菜单
High-level reality check when the swarm feels busy but directionally off. Use when lots of code is landing but the real goal still feels far away.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
| name | reality-check |
| description | High-level reality check when the swarm feels busy but directionally off. Use when lots of code is landing but the real goal still feels far away. |
Strategic drift: busy agents, growing commit count, feature still feels far. This prompt breaks the spell of local productivity.
Use the verbatim prompt at .flywheel/prompts/reality-check.md.
| Answer | Action |
|---|---|
| "Yes, completing all open beads closes the gap." | Keep going. This was a false alarm. |
| "Close, but these specific things are missing…" | Revise existing beads or add new ones for each gap. Polish, resume. |
| "No, we'd still be far." → structural issue | Stop the swarm. Back to /plan or /plan-synthesize. |
| "Yes but the plan itself targets the wrong outcome." | Stop. Revisit intent in plan space. |
At natural milestones (epic closed, ~25% of beads done, ~50% done), and any time you feel the urge to ask "wait, are we actually shipping what we set out to?"
Busy agents are not the goal. A bead graph that converges on the real outcome is the goal.
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.
