菜单
Use when starting any conversation — establishes how to find and use skills, requiring skill invocation before ANY response including clarifying questions. 啟話即用：定技能發現與調用之法，凡答（含反問）前必先察技能。
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
Record + replay an interactive guided demo overlay on the live app — self-advancing walkthrough, labelled gesture affordances. 瀏覽器互動導覽。 Use when: demo a feature, create walkthrough, run walkthrough tool, guided product tour, replay a flow, show what was built, onboarding tour, step-by-step demo overlay.
Adversarial cooperation loop for code quality verification with plan adjustment at each phase. 對抗協作循環：實施者與驗證者互相制衡，逐階調整計劃，確保代碼品質。 Use when: execute dart task, run quality pipeline, adversarial review, implement with verification
Search + synthesize prior agent session history. Claude Code primary (~/.claude/projects/); Codex/Cursor secondary. Discover -> branch/keyword filter -> extract skeleton/errors -> session-historian subagent. Use when: what was tried, prior attempts, past decisions, recent work, why approach abandoned, cross-session context, even without saying 'sessions'.
Canonical verdict-only return schema for adversarial reviewer subagents. Reviewers emit compact ≤30-line verdict block; main loop parses verdict + blockers only. Use when: authoring or updating reviewer agent, building verdict consumer, debugging reviewer report parsing.
Diagnose + configure compound-engineering env — CLI deps, plugin version, repo-local config. Guided install for missing tools, gitignore + config bootstrap. Use when: troubleshoot missing tools, verify CE setup, onboard repo, refresh config template, /ce-setup re-check. Skip: unrelated to compound-engineering tooling.
Resolve review feedback by evaluating validity and fixing issues in parallel. Pre-commit-first. Use when: applying review feedback from local reviewer agent output, resolving pre-commit review findings, working through code-review agent suggestions, addressing reviewer-flagged issues before commit. Also: PR comments via gh as secondary path. 解決審查回饋：本地審查代理輸出優先，PR次之。
| name | using-superpowers |
| description | Use when starting any conversation — establishes how to find and use skills, requiring skill invocation before ANY response including clarifying questions. 啟話即用：定技能發現與調用之法，凡答（含反問）前必先察技能。 |
技能適任時，汝無選擇餘地 —— 必用之。此不可議，不可自圓其說繞過。
於任何回應或動作前，調相關/請求之 skill —— 含反問、探 codebase、查檔。若後證不合，不必續用。
入 plan mode 前： 若任務仍需構想、設計收斂或策略選擇，先調 ideation skill。
then 宣告 "Using [skill] to [purpose]"，遵 skill exact。有 checklist 則每項一 todo。
多 skill 適時，process skills 先 —— 定法，後由 implementation skills 執。ideation 與 systematic-debugging 為最常之 process skill，然此則通用於任一。
ideation 先收斂目的、策略、principles/decision tree/plan draft，後 implementation skills。systematic-debugging 先，後 domain skills。test-driven-development（red→green→refactor）。規劃、執行、branch-finishing 委予 worktrack-loop plugin，非 superpowers：
/worktrack:plan（worktrack-scope-tasks skill）—— 上下文預算切分、最輕 model+effort、測試先行、fresh-context 對抗審。/worktrack:start（或 /loop /worktrack:start --max-tasks 1 乾隊）。/worktrack:task。ideation 之 terminal handoff 可指 /worktrack:plan，代舊 writing-plans → executing-plans 鏈。
下列由 harness 原生具，用之，勿另立 skill：
| 需 | 原生 |
|---|---|
| 驗證完成 | /verify skill |
| Code review | /code-review（含 ultra cloud）、/review、/security-review |
| 平行 subagent | Agent tool（run_in_background） |
| Git worktree 隔離 | EnterWorktree/ExitWorktree、Agent isolation:"worktree" |
| Plan mode | EnterPlanMode/ExitPlanMode |
| Skill 發現 | Skill tool、find-skills |
| 建/審 skill | plugin-dev:skill-development、dev-standards:add-skill |
下列念頭 = STOP，汝在自圓：
| 念頭 | 實 |
|---|---|
| "此僅簡單問題" | 問即任。察 skill。 |
| "須先取 context" | Skill 察先於反問。 |
| "先探 codebase" | Skill 教汝如何探。先察。 |
| "快查 git/檔" | 檔缺對話 context。察 skill。 |
| "先蒐資訊" | Skill 教汝如何蒐。 |
| "此不需正式 skill" | Skill 存則用。 |
| "我記得此 skill" | Skill 會演進。讀現版。 |
| "此不算任務" | 動作即任。察 skill。 |
| "Skill 過度" | 簡事會轉繁。用之。 |
| "先做這一件" | 動前先察。 |
User instructions（CLAUDE.md、AGENTS.md、直接請求）勝 skills，skills 勝 default behavior。唯 human partner 明令方可略 skill workflow。