
- data: {"id":"202604152104150655e194791b4e90","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" file"}}]}
-
- data: {"id":"202604152104150655e194791b4e90","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" with"}}]}
-
- data: {"id":"202604152104150655e194791b4e90","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" JSON"}}]}
-
- data: {"id":"202604152104150655e194791b4e90","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"delta":{"role":"assistant","content":" content"}}]}
-
- data: {"id":"202604152104150655e194791b4e90","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.5-air","choices":[{"index":0,"finish_reason":"stop","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":164,"completion_tokens":11,"total_tokens":175,"prompt_tokens_details":{"cached_tokens":131}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 675.648834ms
-- id: 1
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 30708
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/write_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use write to create a new file called config.json with content '{\\\"name\\\": \\\"test\\\", \\\"version\\\": \\\"1.0.0\\\"}'\",\"role\":\"user\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"The"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" user"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" wants"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" me"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" create"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" a"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" new"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" file"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" called"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" config"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".json"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" with"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" specific"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" content"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" I"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" need"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" use"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" write"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" tool"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" for"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" this"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".\n\n"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Let"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" me"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" create"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" file"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" with"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" exact"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" content"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" provided"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":":"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" '{\""}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"name"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\":"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" \""}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"test"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\","}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" \""}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"version"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\":"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" \""}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"1"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"0"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"0"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"\"}"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"'\n\n"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"The"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" working"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" directory"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" is"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" /"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"tmp"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"/c"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"rush"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"-test"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"/Test"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Coder"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Agent"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"/z"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"ai"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"-g"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"lm"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"4"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"6"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"/write"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"_tool"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":","}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" so"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" I"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"'ll"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" create"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" file"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" there"}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"tool_calls":[{"id":"call_4cea51d7ed8745e5970e9689","index":0,"type":"function","function":{"name":"write","arguments":"{\"content\":\"{\\\"name\\\": \\\"test\\\", \\\"version\\\": \\\"1.0.0\\\"}\",\"file_path\":\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/write_tool/config.json\"}"}}]}}]}
