spec-miner

To document an undocumented legacy codebase, you extract software specifications by mapping code dependencies, tracing data flows, and documenting observed behaviors in EARS format. This reverse engineers inherited projects into foundational architecture and API documentation.

The best way to reverse engineer specifications from existing code is to map module dependencies, trace data flows, and identify key APIs. Using Glob, Grep, and Read tools for codebase exploration accurately documents the observed behaviors.

Yes, you can generate EARS format documentation for an inherited project by analyzing the codebase to extract observed behaviors. This maps underlying data flows and dependencies directly into structured software specifications.

You trace data flows and map dependencies in a complex codebase by using exploration tools like Glob, Grep, and Read to analyze source code. This identifies key modules and reconstructs the architectural relationships.

Yes, codebase exploration works for identifying API endpoints in legacy systems by analyzing existing source code structure. It traces data flows and maps dependencies to reveal undocumented APIs and business logic.

You should not use automated code analysis for specification generation when the codebase lacks accessible source code or when precise runtime behavior cannot be statically determined. It works best for inherited projects requiring foundational reverse engineering.