architecture-mapper

To document legacy codebase architecture, you map unfamiliar repositories into a cross-linked knowledge base by reverse engineering module boundaries, boot paths, and data flows. This process generates Markdown artifacts with file-level citations, confidence labels, and clickable diagrams.

Reverse engineering module boundaries in a polyrepo system involves read-only analysis of application code to extract entry points, entities, jobs, and external integrations. It yields cross-linked documentation with stable anchors and backlinks for easy navigation.

Generating cross-linked architecture documentation with confidence labels requires applying evidence-first analysis to source code. Findings are categorized as confirmed, likely, or speculative, ensuring all architectural claims remain grounded in file-level citations.

Yes, you can trace end-to-end data flow for debugging without modifying the codebase through disciplined read-only analysis. This mapping process identifies how the system boots and moves data, outputting drillable system flowcharts for planning or debugging.

Refreshing stale architecture docs after codebase changes involves remapping the updated source code to produce new Markdown artifacts. This post-change analysis captures current module boundaries and external integrations, replacing outdated documentation with evidence-backed maps.