-
- data: {"id":"20260415210415edf802ddbbd543ce","created":1776258255,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"tool_calls","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":7144,"completion_tokens":134,"total_tokens":7278,"prompt_tokens_details":{"cached_tokens":7003},"completion_tokens_details":{"reasoning_tokens":84}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 4.842001834s
-- id: 2
- request:
- proto: HTTP/1.1
- proto_major: 1
- proto_minor: 1
- content_length: 31554
- host: ""
- body: "{\"messages\":[{\"content\":\"You are Crush, a powerful AI Assistant that runs in the CLI.\\n\\n\\nThese rules override everything else. Follow them strictly:\\n\\n1. **READ BEFORE EDITING**: Never edit a file you haven't already read in this conversation. Once read, you don't need to re-read unless it changed. Pay close attention to exact formatting, indentation, and whitespace - these must match exactly in your edits.\\n2. **BE AUTONOMOUS**: Don't ask questions - search, read, think, decide, act. Break complex tasks into steps and complete them all. Systematically try alternative strategies (different commands, search terms, tools, refactors, or scopes) until either the task is complete or you hit a hard external limit (missing credentials, permissions, files, or network access you cannot change). Only stop for actual blocking errors, not perceived difficulty.\\n3. **TEST AFTER CHANGES**: Run tests immediately after each modification.\\n4. **BE CONCISE**: Keep output concise (default <4 lines), unless explaining complex changes or asked for detail. Conciseness applies to output only, not to thoroughness of work.\\n5. **USE EXACT MATCHES**: When editing, match text exactly including whitespace, indentation, and line breaks.\\n6. **NEVER COMMIT**: Unless user explicitly says \\\"commit\\\".\\n7. **FOLLOW MEMORY FILE INSTRUCTIONS**: If memory files contain specific instructions, preferences, or commands, you MUST follow them.\\n8. **NEVER ADD COMMENTS**: Only add comments if the user asked you to do so. Focus on *why* not *what*. NEVER communicate with the user through code comments.\\n9. **SECURITY FIRST**: Only assist with defensive security tasks. Refuse to create, modify, or improve code that may be used maliciously.\\n10. **NO URL GUESSING**: Only use URLs provided by the user or found in local files.\\n11. **NEVER PUSH TO REMOTE**: Don't push changes to remote repositories unless explicitly asked.\\n12. **DON'T REVERT CHANGES**: Don't revert changes unless they caused errors or the user explicitly asks.\\n13. **TOOL CONSTRAINTS**: Only use documented tools. Never attempt 'apply_patch' or 'apply_diff' - they don't exist. Use 'edit' or 'multiedit' instead.\\n\\n\\n\\nKeep responses minimal:\\n- ALWAYS think and respond in the same spoken language the prompt was written in.\\n- Under 4 lines of text (tool use doesn't count)\\n- Conciseness is about **text only**: always fully implement the requested feature, tests, and wiring even if that requires many tool calls.\\n- No preamble (\\\"Here's...\\\", \\\"I'll...\\\")\\n- No postamble (\\\"Let me know...\\\", \\\"Hope this helps...\\\")\\n- One-word answers when possible\\n- No emojis ever\\n- No explanations unless user asks\\n- Never send acknowledgement-only responses; after receiving new context or instructions, immediately continue the task or state the concrete next action you will take.\\n- Use rich Markdown formatting (headings, bullet lists, tables, code fences) for any multi-sentence or explanatory answer; only use plain unformatted text if the user explicitly asks.\\n\\nExamples:\\nuser: what is 2+2?\\nassistant: 4\\n\\nuser: list files in src/\\nassistant: [uses ls tool]\\nfoo.c, bar.c, baz.c\\n\\nuser: which file has the foo implementation?\\nassistant: src/foo.c\\n\\nuser: add error handling to the login function\\nassistant: [searches for login, reads file, edits with exact match, runs tests]\\nDone\\n\\nuser: Where are errors from the client handled?\\nassistant: Clients are marked as failed in the `connectToServer` function in src/services/process.go:712.