メニュー
Overview of Forge's Agentic Coding Flywheel methodology — plan space, bead space, code space. Invoke this skill when you need to know which Flywheel skill or prompt applies to the current phase.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
Deep review — random exploration + cross-agent review, alternating. Finds integration bugs that self-review misses. Run every 30–60 min during active implementation.
Overview of Forge's Agentic Coding Flywheel methodology. Use to route between plan, bead, and code space. See FLYWHEEL.md at the repo root.
End-of-session ritual. Close finished beads, file new ones for remaining work, run verify, commit in logical groupings, push. Work is not complete until git push succeeds.
Start a new Flywheel planning round. Produce initial markdown plan, route to competing frontier models, synthesize hybrid, refine 4–5×. Use before any coding begins on a new feature.
Translate a polished markdown plan into self-contained beads in .beads/beads.jsonl. Every bead must be executable without reopening the plan. Use after the plan has converged.
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.
SOC 職業分類に基づく
| name | flywheel |
| description | Overview of Forge's Agentic Coding Flywheel methodology — plan space, bead space, code space. Invoke this skill when you need to know which Flywheel skill or prompt applies to the current phase. |
Forge follows the Agentic Coding Flywheel methodology: plan obsessively in markdown, translate polished plans into self-contained beads, and only then let a swarm execute.
| If you are… | Use |
|---|---|
| starting a new feature / subsystem | /plan → write the initial plan |
| you have 2+ competing plans from different models | /plan-synthesize → GPT Pro synthesizes hybrid |
| refining a plan | .flywheel/prompts/plan-refine.md (fresh chats, 4–5×) |
| plan feels stable and complete | /plan-to-beads → create .beads/beads.jsonl |
| beads were just created | /polish-beads (run 4–6×) |
| polishing has flatlined | /fresh-eyes → new session, reload docs |
| ≥ 20 new beads were just batched in | /dedupe-beads |
| adding features to existing Forge | /idea-wizard (30 → 5 → 15 funnel) |
| swarm is running, agents finished a bead | self-review inline, then /deep-review every hour |
| swarm feels busy but directionally off | /reality-check |
| end of session | /landing → close beads, commit, push |
| agent confused after compaction | "Reread AGENTS.md, CLAUDE.md, and FLYWHEEL.md" |
.flywheel/plans/*.md (where architecture lives).beads/beads.jsonl (where execution lives)src/ + test/ (where implementation lives).claude/, .codex/, .cursor/, .agents/
(how agents behave)Once polished beads exist, the beads are the source of truth for execution. You don't keep editing the plan while agents implement.
AGENTS.md re-reads, beads).See FLYWHEEL.md at the repo root for the full methodology.