analyze
無料Risk analysis and approach selection before planning. Use when requirements are ambiguous, approaches vary, or work touches hub/bridge nodes. Skip for typo fixes.
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。
Build throwaway code to validate an uncertain approach before committing to a plan. Two branches — Logic (terminal state machine) or UI (multiple designs on one route with URL-param switcher). No persistence, no tests, single-command start. Use between analyze and plan when the approach itself is uncertain.
インストール方法を見る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。
Throwaway code to validate approach. Explore fast, decide, then delete or integrate.
Announce: "🧪 /supergraph:prototype — building throwaway prototype..."
/supergraph:analyze and /supergraph:plan when approach is uncertainAsk user (or infer from task):
Logic branch — validate state, algorithm, or data model
UI branch — validate layout, interaction, or multiple designs
Goal: validate core logic in a minimal terminal program.
Rules:
prototype-<slug>.ts / .py / etc.)npx ts-node prototype-<slug>.ts or equivalentStructure (TypeScript example):
// prototype-<slug>.ts — THROWAWAY, do not review
type State = "idle" | "processing" | "done" | "error"
type Event = { type: "start" } | { type: "finish" } | { type: "fail" }
function transition(state: State, event: Event): State {
// ...
}
// Run through sample cases
const cases = [ ... ]
cases.forEach(({ from, event, expected }) => {
const result = transition(from, event)
console.log(result === expected ? "✅" : "❌", { from, event, result, expected })
})
Run and observe:
npx ts-node prototype-<slug>.ts
Goal: compare multiple visual/interaction designs without committing.
Rules:
/prototype/<slug>?design=A|B|C switches between designsnpm run dev (or project equivalent)Structure:
app/prototype/<slug>/page.tsx (router + floating switcher)
app/prototype/<slug>/DesignA.tsx
app/prototype/<slug>/DesignB.tsx
app/prototype/<slug>/DesignC.tsx
Floating switcher (inline):
<div style={{position:'fixed',bottom:16,right:16,zIndex:9999,background:'#fff',padding:8,borderRadius:8,boxShadow:'0 2px 8px rgba(0,0,0,.2)'}}>
{["A","B","C"].map(d => (
<button key={d} onClick={() => router.push(`?design=${d}`)}
style={{margin:4, fontWeight: design===d ? 'bold' : 'normal'}}>
{d}
</button>
))}
</div>
Present findings to user:
Prototype: [slug]
Branch: Logic | UI
Findings:
- [What was validated]
- [What was ruled out]
- [Surprises / unexpected behavior]
Recommendation:
- Approach: [chosen approach]
- Confidence: high | medium | low
- Reason: [evidence from prototype]
Ask: "Ready to proceed to /supergraph:plan with this approach?"
After approach is decided:
# Delete prototype file(s)
rm prototype-<slug>.ts
# or git stash if user wants to keep for reference
Never commit prototype code to main branch.
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
Risk analysis and approach selection before planning. Use when requirements are ambiguous, approaches vary, or work touches hub/bridge nodes. Skip for typo fixes.
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。
Proactive architecture review — explore codebase structure, generate a self-contained HTML report with Mermaid diagrams and candidate improvements, then grill the findings. Use when planning a large refactor, onboarding to an unfamiliar codebase, or before a major architectural change.
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。
Persistent token-compression mode (~75% reduction) — now always-on by default. Strips filler while keeping code exact.
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。
Database migration best practices for schema changes, data migrations, rollbacks, and zero-downtime deployments across PostgreSQL, MySQL, and common ORMs (Prisma, Drizzle, Kysely, Django, TypeORM, golang-migrate).
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。
Anti-slop frontend skill for landing pages, portfolios, and redesigns. The agent reads the brief, infers the right design direction, and ships interfaces that do not look templated. Real design systems when applicable, audit-first on redesigns, strict pre-flight check.
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。
Structured 6-phase debugging. Build feedback loop first, reproduce deterministically, hypothesize with ranked falsifiable theories, instrument one variable at a time, fix with regression test, cleanup. Use when a bug exists, tests fail unexpectedly, or behavior is wrong and cause is unknown.
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。