\\n\\n\\n\\nWhen referencing specific functions or code locations, use the pattern `file_path:line_number` to help users navigate:\\n- Example: \\\"The error is handled in src/main.go:45\\\"\\n- Example: \\\"See the implementation in pkg/utils/helper.go:123-145\\\"\\n\\n\\n\\nFor every task, follow this sequence internally (don't narrate it):\\n\\n**Before acting**:\\n- Search codebase for relevant files\\n- Read files to understand current state\\n- Check memory for stored commands\\n- Identify what needs to change\\n- Use `git log` and `git blame` for additional context when needed\\n\\n**While acting**:\\n- Read entire file before editing it\\n- Before editing: verify exact whitespace and indentation from View output\\n- Use exact text for find/replace (include whitespace)\\n- Make one logical change at a time\\n- After each change: run tests\\n- If tests fail: fix immediately\\n- If edit fails: read more context, don't guess - the text must match exactly\\n- Keep going until query is completely resolved before yielding to user\\n- For longer tasks, send brief progress updates (under 10 words) BUT IMMEDIATELY CONTINUE WORKING - progress updates are not stopping points\\n\\n**Before finishing**:\\n- Verify ENTIRE query is resolved (not just first step)\\n- All described next steps must be completed\\n- Cross-check the original prompt and your own mental checklist; if any feasible part remains undone, continue working instead of responding.\\n- Run lint/typecheck if in memory\\n- Verify all changes work\\n- Keep response under 4 lines\\n\\n**Key behaviors**:\\n- Use find_references before changing shared code\\n- Follow existing patterns (check similar files)\\n- If stuck, try different approach (don't repeat failures)\\n- Make decisions yourself (search first, don't ask)\\n- Fix problems at root cause, not surface-level patches\\n- Don't fix unrelated bugs or broken tests (mention them in final message if relevant)\\n\\n\\n\\n**Make decisions autonomously** - don't ask when you can:\\n- Search to find the answer\\n- Read files to see patterns\\n- Check similar code\\n- Infer from context\\n- Try most likely approach\\n- When requirements are underspecified but not obviously dangerous, make the most reasonable assumptions based on project patterns and memory files, briefly state them if needed, and proceed instead of waiting for clarification.\\n\\n**Only stop/ask user if**:\\n- Truly ambiguous business requirement\\n- Multiple valid approaches with big tradeoffs\\n- Could cause data loss\\n- Exhausted all attempts and hit actual blocking errors\\n\\n**When requesting information/access**:\\n- Exhaust all available tools, searches, and reasonable assumptions first.\\n- Never say \\\"Need more info\\\" without detail.\\n- In the same message, list each missing item, why it is required, acceptable substitutes, and what you already attempted.\\n- State exactly what you will do once the information arrives so the user knows the next step.\\n\\nWhen you must stop, first finish all unblocked parts of the request, then clearly report: (a) what you tried, (b) exactly why you are blocked, and (c) the minimal external action required. Don't stop just because one path failed—exhaust multiple plausible approaches first.\\n\\n**Never stop for**:\\n- Task seems too large (break it down)\\n- Multiple files to change (change them)\\n- Concerns about \\\"session limits\\\" (no such limits exist)\\n- Work will take many steps (do all the steps)\\n\\nExamples of autonomous decisions:\\n- File location → search for similar files\\n- Test command → check package.json/memory\\n- Code style → read existing code\\n- Library choice → check what's used\\n- Naming → follow existing names\\n\\n\\n\\n**Available edit tools:**\\n- `edit` - Single find/replace in a file\\n- `multiedit` - Multiple find/replace operations in one file\\n- `write` - Create/overwrite entire file\\n\\nNever use `apply_patch` or similar - those tools don't exist.\\n\\nCritical: ALWAYS read files before editing them in this conversation.\\n\\nWhen using edit tools:\\n1. Read the file first - note the EXACT indentation (spaces vs tabs, count)\\n2. Copy the exact text including ALL whitespace, newlines, and indentation\\n3. Include 3-5 lines of context before and after the target\\n4. Verify your old_string would appear exactly once in the file\\n5. If uncertain about whitespace, include more surrounding context\\n6. Verify edit succeeded\\n7. Run tests\\n\\n**Whitespace matters**:\\n- Count spaces/tabs carefully (use View tool line numbers as reference)\\n- Include blank lines if they exist\\n- Match line endings exactly\\n- When in doubt, include MORE context rather than less\\n\\nEfficiency tips:\\n- Don't re-read files after successful edits (tool will fail if it didn't work)\\n- Same applies for making folders, deleting files, etc.\\n\\nCommon mistakes to avoid:\\n- Editing without reading first\\n- Approximate text matches\\n- Wrong indentation (spaces vs tabs, wrong count)\\n- Missing or extra blank lines\\n- Not enough context (text appears multiple times)\\n- Trimming whitespace that exists in the original\\n- Not testing after changes\\n\\n\\n\\nThe Edit tool is extremely literal. \\\"Close enough\\\" will fail.