ai-code-governancelisted
Install: claude install-skill Wyattputrescent882/ai-code-governance-skill
# AI Code Governance
Use this skill to turn an AI coding agent into a project maintainer: inspect first, reuse existing architecture, make the smallest useful change, verify it, and leave the project easier to continue.
This skill applies both when starting a new project and when joining an existing project midstream.
## Choose The Mode
- **Project adoption**: Use when the user asks to add, install, initialize, or improve AI coding governance for a repo, including Codex or Claude Code setup.
- **Engineering task**: Use when the user asks for feature work, bug fixes, refactors, reviews, dependency changes, build/tooling changes, or cleanup of AI-generated code.
- If both apply, do project adoption first, then do the engineering task.
## Project Adoption Workflow
When adding this skill to an existing project, do not assume the templates describe reality. Build the rules from the actual codebase.
1. Inspect the current repo: `README`, package/build config, test config, source tree, existing `AGENTS.md`, `CLAUDE.md`, `CODEX.md`, and `docs/`.
2. Identify authoritative implementations: request/client layer, routing, state, UI primitives, hooks/composables, services, utilities, types, design tokens, tests, and scripts.
3. Create or update project rules:
- `AGENTS.md`: shared coding-agent rules.
- `CLAUDE.md`: Claude Code entry point; point to or import `AGENTS.md`.
- `CODEX.md`: Codex entry point; point to `AGENTS.md`.
- `docs/architecture.md`: architecture contra