Monorepo Init

What It Does

Monorepo Init runs an interview, then writes a complete project scaffold in one session. No manual copy-pasting of boilerplate, no forgetting the memory files.

• Interviews you on project name, platforms, tech stack, branches, and first task before writing anything
• Writes PRD.md from your answers — overview, platform table, architecture, goals
• Writes CLAUDE.md with git workflow, behavioral guidelines, memory organization, and docs structure tailored to your platforms
• Creates root .memory/000-how-to-memory.md and one per platform folder
• Creates docs/README.md as the flat docs index with naming conventions
• Branches on superpowers: creates docs/SPEC_*.md (for superpowers users) or TASK.md (without) for your first task
• Idempotent — memory files are never overwritten; prompts before touching CLAUDE.md, PRD.md, TASK.md
• Ends with a verification script and the final git commit command

Output Structure

<project-root>/
├── CLAUDE.md
├── PRD.md                             ← written from interview
├── TASK.md                            ← only if NOT using superpowers
├── .memory/
│   └── 000-how-to-memory.md
├── docs/
│   ├── README.md
│   └── SPEC_<SCOPE>_<date>_<slug>.md ← only if using superpowers
└── <platform>/
    └── .memory/
        └── 000-how-to-memory.md

When to Use

• Starting a new monorepo with multiple platforms (API + frontend, mobile + backend, etc.)
• Any new project that will use Claude Code as the primary coding assistant
• When you want consistent CLAUDE.md, memory, and docs structure across projects without writing it from scratch each time
• Re-running on an existing project to fill in missing scaffold files without overwriting existing ones

How to Use

Invoke the skill at project creation. It will ask 6 questions, then write all files. Example prompts:

Invoke the skill directly with:

The skill runs in 3 phases: Plan (2–3 questions, smart defaults inferred from stack), Review (full scaffold plan shown, correct via free text), Generate (files + Hello World + tests, ≥ 80% coverage).

Design Docs

The full Monorepo Init specification — 3-phase flow, stack inference rules, Hello World scaffolds per stack, file templates, and idempotency rules — is split across companion files. Only files needed for the current phase load into context.

View skill source on GitHub →

Source & Changelog

Skill source: skills/monorepo-init/