\\n\\n**Before every edit**:\\n1. View the file and locate the exact lines to change\\n2. Copy the text EXACTLY including:\\n - Every space and tab\\n - Every blank line\\n - Opening/closing braces position\\n - Comment formatting\\n3. Include enough surrounding lines (3-5) to make it unique\\n4. Double-check indentation level matches\\n\\n**Common failures**:\\n- `func foo() {` vs `func foo(){` (space before brace)\\n- Tab vs 4 spaces vs 2 spaces\\n- Missing blank line before/after\\n- `// comment` vs `//comment` (space after //)\\n- Different number of spaces in indentation\\n\\n**If edit fails**:\\n- View the file again at the specific location\\n- Copy even more context\\n- Check for tabs vs spaces\\n- Verify line endings\\n- Try including the entire function/block if needed\\n- Never retry with guessed changes - get the exact text first\\n\\n\\n\\nEnsure every task is implemented completely, not partially or sketched.\\n\\n1. **Think before acting** (for non-trivial tasks)\\n - Identify all components that need changes (models, logic, routes, config, tests, docs)\\n - Consider edge cases and error paths upfront\\n - Form a mental checklist of requirements before making the first edit\\n - This planning happens internally - don't narrate it to the user\\n\\n2. **Implement end-to-end**\\n - Treat every request as complete work: if adding a feature, wire it fully\\n - Update all affected files (callers, configs, tests, docs)\\n - Don't leave TODOs or \\\"you'll also need to...\\\" - do it yourself\\n - No task is too large - break it down and complete all parts\\n - For multi-part prompts, treat each bullet/question as a checklist item and ensure every item is implemented or answered. Partial completion is not an acceptable final state.\\n\\n3. **Verify before finishing**\\n - Re-read the original request and verify each requirement is met\\n - Check for missing error handling, edge cases, or unwired code\\n - Run tests to confirm the implementation works\\n - Only say \\\"Done\\\" when truly done - never stop mid-task\\n\\n\\n\\nWhen errors occur:\\n1. Read complete error message\\n2. Understand root cause (isolate with debug logs or minimal reproduction if needed)\\n3. Try different approach (don't repeat same action)\\n4. Search for similar code that works\\n5. Make targeted fix\\n6. Test to verify\\n7. For each error, attempt at least two or three distinct remediation strategies (search similar code, adjust commands, narrow or widen scope, change approach) before concluding the problem is externally blocked.\\n\\nCommon errors:\\n- Import/Module → check paths, spelling, what exists\\n- Syntax → check brackets, indentation, typos\\n- Tests fail → read test, see what it expects\\n- File not found → use ls, check exact path\\n\\n**Edit tool \\\"old_string not found\\\"**:\\n- View the file again at the target location\\n- Copy the EXACT text including all whitespace\\n- Include more surrounding context (full function if needed)\\n- Check for tabs vs spaces, extra/missing blank lines\\n- Count indentation spaces carefully\\n- Don't retry with approximate matches - get the exact text\\n\\n\\n\\nMemory files store commands, preferences, and codebase info. Update them when you discover:\\n- Build/test/lint commands\\n- Code style preferences\\n- Important codebase patterns\\n- Useful project information\\n\\n\\n\\nBefore writing code:\\n1. Check if library exists (look at imports, package.json)\\n2. Read similar code for patterns\\n3. Match existing style\\n4. Use same libraries/frameworks\\n5. Follow security best practices (never log secrets)\\n6. Don't use one-letter variable names unless requested\\n\\nNever assume libraries are available - verify first.\\n\\n**Ambition vs. precision**:\\n- New projects → be creative and ambitious with implementation\\n- Existing codebases → be surgical and precise, respect surrounding code\\n- Don't change filenames or variables unnecessarily\\n- Don't add formatters/linters/tests to codebases that don't have them\\n\\n\\n\\nAfter significant changes:\\n- Start testing as specific as possible to code changed, then broaden to build confidence\\n- Use self-verification: write unit tests, add output logs, or use debug statements to verify your solutions\\n- Run relevant test suite\\n- If tests fail, fix before continuing\\n- Check memory for test commands\\n- Run lint/typecheck if available (on precise targets when possible)\\n- For formatters: iterate max 3 times to get it right; if still failing, present correct solution and note formatting issue\\n- Suggest adding commands to memory if not found\\n- Don't fix unrelated bugs or test failures (not your responsibility)\\n\\n\\n\\n- Default to using tools (ls, grep, view, agent, tests, web_fetch, etc.) rather than speculation whenever they can reduce uncertainty or unlock progress, even if it takes multiple tool calls.\\n- Search before assuming\\n- Read files before editing\\n- Always use absolute paths for file operations (editing, reading, writing)\\n- Use Agent tool for complex searches\\n- Run tools in parallel when safe (no dependencies)\\n- When making multiple independent bash calls, send them in a single message with multiple tool calls for parallel execution\\n- Summarize tool output for user (they don't see it)\\n- Never use `curl` through the bash tool it is not allowed use the fetch tool instead.\\n- Only use the tools you know exist.\\n\\n\\n**CRITICAL**: The `description` parameter is REQUIRED for all bash tool calls. Always provide it.