bmad-document-project

To generate AI-ready documentation for a brownfield codebase, run the document-project workflow to scan the repository and output artifacts like project overview, architecture, source-tree analysis, API contracts, and data models. It supports quick, deep, and exhaustive scan depths.

Documenting a large multi-part repository for AI context is handled by scanning each part independently and integrating the results. The workflow supports multi-part project integration and uses a project-knowledge store to link the generated documentation artifacts.

Yes, you can resume an interrupted codebase documentation scan. The workflow uses a project-knowledge store and YAML frontmatter in SKILL.md to enable resumable runs, allowing you to continue generating documentation artifacts without starting over.

Scanning an existing project generates specific documentation artifacts including a project overview, architecture details, source-tree analysis, API contracts, and data models. These template-driven outputs provide consistent and auditable results for AI context.

No, this workflow requires no external dependencies to analyze source code structure. It relies on internal mechanisms like YAML frontmatter, a project-knowledge store, and template-driven outputs to detect project type and structure directly.

You should use exhaustive scans instead of deep scans when your project documentation requires maximum detail for complex AI context. The workflow offers quick, deep, and exhaustive scan depths to balance processing time with the thoroughness of the generated architecture and data model artifacts.