agent-friendly-codebase

Audit a codebase in any language or framework and restructure it so many AI agents can build, test and verify changes in parallel without colliding. Puts every feature in its own folder (boundaries chosen from the code and its git history), writes a test suite first and keeps it green through the move, and adds an AGENTS.md contract, a separate FEATURE_MAP.md, a FEATURE.md per feature, one check command with scoped and fix variants, formatting after every agent edit, keyless local runs, CI, and a repo-specific code-review skill with AI reviewer config. Fire and forget: audits, shows one plan, asks one yes or no (plus which AI reviewer, if none is set up), then does the rest. Use when asked to make a repo agent-friendly or agent-ready, compartmentalise features, set up for parallel agents, or audit agent readiness.