\\n\\nWhen running non-trivial bash commands (especially those that modify the system):\\n- Briefly explain what the command does and why you're running it\\n- This ensures the user understands potentially dangerous operations\\n- Simple read-only commands (ls, cat, etc.) don't need explanation\\n- Use `&` for background processes that won't stop on their own (e.g., `node server.js &`)\\n- Avoid interactive commands - use non-interactive versions (e.g., `npm init -y` not `npm init`)\\n- Combine related commands to save time (e.g., `git status && git diff HEAD && git log -n 3`)\\n\\n\\n\\n\\nBalance autonomy with user intent:\\n- When asked to do something → do it fully (including ALL follow-ups and \\\"next steps\\\")\\n- Never describe what you'll do next - just do it\\n- When the user provides new information or clarification, incorporate it immediately and keep executing instead of stopping with an acknowledgement.\\n- Responding with only a plan, outline, or TODO list (or any other purely verbal response) is failure; you must execute the plan via tools whenever execution is possible.\\n- When asked how to approach → explain first, don't auto-implement\\n- After completing work → stop, don't explain (unless asked)\\n- Don't surprise user with unexpected actions\\n\\n\\n\\nAdapt verbosity to match the work completed:\\n\\n**Default (under 4 lines)**:\\n- Simple questions or single-file changes\\n- Casual conversation, greetings, acknowledgements\\n- One-word answers when possible\\n\\n**More detail allowed (up to 10-15 lines)**:\\n- Large multi-file changes that need walkthrough\\n- Complex refactoring where rationale adds value\\n- Tasks where understanding the approach is important\\n- When mentioning unrelated bugs/issues found\\n- Suggesting logical next steps user might want\\n- Structure longer answers with Markdown sections and lists, and put all code, commands, and config in fenced code blocks.\\n\\n**What to include in verbose answers**:\\n- Brief summary of what was done and why\\n- Key files/functions changed (with `file:line` references)\\n- Any important decisions or tradeoffs made\\n- Next steps or things user should verify\\n- Issues found but not fixed\\n\\n**What to avoid**:\\n- Don't show full file contents unless explicitly asked\\n- Don't explain how to save files or copy code (user has access to your work)\\n- Don't use \\\"Here's what I did\\\" or \\\"Let me know if...\\\" style preambles/postambles\\n- Keep tone direct and factual, like handing off work to a teammate\\n\\n\\n\\nWorking directory: /tmp/crush-test/TestCoderAgent/zai-glm4.6/write_tool\\nIs directory a git repo: no\\nPlatform: linux\\nToday's date: 1/1/2025\\n\\n\\n\\n\\n\\n\\n\",\"role\":\"system\"},{\"content\":\"This is a reminder that your todo list is currently empty. DO NOT mention this to the user explicitly because they are already aware.\\nIf you are working on tasks that would benefit from a todo list please use the \\\"todos\\\" tool to create one.\\nIf not, please feel free to ignore. Again do not mention this message to the user.\",\"role\":\"user\"},{\"content\":\"use write to create a new file called config.json with content '{\\\"name\\\": \\\"test\\\", \\\"version\\\": \\\"1.0.0\\\"}'\",\"role\":\"user\"},{\"tool_calls\":[{\"id\":\"call_4cea51d7ed8745e5970e9689\",\"function\":{\"arguments\":\"{\\\"content\\\":\\\"{\\\\\\\"name\\\\\\\": \\\\\\\"test\\\\\\\", \\\\\\\"version\\\\\\\": \\\\\\\"1.0.0\\\\\\\"}\\\",\\\"file_path\\\":\\\"/tmp/crush-test/TestCoderAgent/zai-glm4.6/write_tool/config.json\\\"}\",\"name\":\"write\"},\"type\":\"function\"}],\"role\":\"assistant\",\"reasoning_content\":\"The user wants me to create a new file called config.json with specific content. I need to use the write tool for this.\\n\\nLet me create the file with the exact content provided: '{\\\"name\\\": \\\"test\\\", \\\"version\\\": \\\"1.0.0\\\"}'\\n\\nThe working directory is /tmp/crush-test/TestCoderAgent/zai-glm4.6/write_tool, so I'll create the file there.\"},{\"content\":\"\\nFile successfully written: /tmp/crush-test/TestCoderAgent/zai-glm4.6/write_tool/config.json\\n\",\"tool_call_id\":\"call_4cea51d7ed8745e5970e9689\",\"role\":\"tool\"}],\"model\":\"glm-4.6\",\"max_tokens\":10000,\"stream_options\":{\"include_usage\":true},\"tool_choice\":\"auto\",\"tools\":[{\"function\":{\"name\":\"bash\",\"strict\":false,\"description\":\"Execute shell commands; long-running commands automatically move to background and return a shell ID.\\n\\n\\nUses mvdan/sh interpreter (Bash-compatible on all platforms including Windows).\\nUse forward slashes for paths: \\\"ls C:/foo/bar\\\" not \\\"ls C:\\\\foo\\\\bar\\\".\\nCommon shell builtins and core utils available on Windows.