bmad-document-project

The best way to document a brownfield project for AI context is to automate the generation of an index, source-tree analysis, architecture docs, data models, API contracts, and a JSON metadata file describing the system parts.

Yes, you can automatically generate API contracts and data models for client/server monoliths by running conventions and runtime workflow steps that analyze the codebase and produce per-part deep-dive documentation.

To start automating project documentation, invoke the skill on a brownfield project root to trigger the activation flow, which loads configuration, determines project parts, and outputs comprehensive architecture artifacts.

Automated brownfield documentation supports both client/server monoliths and multi-part systems, providing per-part documentation and deep-dive outputs including architecture, data models, and API contracts for each identified component.