bmad-document-project

To generate documentation for an existing brownfield codebase, run an initial project scan to analyze the source tree and produce structured markdown artifacts like index.md and architecture.md for immediate developer and AI consumption.

Yes, you can pause and resume a deep-dive repository scan by leveraging stateful workflow tracking. The process maintains a project-scan-report.json file, allowing batched file processing to be safely stopped and resumed across quick, deep, or exhaustive modes.

The best way to extract API contracts from a multi-part repository is to perform an exhaustive scan that supports batched file reading and dependency graph tracing, outputting validated machine-readable api-contracts.md files for architecture review.

Yes, this brownfield project documentation tool supports both monoliths and multi-part repositories. It generates per-part architecture files and exhaustive per-file deep-dive reports tailored to the specific structure of your codebase during full rescans.

When scanning a source tree for architecture summaries, the outputs are validated markdown files including index.md, architecture.md, api-contracts.md, and specific deep-dive-*.md reports, generated using provided templates and checklist-driven validation.

You need an exhaustive deep-dive analysis for a brownfield PRD when planning new features on top of an existing codebase, as it traces data flows and dependencies to produce actionable insights and per-part architecture docs for feature planning.