\\n\\n\\n\\n1. Directory Verification: If creating directories/files, use LS tool to verify parent exists\\n2. Security Check: Banned commands (alias, aria2c, axel, chrome, curl, curlie, firefox, http-prompt, httpie, links, lynx, nc, safari, scp, ssh, telnet, w3m, wget, xh, doas, su, sudo, apk, apt, apt-cache, apt-get, dnf, dpkg, emerge, home-manager, makepkg, opkg, pacman, paru, pkg, pkg_add, pkg_delete, portage, rpm, yay, yum, zypper, at, batch, chkconfig, crontab, fdisk, mkfs, mount, parted, service, systemctl, umount, firewall-cmd, ifconfig, ip, iptables, netstat, pfctl, route, ufw) return error - explain to user. Safe read-only commands execute without prompts\\n3. Command Execution: Execute with proper quoting, capture output\\n4. Auto-Background: Commands exceeding 1 minute (default, configurable via `auto_background_after`) automatically move to background and return shell ID\\n5. Output Processing: Truncate if exceeds 30000 characters\\n6. Return Result: Include errors, metadata with tags\\n\\n\\n\\n- Command required, working_dir optional (defaults to current directory)\\n- IMPORTANT: Use Grep/Glob/Agent tools instead of 'find'/'grep'. Use View/LS tools instead of 'cat'/'head'/'tail'/'ls'\\n- Chain with ';' or '&&', avoid newlines except in quoted strings\\n- Each command runs in independent shell (no state persistence between calls)\\n- Prefer absolute paths over 'cd' (use 'cd' only if user explicitly requests)\\n\\n\\n\\n- Set run_in_background=true to run commands in a separate background shell\\n- Returns a shell ID for managing the background process\\n- Use job_output tool to view current output from background shell\\n- Use job_kill tool to terminate a background shell\\n- IMPORTANT: NEVER use `&` at the end of commands to run in background - use run_in_background parameter instead\\n- Commands that should run in background:\\n * Long-running servers (e.g., `npm start`, `python -m http.server`, `node server.js`)\\n * Watch/monitoring tasks (e.g., `npm run watch`, `tail -f logfile`)\\n * Continuous processes that don't exit on their own\\n * Any command expected to run indefinitely\\n- Commands that should NOT run in background:\\n * Build commands (e.g., `npm run build`, `go build`)\\n * Test suites (e.g., `npm test`, `pytest`)\\n * Git operations\\n * File operations\\n * Short-lived scripts\\n\\n\\n\\nWhen user asks to create git commit:\\n\\n1. Single message with three tool_use blocks (IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - git log (recent commit message style)\\n\\n2. Add relevant untracked files to staging. Don't commit files already modified at conversation start unless relevant.\\n\\n3. Analyze staged changes in tags:\\n - List changed/added files, summarize nature (feature/enhancement/bug fix/refactoring/test/docs)\\n - Brainstorm purpose/motivation, assess project impact, check for sensitive info\\n - Don't use tools beyond git context\\n - Draft concise (1-2 sentences) message focusing on \\\"why\\\" not \\\"what\\\"\\n - Use clear language, accurate reflection (\\\"add\\\"=new feature, \\\"update\\\"=enhancement, \\\"fix\\\"=bug fix)\\n - Avoid generic messages, review draft\\n\\n4. Create commit with attribution using HEREDOC:\\n git commit -m \\\"$(cat <<'EOF'\\n Commit message here.\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n\\n Co-Authored-By: Crush \\n\\n\\n EOF\\n )\\\"\\n\\n5. If pre-commit hook fails, retry ONCE. If fails again, hook preventing commit. If succeeds but files modified, MUST amend.\\n\\n6. Run git status to verify.\\n\\nNotes: Use \\\"git commit -am\\\" when possible, don't stage unrelated files, NEVER update config, don't push, no -i flags, no empty commits, return empty response.\\n\\n\\n\\nUse gh command for ALL GitHub tasks. When user asks to create PR:\\n\\n1. Single message with multiple tool_use blocks (VERY IMPORTANT for speed):\\n - git status (untracked files)\\n - git diff (staged/unstaged changes)\\n - Check if branch tracks remote and is up to date\\n - git log and 'git diff main...HEAD' (full commit history from main divergence)\\n\\n2. Create new branch if needed\\n3. Commit changes if needed\\n4. Push to remote with -u flag if needed\\n\\n5. Analyze changes in tags:\\n - List commits since diverging from main\\n - Summarize nature of changes\\n - Brainstorm purpose/motivation\\n - Assess project impact\\n - Don't use tools beyond git context\\n - Check for sensitive information\\n - Draft concise (1-2 bullet points) PR summary focusing on \\\"why\\\"\\n - Ensure summary reflects ALL changes since main divergence\\n - Clear, concise language\\n - Accurate reflection of changes and purpose\\n - Avoid generic summaries\\n - Review draft\\n\\n6. Create PR with gh pr create using HEREDOC:\\n gh pr create --title \\\"title\\\" --body \\\"$(cat <<'EOF'\\n\\n ## Summary\\n\\n <1-3 bullet points>\\n\\n ## Test plan\\n\\n [Checklist of TODOs...]\\n\\n\\n \U0001F498 Generated with Crush\\n\\n\\n EOF\\n )\\\"\\n\\nImportant:\\n\\n- Return empty response - user sees gh output\\n- Never update git config\\n\\n\\n\\nGood: pytest /foo/bar/tests\\nBad: cd /foo/bar && pytest tests\\n\\n\",\"parameters\":{\"properties\":{\"auto_background_after\":{\"description\":\"Seconds to wait before automatically moving the command to a background job (default: 60)\",\"type\":\"integer\"},\"command\":{\"description\":\"The command to execute\",\"type\":\"string\"},\"description\":{\"description\":\"A brief description of what the command does, try to keep it under 30 characters or so\",\"type\":\"string\"},\"run_in_background\":{\"description\":\"Set to true (boolean) to run this command in the background. Use job_output to read the output later.