What problem does it solve? New repositories often lack the documentation layer AI agents and human collaborators need to work effectively, forcing every session to rediscover architecture, interfaces, and data models from scratch. ## Core Features & Use Cases - Architecture Panorama: Creates docs/architecture.md with three Mermaid diagrams covering layered architecture, internal dependency graphs with cycle highlighting, and a tri-class partition of internal artifacts, middleware, and external services. - Interface and Data Model Documentation: Produces docs/api-list.md grouped by module with External/Internal tagging, plus docs/data-model.md and a matching ER diagram as docs/data-model-er.svg grounded in the actual persistence layer. - Reconciliation and Agent Entry Point: Diffs claims across generated docs until no contradictions remain, writes a concise root CLAUDE.md with Guardrails and Debt placeholder sections, and scaffolds a read-only docs-auto-sync drift-check skill. - Use Case: Point it at an unfamiliar monorepo and receive a complete documentation kit—architecture diagrams, API catalog, ER diagram, and CLAUDE.md—plus a summary of judgment calls needing human confirmation. ## Quick Start Ask the agent to bootstrap the full AI collaboration documentation infrastructure for this repository, including architecture diagrams, API catalog, data model, and CLAUDE.md.
