bmad-document-project

To document a brownfield project for AI context, you need to perform deep code analysis to generate structured artifacts covering architecture, source tree, APIs, and development workflows. This process automates transforming complex codebases into understandable documentation for AI agents and human developers.

The best way to generate architecture and API documentation for a legacy codebase is to use an automated code analysis tool that scans the project structure, technology stack, and dependencies. This produces detailed reports on API contracts and architecture specifically formatted for AI consumption.

Yes, you can run a targeted deep-dive analysis into specific areas of your source tree. The documentation process supports initial scans, full rescans, and focused deep-dives to generate comprehensive reports on particular project sections without rescanning the entire codebase.

Automated project documentation works effectively for onboarding new developers to complex software by generating up-to-date guides that explain project architecture, key components, and getting started workflows. It transforms intricate brownfield codebases into easily understandable development references.

You do not need to manually provide technology stack or dependency information before scanning a codebase. The automated analysis process identifies the project structure, technology stack, architecture, and dependencies directly from the existing source code during the initial scan.

Your AI agent struggles to understand existing codebases without structured documentation because raw code lacks the architectural context and workflow explanations AI requires. Generating structured documentation artifacts like source tree analysis and API contracts provides the necessary AI context for accurate interaction.