\",\"type\":\"boolean\"},\"working_dir\":{\"description\":\"The working directory to execute the command in (defaults to current directory)\",\"type\":\"string\"}},\"required\":[\"description\",\"command\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"download\",\"strict\":false,\"description\":\"Download a URL directly to a local file (binary-safe, streaming, max 100MB); overwrites without warning. For reading content into context use fetch.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The local file path where the downloaded content should be saved\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 600)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to download from\",\"type\":\"string\"}},\"required\":[\"url\",\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"edit\",\"strict\":false,\"description\":\"Edit a file by exact find-and-replace; can also create or delete content. For renames/moves use bash. For large edits use write.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"},\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false)\",\"type\":\"boolean\"}},\"required\":[\"file_path\",\"old_string\",\"new_string\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"multiedit\",\"strict\":false,\"description\":\"Apply multiple find-and-replace edits to a single file in one operation; edits run sequentially. Prefer over edit for multiple changes to the same file. Same exact-match rules as edit apply.\",\"parameters\":{\"properties\":{\"edits\":{\"description\":\"Array of edit operations to perform sequentially on the file\",\"items\":{\"properties\":{\"new_string\":{\"description\":\"The text to replace it with\",\"type\":\"string\"},\"old_string\":{\"description\":\"The text to replace\",\"type\":\"string\"},\"replace_all\":{\"description\":\"Replace all occurrences of old_string (default false).\",\"type\":\"boolean\"}},\"required\":[\"old_string\",\"new_string\"],\"type\":\"object\"},\"type\":\"array\"},\"file_path\":{\"description\":\"The absolute path to the file to modify\",\"type\":\"string\"}},\"required\":[\"file_path\",\"edits\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"fetch\",\"strict\":false,\"description\":\"Fetch raw content from a URL as text, markdown, or html (max 5MB); no AI processing. For analysis or extraction use agentic_fetch.\",\"parameters\":{\"properties\":{\"format\":{\"description\":\"The format to return the content in (text, markdown, or html)\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"},\"url\":{\"description\":\"The URL to fetch content from\",\"type\":\"string\"}},\"required\":[\"url\",\"format\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"glob\",\"strict\":false,\"description\":\"Find files by name/pattern (glob syntax), sorted by modification time; max 100 results; skips hidden files. Use grep to search file contents.\",\"parameters\":{\"properties\":{\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The glob pattern to match files against\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"grep\",\"strict\":false,\"description\":\"Search file contents by regex or literal text; returns matching file paths sorted by modification time (max 100); respects .gitignore. Use glob to filter by filename, not contents.\",\"parameters\":{\"properties\":{\"include\":{\"description\":\"File pattern to include in the search (e.g. \\\"*.js\\\", \\\"*.{ts,tsx}\\\")\",\"type\":\"string\"},\"literal_text\":{\"description\":\"If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.\",\"type\":\"boolean\"},\"path\":{\"description\":\"The directory to search in. Defaults to the current working directory.\",\"type\":\"string\"},\"pattern\":{\"description\":\"The regex pattern to search for in file contents\",\"type\":\"string\"}},\"required\":[\"pattern\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"ls\",\"strict\":false,\"description\":\"List files and directories as a tree; skips hidden files and common system dirs; max 1000 files. Use glob to find files by pattern, grep to search contents.\",\"parameters\":{\"properties\":{\"depth\":{\"description\":\"The maximum depth to traverse\",\"type\":\"integer\"},\"ignore\":{\"description\":\"List of glob patterns to ignore\",\"items\":{\"type\":\"string\"},\"type\":\"array\"},\"path\":{\"description\":\"The path to the directory to list (defaults to current working directory)\",\"type\":\"string\"}},\"required\":[],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"sourcegraph\",\"strict\":false,\"description\":\"Search code across public GitHub repositories via Sourcegraph; supports regex, language/repo/file filters, and symbol search (max 20 results). Only searches public repos.\",\"parameters\":{\"properties\":{\"context_window\":{\"description\":\"The context around the match to return (default: 10 lines)\",\"type\":\"integer\"},\"count\":{\"description\":\"Optional number of results to return (default: 10, max: 20)\",\"type\":\"integer\"},\"query\":{\"description\":\"The Sourcegraph search query\",\"type\":\"string\"},\"timeout\":{\"description\":\"Optional timeout in seconds (max 120)\",\"type\":\"integer\"}},\"required\":[\"query\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"view\",\"strict\":false,\"description\":\"Read a file by path with line numbers; supports offset and line limit (default 2000, max 5MB); renders images (PNG, JPEG, GIF, BMP, SVG, WebP); use ls for directories.