From 77e68b4e41fa72d301b2f4ca948f3c4806403fba Mon Sep 17 00:00:00 2001 From: Ben Brandt Date: Mon, 23 Feb 2026 17:32:45 +0100 Subject: [PATCH] agent: Remove warning for spawn_agent (#49903) It seems this was working too well and the model wasn't choosing to use subagents at all. Release Notes: - N/A --- crates/agent/src/tools/spawn_agent_tool.rs | 2 -- 1 file changed, 2 deletions(-) diff --git a/crates/agent/src/tools/spawn_agent_tool.rs b/crates/agent/src/tools/spawn_agent_tool.rs index 7392a85cfe256fbfdf53a808925528923dfa93d2..e2dd78d4476de48465cb5c48e225e2ae5a0a7767 100644 --- a/crates/agent/src/tools/spawn_agent_tool.rs +++ b/crates/agent/src/tools/spawn_agent_tool.rs @@ -17,8 +17,6 @@ use crate::{AgentTool, Thread, ThreadEnvironment, ToolCallEventStream}; /// - Complete a self-contained task where you need to know if it succeeded or failed (and how), but none of its intermediate output. /// - Perform an investigation where all you need to know is the outcome, not the research that led to that outcome. /// -/// Do NOT use this tool for simple tasks you could accomplish directly with one or two tool calls (e.g. reading a file, running a single command). Each agent has startup overhead. -/// /// You control what the agent does by providing a prompt describing what the agent should do. The agent has access to the same tools you do, but does NOT see your conversation history or any context the user attached. You must include all relevant context (file paths, requirements, constraints) in the prompt. /// /// You will receive only the agent's final message as output.