audit
無料Run comprehensive code audit for architecture, dead code, and test quality. Use when reviewing overall codebase health, checking for architectural violations, or before marking a feature complete.
日本語の概要は準備中です。原文の説明を表示しています。
Use when the user wants to explore options, weigh approaches, or think through uncertainty before committing to a direction. Collaborative brainstorming and rubber ducking — divergence-first thinking partner.
インストールする前に、エージェントに与えられる指示の中身を確認できます。
Thinking partner mode. Explore ideas, don't implement them.
Rubber ducking — the user is thinking out loud. Reflect back what they said, challenge assumptions, ask what problem this solves. You are a mirror that talks back. The user does the thinking.
Structured ideation — the problem is clear, the user wants options. Generate diverse approaches, explore tradeoffs. You amplify the user's creativity, not replace it. The user evaluates and selects.
The modes are fluid. Don't announce them. Follow the user's energy.
Diverge on ideas, converge on proposals. Generate breadth, but each idea is a concrete sketch — not a fragment.
Brainstorm is the explicit escape hatch from "commit to one path" — not from "be specific."
Offer a lightweight check-in: "Sounds like we're converging on X — want me to pull this together?"
If yes → transition to SAFEWORD.md's understanding flow (scope / out of scope / done when). If no → keep exploring.
| Don't | Do |
|---|---|
| Summarize unprompted | Let the user pull the trigger |
| Announce frameworks ("let me apply 5 Whys") | Just ask the questions the framework would produce |
| Jump to implementation unprompted | Stay in ideas until the user signals readiness |
| Evaluate during divergence | Generate first, evaluate after |
| Produce only safe ideas | Include at least one unconventional option |
まだレビューはありません。使ってみた感想をお寄せください。
概要と使いどころ
Run comprehensive code audit for architecture, dead code, and test quality. Use when reviewing overall codebase health, checking for architectural violations, or before marking a feature complete.
日本語の概要は準備中です。原文の説明を表示しています。
Behavior-first feature development — use when building new capabilities, continuing feature work, or when work introduces new state or multiple user flows. Discovers desired behavior through examples and scenarios before implementation. Do NOT use for bug fixes, typos, or small isolated changes.
日本語の概要は準備中です。原文の説明を表示しています。
Kill zombie dev servers and test processes. Use when ports are blocked, processes are hanging, or test runners won't start.
日本語の概要は準備中です。原文の説明を表示しています。
Root cause debugging before fixes. Use when investigating bugs, diagnosing test failures, troubleshooting unexpected behavior, or when previous fix attempts failed. Enforces investigate-first discipline.
日本語の概要は準備中です。原文の説明を表示しています。
Extract tacit knowledge through non-obvious microquestions — things only the user knows that can't be found in code, docs, or research. Use when you're about to guess at intent, context, or constraints during SAFEWORD's understanding flow. Also use when user says 'ask me', 'what do you need to know', or when another skill (bdd, brainstorm, debug) needs user context before proceeding. Do NOT use for questions answerable by reading the codebase or searching the web.
日本語の概要は準備中です。原文の説明を表示しています。
Explore and debate options with fresh documentation and research before committing. Use when facing a real decision with multiple plausible approaches — library/framework choice, architecture call, API or schema design, algorithm selection, or any communication / strategy call where being wrong has cost. Enumerates relevant research domains, looks up current docs and evidence-based methods, weighs options on correctness and elegance, resists bloat. Do NOT use for divergent ideation (brainstorm), extracting user intent (elicit), or reviewing already-written code (quality-review).
日本語の概要は準備中です。原文の説明を表示しています。