bmad-document-project

To document an existing codebase for AI context, run a workflow that scans your repository structure, analyzes architecture, and generates AI-ready documentation artifacts directly to disk for agent consumption.

Brownfield analysis scans an existing, unfamiliar codebase to capture its structure, architecture, and integration details. You need it when preparing AI context for downstream planning tasks like generating brownfield PRDs.

Yes, you can control codebase scanning depth using quick, deep, and exhaustive scan modes. These options balance analysis speed against the level of architectural detail captured for AI context generation.

If the documentation workflow is interrupted, it resumes safely. The process writes and updates a state file during execution, allowing incremental updates to resume exactly where the previous run stopped.

Yes, the project documentation process validates generated files. It performs validation after each document is written immediately to disk, ensuring output quality before proceeding to the next artifact.

The best way to prepare repository documentation for planning is to generate a comprehensive documentation set including source-tree analysis and architecture docs. This provides the necessary AI context for downstream planning tasks like brownfield PRDs.