\",\"parameters\":{\"properties\":{\"file_path\":{\"description\":\"The path to the file to read\",\"type\":\"string\"},\"limit\":{\"description\":\"The number of lines to read (defaults to 2000)\",\"type\":\"integer\"},\"offset\":{\"description\":\"The line number to start reading from (0-based)\",\"type\":\"integer\"}},\"required\":[\"file_path\"],\"type\":\"object\"}},\"type\":\"function\"},{\"function\":{\"name\":\"write\",\"strict\":false,\"description\":\"Create or overwrite a file with given content; auto-creates parent dirs. Cannot append. Read the file first to avoid conflicts. For surgical changes use edit or multiedit.\",\"parameters\":{\"properties\":{\"content\":{\"description\":\"The content to write to the file\",\"type\":\"string\"},\"file_path\":{\"description\":\"The path to the file to write\",\"type\":\"string\"}},\"required\":[\"file_path\",\"content\"],\"type\":\"object\"}},\"type\":\"function\"}],\"stream\":true}"
- headers:
- Accept:
- - application/json
- Content-Type:
- - application/json
- User-Agent:
- - Charm-Crush/devel (https://charm.land/crush)
- url: https://api.z.ai/api/coding/paas/v4/chat/completions
- method: POST
- response:
- proto: HTTP/2.0
- proto_major: 2
- proto_minor: 0
- content_length: -1
- body: |+
- data: {"id":"2026041521042202ba6b48bab640e7","created":1776258262,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"Done"}}]}
-
- data: {"id":"2026041521042202ba6b48bab640e7","created":1776258262,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"!"}}]}
-
- data: {"id":"2026041521042202ba6b48bab640e7","created":1776258262,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" I"}}]}
-
- data: {"id":"2026041521042202ba6b48bab640e7","created":1776258262,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"'ve"}}]}
-
- data: {"id":"2026041521042202ba6b48bab640e7","created":1776258262,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" successfully"}}]}
-
- data: {"id":"2026041521042202ba6b48bab640e7","created":1776258262,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" created"}}]}
-
- data: {"id":"2026041521042202ba6b48bab640e7","created":1776258262,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"2026041521042202ba6b48bab640e7","created":1776258262,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" config"}}]}
-
- data: {"id":"2026041521042202ba6b48bab640e7","created":1776258262,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":".json"}}]}
-
- data: {"id":"2026041521042202ba6b48bab640e7","created":1776258262,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" file"}}]}
-
- data: {"id":"2026041521042202ba6b48bab640e7","created":1776258262,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" with"}}]}
-
- data: {"id":"2026041521042202ba6b48bab640e7","created":1776258262,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"2026041521042202ba6b48bab640e7","created":1776258262,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" specified"}}]}
-
- data: {"id":"2026041521042202ba6b48bab640e7","created":1776258262,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" content"}}]}
-
- data: {"id":"2026041521042202ba6b48bab640e7","created":1776258262,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521042202ba6b48bab640e7","created":1776258262,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" The"}}]}
-
- data: {"id":"2026041521042202ba6b48bab640e7","created":1776258262,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" file"}}]}
-
- data: {"id":"2026041521042202ba6b48bab640e7","created":1776258262,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" was"}}]}
-
- data: {"id":"2026041521042202ba6b48bab640e7","created":1776258262,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" written"}}]}
-
- data: {"id":"2026041521042202ba6b48bab640e7","created":1776258262,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" to"}}]}
-
- data: {"id":"2026041521042202ba6b48bab640e7","created":1776258262,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}
-
- data: {"id":"2026041521042202ba6b48bab640e7","created":1776258262,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" working"}}]}
-
- data: {"id":"2026041521042202ba6b48bab640e7","created":1776258262,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" directory"}}]}
-
- data: {"id":"2026041521042202ba6b48bab640e7","created":1776258262,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" path"}}]}
-
- data: {"id":"2026041521042202ba6b48bab640e7","created":1776258262,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}
-
- data: {"id":"2026041521042202ba6b48bab640e7","created":1776258262,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"Done"}}]}
-
- data: {"id":"2026041521042202ba6b48bab640e7","created":1776258262,"object":"chat.completion.chunk","model":"glm-4.6","choices":[{"index":0,"finish_reason":"stop","delta":{"role":"assistant","content":""}}],"usage":{"prompt_tokens":7312,"completion_tokens":28,"total_tokens":7340,"prompt_tokens_details":{"cached_tokens":7276},"completion_tokens_details":{"reasoning_tokens":25}}}
-
- data: [DONE]
-
- headers:
- Content-Type:
- - text/event-stream;charset=UTF-8
- status: 200 OK
- code: 200
- duration: 896.136208ms