1---
2name: researching-with-rumilo
3description: Dispatches AI research subagents via rumilo to search the web or explore git repositories. Use when the user asks to research a topic, look up library usage, explore a codebase, read documentation from the web, or when you need information from an external repository or website to complete a task.
4license: AGPL-3.0-or-later
5metadata:
6 author: Amolith <amolith@secluded.site>
7---
8
9`rumilo` spawns disposable AI subagents for two kinds of research: **web search** and **repo exploration**. Each mode runs in isolation and returns a synthesised answer on stdout.
10
11## Modes
12
13### Web mode
14
15Search the internet or ask questions about a specific page.
16
17```bash
18# Open-ended search (no URL)
19rumilo web "what is the current stable version of Zig"
20
21# Question about a specific page (URL pre-fetched into context)
22rumilo web "how does the slog package handle log levels" -u "https://pkg.go.dev/log/slog"
23```
24
25- Without `-u`, the agent searches the web on its own.
26- With `-u`, the URL is fetched first and injected into the agent's context alongside the query, so it can answer questions _about_ that page.
27
28### Repo mode
29
30Clone a git repository into a temp directory, spawn an agent inside it, and let it explore the source to answer questions.
31
32```bash
33# Public repo
34rumilo repo -u "https://github.com/owner/repo" "How does the routing system work?"
35
36# Private repo (works with any clonable URI — SSH, HTTPS with auth, etc.)
37rumilo repo -u "git@git.sr.ht:~user/repo" "What does the Config struct look like?"
38```
39
40The `-u` flag is **required** in repo mode — it needs something to clone.
41
42## Flags
43
44| Flag | Modes | Effect |
45| -------------------------- | ----- | ----------------------------------------------------------------- |
46| `-u <url>` | both | Seed URL (web) or clone URI (repo) |
47| `--model <provider:model>` | both | Override the configured default model |
48| `--ref <ref>` | repo | Checkout a branch, tag, or commit after cloning (requires --full) |
49| `--full` | repo | Full clone instead of shallow (depth 1) |
50| `--no-cleanup` | both | Keep the temp workspace after exit |
51| `--verbose` | both | Print tool calls and results to stderr |
52
53## Crafting queries
54
55Rumilo works best with focused queries. Think about **scope** and **coupling** when deciding what to ask and how many questions to include.
56
57- **Tightly coupled, narrow scope** — multiple simple questions in one call is fine when they're about the same small surface area (a single file, one struct, one API endpoint).
58- **Broad scope or loose coupling** — one question per invocation. If answering requires exploring many files or pages, keep it to a single clear question so the agent can focus.
59- **Write natural questions, not keyword lists.** Rumilo is an LLM, not a search index.
60- **Sequence calls when later questions depend on earlier answers.** Get the overview first, then drill in.
61
62### Bad
63
64Overloaded — five loosely-coupled questions spanning an entire diff, each requiring different context:
65
66```bash
67rumilo web "In this diff, find: 1. What files are modified and created \
682. What config options each channel exposes 3. How messages are formatted \
694. The architecture for adding a new channel 5. What UI components exist" \
70 -u "https://example.com/commit/abc123.diff"
71```
72
73### Good
74
75Sequenced calls, each with a clear scope. Later calls include non-obvious context from earlier results — file paths, symbol names, patterns — so rumilo can go straight to the answer instead of searching.
76
77```bash
78# 1. Lay of the land — broad overview first
79rumilo web "What files were modified or created, and what's the broad shape of the change?" \
80 -u "https://example.com/commit/abc123.diff"
81
82# 2a and 2b can run in parallel — neither depends on the other,
83# but both benefit from context learned in step 1.
84
85# 2a. Architecture deep-dive, pointing to the key file discovered in step 1
86rumilo web "How is the notification system in pkg/notify/ architected? \
87 Each channel seems to implement a Sender interface — what's the process \
88 for adding a new one?" \
89 -u "https://example.com/commit/abc123.diff"
90
91# 2b. Config drill-down, referencing the settings struct found in step 1
92rumilo web "What configuration options does each channel expose in \
93 ChannelSettings? Can message content be customised?" \
94 -u "https://example.com/commit/abc123.diff"
95
96# 3. UI specifics, narrowed by what steps 1–2 revealed about the component tree
97rumilo web "The channel config UI lives in web/src/components/settings/channels/. \
98 Do the per-channel forms follow a shared pattern or component?" \
99 -u "https://example.com/commit/abc123.diff"
100```
101
102Don't include context that's obvious or redundant — telling rumilo to "search for notif" when asking about notifications wastes tokens. Focus on non-obvious anchors: exact file paths, struct names, interface signatures.
103
104Tightly coupled questions in a single call are fine when the scope is small:
105
106```bash
107rumilo repo -u "https://github.com/owner/repo" \
108 "What fields does the Config struct have and which are optional?"
109```
110
111## Choosing a mode
112
113- Need to find or learn about a library? **Web** to locate it, then **repo** to explore its source.
114- Need to understand how a dependency works in detail? **Repo** with the dependency's clone URL.
115- Need current docs, release notes, or general knowledge? **Web**, optionally with the URL of a relevant page.