bmad-document-project

To generate AI-ready documentation for a brownfield codebase, you run a project scanning workflow that analyzes your source tree and outputs Markdown artifacts like architecture docs and master indexes. It turns messy existing code into clear, AI-consumable context.

The best way to document API contracts and data models for AI context is using an exhaustive deep-dive scan. This process detects project requirements and produces validated Markdown outputs, ensuring AI agents can understand your architecture without guessing.

Yes, you can resume a brownfield documentation workflow if project scanning is interrupted. The process supports state tracking and resume prompts, allowing you to pause and continue fresh scans, full rescans, or deep-dive runs without losing progress.

Yes, brownfield project documentation supports configurable scan depth. You can resolve runtime configuration and workflow customization to select your scan mode, controlling whether you perform a quick initial scan or an exhaustive deep-dive documentation run.

When mapping architecture for an existing repository, the generated outputs include a master index, source-tree analysis, architecture docs, and development or deployment guides. It optionally produces API contracts and data models depending on detected project requirements.

You should not use an automated project scanner for brownfield documentation if your repository lacks persistent facts or requires manual architectural interpretation. The workflow relies on resolving runtime config and loading workflow customization to generate accurate AI context.