Draft → refine → self-improve → mark ready → halt for a user-supplied small task.
Use TaskCreate to create a task for each phase:
- Run
/draft-planskill - Run
/refine-planskill - Run
/self-improveskill - Mark plan ready
- Halt with next-step instructions
Run the /draft-plan skill with the user's task description. This produces a new plan at .turbo/plans/<slug>.md. If a slug was passed, forward it. Capture the resolved plan path from /draft-plan's output for the next phases.
Run the /refine-plan skill with <path> from Phase 1. This loops review, evaluation, and application until the plan stabilizes or the iteration cap is hit.
Run the /self-improve skill to compound planning learnings.
Update the plan's YAML frontmatter to status: ready.
Halt with this message:
Plan ready at
<plan path>.Planning context is likely full, and the plan is comprehensive enough to continue fresh. Run
/clear, then/implement-plan <slug>to implement.
- Do not edit the plan file directly. Revisions go through
/refine-planor/draft-plan. - Do not attempt to auto-implement. The user drives implementation with
/implement-planin a fresh session.