Commencer
Japanese Spec Kit — Idea to 5 Definition Docs

Japanese Spec Kit — Idea to 5 Definition Docs

Japanese-language skill for startup CTOs, PMs, and contract developers. Drop a one-line app idea — get five aligned definition docs in Japanese: 要求定義書 (Requirements), 要件定義書 (Functional Spec), 技術定義書 (Technical Design), 工程定義書 (Project Plan), 品質定義書 (Quality Plan). Each doc is internally consistent. Mermaid diagrams embedded automatically. Pre-tuned for Cloud Run / Next.js / Python stacks. Output is Markdown, ready to paste into Notion, GitHub, or Cursor. Cuts a week of spec work to under an hour.
#Ingénierie#Produit#Productivité
Note
Davantage d’évaluations nécessaires
Vendus
2
Mode d’utilisation
Exécuter sur Capafy
Aussi sur des applications externes
Fourni par l’éditeur
Claude Sonnet 4.6

What this skill does

Drop a one-line app idea in Japanese and get five aligned definition documents — written in Japanese, structured for a Japanese development team, ready to hand off.

The skill runs as a pipeline:

  1. You provide: a one-sentence app idea (e.g. "中小企業向けのSaaS契約管理ツールを作りたい")
  2. The skill asks: 3–5 targeted clarifying questions (target users, scale, integration constraints)
  3. You receive: 5 internally-consistent Markdown documents

The 5 documents

Document Japanese name What's inside
Requirements 要求定義書 User personas, jobs-to-be-done, business goals, success metrics
Functional Spec 要件定義書 Feature list, user flows, screen specs, acceptance criteria
Technical Design 技術定義書 Architecture, stack rationale, data model, API design, Mermaid diagrams
Project Plan 工程定義書 Milestones, dependencies, estimated timeline, team composition
Quality Plan 品質定義書 Test strategy, KPIs, monitoring, release criteria

Each document references the others — no rewriting the same assumptions five times.

Built-in stack assumptions

Pre-tuned for the stack that dominates Japanese product startups:

  • Frontend: Next.js (App Router) + TypeScript
  • Backend: Python (FastAPI) or TypeScript (Bun / Hono)
  • Infra: Google Cloud Run, Cloud SQL, Cloud Storage
  • Auth: Firebase Auth or Auth.js
  • AI/Agent layer: Anthropic Claude API, optional MCP servers

You can override any of these in clarifying questions — the skill adapts.

Who this is for

  • Startup CTOs drafting specs before their first engineering hire
  • Product managers translating ideas into developer-ready specs
  • Contract developers (受託開発) needing a fast first-draft spec for client review
  • In-house DX teams at non-tech companies starting an internal tool project
  • Consultants producing upstream documentation deliverables

What you get

  • 5 Markdown files (one per definition document)
  • Mermaid diagrams embedded inline (renders in Notion, GitHub, GitLab, VS Code)
  • Internal cross-references between documents
  • Ready to paste into Notion / GitHub / Cursor

What this skill is NOT

  • Not a code generator — it produces specs, not working code
  • Not for non-Japanese projects — all output is in Japanese
  • Not a one-shot prompt — it asks clarifying questions to ground the output in your actual context
  • Not a substitute for product discovery — it accelerates documentation, not customer research

FAQ

Q: Can I edit the output before using it?
A: Yes, output is plain Markdown — edit freely in any text editor.

Q: What if my stack isn't Next.js / Python?
A: Tell the skill in the clarifying-question phase. It will adapt to Go, Rails, Java, etc.

Q: How long does generation take?
A: Typically 5–10 minutes end-to-end including clarification turns.

Q: Is the output good enough to send to a client?
A: It's a strong first draft. Review and edit before client delivery — but expect to save 80%+ of the writing time.