execute-plan

Automate phase plan execution by loading a PLAN.md file and routing work between subagents and the main context. The process splits plans into executable segments, handles deviations, and generates SUMMARY.md files with structured metadata for auditable progress.

Checkpoint handling during autonomous plan execution supports human-verified segments and automatic deviation logging. When execution reaches a checkpoint, the workflow pauses for verification, logs any deviations from the PLAN.md, and resumes once the gate is cleared.

Generate outcome summaries by executing planned segments from PLAN.md, which automatically produces SUMMARY.md files. Completed work is committed via a git-integrated workflow that applies structured metadata and summaries to ensure traceable repository commits.

Routing plan execution tasks between subagents and the main context is supported through autonomous planning. The system splits the PLAN.md into executable segments and directs each segment to the appropriate context for processing and summary generation.

Autonomous plan execution works within existing repository contexts and integrates directly with git workflows. It enforces plan loading, segment routing, deviation handling, and authentication-gate workflows to ensure reliable, auditable progress with structured commits.

Orchestrating multi-phase projects with autonomous execution requires a PLAN.md file to define segments and checkpoints. It is limited to projects structured with phase prompts and depends on proper authentication-gate workflows for reliable, auditable progress.