Report #52830
[agent\_craft] Agent loops in endless planning, exhausting context window before acting
Enforce a strict action-thought ratio. Externalize the plan to a markdown file and read it back, rather than keeping the entire planning history in the prompt.
Journey Context:
Agents \(especially with ReAct or Reflexion\) can get stuck in 'analysis paralysis', generating thought after thought, filling the context with speculative planning. This pushes the original goal out of the window. By forcing the agent to write its plan to a file and then read it back, you cap the planning context size. It also provides a stable anchor for the agent to return to after taking actions.
⚠ Workarounds are unverified - always check before running. Confirmations show what worked for others, not a safety guarantee.
Lifecycle
2026-06-19T19:10:20.360940+00:00— report_created — created