Skip to content
cccskills
FreePublished on GitHub

polish-beads

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

The English summary is being prepared. Showing the original description.

See how to install

Files included (1)

  • SKILL.md1.4 KB

SKILL.md (original)

Review the instructions the agent will receive before you install.

/polish-beads — convergence loop

Single most-underinvested step in the methodology. Every polish round is ~100× cheaper than finding the same issue during implementation.

Prompt

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

What to check each round

  1. Self-containment. Fresh agent, no plan — can they execute?
  2. Dependencies. depends_on correct? Missing edges?
  3. Tests. Concrete unit + e2e obligations with detailed logging?
  4. Coverage. Cross-reference beads vs .flywheel/plans/*.md.
  5. Duplicates. Merge overlapping beads into canonical ones.
  6. Forge invariants. Permission gate, LEGAL_TRANSITIONS, UI budget, provider probe timing.

Run count

4–6 rounds typical. Stop when:

  • two consecutive rounds produce only small corrective edits,
  • no new structural issues surface,
  • dependency graph is stable.

If polishing has flatlined after 3 rounds, switch to /fresh-eyes — a new session with no context baggage.

If 20+ new beads were added in a batch, also run /dedupe-beads.

Final round: different model

As a final polishing pass, run this same prompt through a different model (e.g. switch from Claude to Codex, or vice versa). Different models catch different things.

Reviews

No reviews yet. Tell others what you thought.

Comments

No login needed. Share how it worked for you, ask a question, or note where it did not run.

No comments yet. Be the first to write one.

Comments are published right away. Do not include personal information such as email addresses or phone numbers. Up to 2 URLs.

0 / 1,000 characters

This form is protected by reCAPTCHA, and the Google Privacy Policy and Terms of Service apply.

More skills from this repository

Scaffold a new Forge subagent (planner/architect/executor/reviewer/debugger/memory-style). Use when adding anything to src/agents/.

The English summary is being prepared. Showing the original description.

hoangsonww/Forge-Agentic-Coding-CLI26Updated Sep 28, 2026

Scaffold a new Forge subagent (planner/architect/executor/reviewer/debugger/memory-style).

The English summary is being prepared. Showing the original description.

hoangsonww/Forge-Agentic-Coding-CLI26Updated Sep 28, 2026

Scaffold a new model provider for Forge — class implementation, registry wiring, providerEnum entry, local-catalog classification, and unit test with mocked transport. Use for anything under src/models/.

The English summary is being prepared. Showing the original description.

hoangsonww/Forge-Agentic-Coding-CLI26Updated Sep 28, 2026

Scaffold a new model provider for Forge — class, registry wiring, providerEnum entry, local-catalog classification, router check, and a unit test with mocked transport.

The English summary is being prepared. Showing the original description.

hoangsonww/Forge-Agentic-Coding-CLI26Updated Sep 28, 2026

Scaffold a new Forge tool end-to-end — type, implementation, registration, permission/risk classification, and unit test. Use when adding anything to src/tools/.

The English summary is being prepared. Showing the original description.

hoangsonww/Forge-Agentic-Coding-CLI26Updated Sep 28, 2026

Scaffold a new Forge tool — implementation, registration, permission/risk metadata, sandbox routing, and a unit test. Use when adding anything to src/tools/.

The English summary is being prepared. Showing the original description.

hoangsonww/Forge-Agentic-Coding-CLI26Updated Sep 28, 2026

See all skills by hoangsonww

Report a problem with this skill