This skill implements a rigorous planning phase for AI agents, ensuring that complex work is properly classified and planned before any code is changed. It prevents 'cowboy coding' by enforcing a clear separation between the PLAN and EXECUTION phases.
Use this skill at the start of any non-trivial implementation work (spanning multiple files, changing APIs) or investigation work that requires a reviewable output.
Designed for high-autonomy agents capable of managing state across multiple files and sessions, such as Claude Code, Codex, and other ACP-compatible harnesses.
This skill has not been reviewed by our automated audit pipeline yet.