本文へ移動
cccskills
無料GitHub で公開

discovery

Phase 0 validation gate — problem framing, competitive analysis, feasibility, and a go/no-go checkpoint before committing to a feature. Use when evaluating whether an idea is worth building, before /story or /design. Trigger on "is this worth building", "validate this idea", "discovery", "should we build".

インストール方法を見る

含まれるファイル(10)

  • SKILL.md5.8 KB
  • templates/checkpoint/patterns.md5.4 KB
  • templates/checkpoint/summary.md1.9 KB
  • templates/competitive/patterns.md3.3 KB
  • templates/competitive/summary.md1.3 KB
  • templates/feasibility/patterns.md4.0 KB
  • templates/feasibility/summary.md1.7 KB
  • templates/overview.md1.1 KB
  • templates/problem/patterns.md2.9 KB
  • templates/problem/summary.md1.3 KB

SKILL.md(原文)

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。

Discovery — Phase 0 Validation Gate

Structured Discovery & Validation before committing to a feature. Prevents building unvalidated ideas by gating entry into the story and design pipeline with four sequential modes.

Usage

/discovery                   # Show overview of all 4 modes
/discovery problem           # Problem framing: who has it, how urgent, what evidence
/discovery competitive       # Competitive landscape: what exists, differentiators, moat
/discovery feasibility       # Feasibility: technical, economic, time constraints
/discovery checkpoint        # Go/no-go gate: 5-criteria verdict before /story create

Workflow

Step 1: Determine Mode

Parse $ARGUMENTS for explicit mode. If no arguments provided, show overview:

no args         → read templates/overview.md
problem         → read templates/problem/summary.md
competitive     → read templates/competitive/summary.md
feasibility     → read templates/feasibility/summary.md
checkpoint      → read templates/checkpoint/summary.md
<anything else> → show usage error + list valid modes + read templates/overview.md

Step 2: Present Guidance (Progressive Disclosure)

Level 1 — Summary (default, ~20 lines): Read and present the summary template. This is the entry point — frames the questions to answer.

Level 2 — Patterns (on request, ~60 lines): When the user asks for more depth, "show me patterns", "how do I score this", or "give me examples" — read the patterns template.

Level 3 — Full Analysis (on request): No template — dive into specific questions with the user based on context provided.

Step 3: Run Modes Sequentially

Recommended sequence: problem → competitive → feasibility → checkpoint

Each mode builds on the previous:

  • problem → frames who, what, and why
  • competitive → validates market gap
  • feasibility → validates buildability
  • checkpoint → synthesizes all three into a go/no-go verdict

You can run individual modes out of order when the user has already done partial research.

Step 4: Checkpoint Verdict

After the checkpoint mode:

  • Go → "Run /story create to define the feature scope."
  • Conditional go → Flag caveats, suggest mitigations, then allow forward.
  • No-go → Explain which criteria failed, suggest what needs to change before re-evaluating.

Template Locations

Summary Templates (~20 lines each — L1 default)

ModeFile
overview~/.claude/skills/discovery/templates/overview.md
problem~/.claude/skills/discovery/templates/problem/summary.md
competitive~/.claude/skills/discovery/templates/competitive/summary.md
feasibility~/.claude/skills/discovery/templates/feasibility/summary.md
checkpoint~/.claude/skills/discovery/templates/checkpoint/summary.md

Patterns Templates (~60 lines each — L2 on request)

ModeFile
problem~/.claude/skills/discovery/templates/problem/patterns.md
competitive~/.claude/skills/discovery/templates/competitive/patterns.md
feasibility~/.claude/skills/discovery/templates/feasibility/patterns.md
checkpoint~/.claude/skills/discovery/templates/checkpoint/patterns.md

Progressive Disclosure Rules

  1. Always start at L1 — never present all 4 modes' full patterns at once
  2. Read templates on demand — use the Read tool; don't memorize template content
  3. User drives depth — "show me how to score", "give me the rubric", "examples?" triggers L2
  4. Don't hard-block the user — if they skip modes, accept partial input but flag explicitly which criteria are scored on assumptions rather than evidence
  5. Checkpoint is the exit gate — nothing moves to /story create without a checkpoint verdict

Integration

Position in pipeline: Phase 0 → before everything else

/discovery (Phase 0)
  → /story create (Phase 1 — only after checkpoint go)
    → /design (Phase 2)
      → /implementation (Phase 3)

Invoked by: User directly (/discovery) Invokes: Nothing — produces validated problem definition consumed by /story create Connected skills:

  • /story create — next step after checkpoint go
  • /design — invoked by user after /story prp generates a PRP; uses the validated problem scope from discovery
  • /product-assess — deeper product assessment for complex ideas (optional, before /discovery checkpoint)

Examples

Full sequence on a new idea:

User: /discovery problem → frames who/urgency/evidence User: /discovery competitive → maps alternatives User: /discovery feasibility → checks buildability User: /discovery checkpoint → verdict: Go → "Great, run /story create to define the feature."

Skipping straight to checkpoint:

User: /discovery checkpoint → Present checkpoint summary, fill criteria based on what user already knows → Flag any missing evidence as risks

L2 escalation:

User: /discovery problem → Show problem/summary.md User: "show me how to score urgency" → Read and show problem/patterns.md (pain-point scoring section)

レビュー

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。

同じリポジトリのスキル

概要と使いどころ

Validate WCAG 2.1 Level AA compliance and accessibility best practices. Use when performing accessibility audits and WCAG certification.

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。

matteocervelli/llms252026年5月22日 更新

analysis

無料

Analyze feature requirements, dependencies, and security considerations. Use when starting feature implementation from GitHub issues to understand scope, technical feasibility, and risks.

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。

matteocervelli/llms252026年5月22日 更新

analytics

無料

Run SQL queries against psql, BigQuery, or MySQL from the terminal, including natural-language-to-SQL and schema exploration. Use when analyzing data, inspecting DB state, or debugging tables. Trigger on "query the database", "SQL", "show me data from", "explore table".

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。

matteocervelli/llms252026年5月22日 更新

Design REST APIs or function contracts with clear request/response specifications, error handling patterns, authentication strategies, and comprehensive documentation.

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。

matteocervelli/llms252026年5月22日 更新

Generate comprehensive API endpoint tests for REST and GraphQL APIs. Creates tests for all HTTP methods, status codes, authentication, and validation.

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。

matteocervelli/llms252026年5月22日 更新

Design component architecture and module structure using established architectural patterns for clean, maintainable, and scalable systems.

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。

matteocervelli/llms252026年5月22日 更新

matteocervelli のスキルをすべて見る

このスキルの問題を報告する