Polish Beads

Polishing pass over all open beads. Run 4–6 times. "Check your beads N times, implement once." Use after plan-to-beads and between refinement rounds.

hoangsonww Updated

File contents

Use the verbatim prompt at .flywheel/prompts/polish-beads.md.

Each round, verify:

  1. Self-containment — a fresh agent can execute without the plan.
  2. Dependencies — depends_on is correct; no missing edges.
  3. Tests — concrete obligations with logging expectations.
  4. Coverage — cross-reference vs .flywheel/plans/*.md.
  5. Duplicates — merge overlapping beads (see dedupe-beads prompt).
  6. Forge invariants — permission gate, LEGAL_TRANSITIONS, UI < 120KB, provider probe ~1.5s.

Stop when two consecutive rounds produce only small corrective edits. If polishing has flatlined after 3 rounds, switch to a fresh session and .flywheel/prompts/fresh-eyes.md.

Final round: run through a different model (e.g. switch Claude ↔ Codex). Different models catch different things.

hoangsonww/forge-agentic-coding-cli/tree/main/.codex/skills/polish-beads commit e6c552eed2

Frequently asked questions

npx skillmds@latest add hoangsonww/polish-beads