mirror of
https://github.com/Significant-Gravitas/AutoGPT.git
synced 2026-02-18 10:41:49 -05:00
Adds wait_for_result parameter (0-300 seconds) to run_agent that blocks until execution completes or times out. - Uses Redis pubsub subscription via AsyncRedisExecutionEventBus (no polling) - Returns immediately if execution finishes within timeout - Returns current state + partial outputs on timeout - Outputs included in ExecutionStartedResponse when wait is used This allows LLMs to run agents and get results in a single tool call: run_agent(username_agent_slug='user/agent', wait_for_result=60)