spec-miner

Reverse-engineer system specifications from legacy code by mapping entry points, routes, data models, and business logic to produce documented requirements. This approach grounds the extracted specifications in actual code behavior, reducing guesswork when original documentation is missing or outdated.

The EARS format is a structured requirements notation that categorizes extracted specifications into Ubiquitous, Event-driven, State-driven, and Optional types. Applying this format to reverse-engineered legacy code yields clear, actionable requirements suitable for implementation planning and modernization alignment.

Document undocumented features in a legacy codebase by tracing data flows and request paths from entry points through downstream logic. This maps end-to-end system behaviors into grounded observations, distinguishing factual code evidence from inferences and flagging uncertainties for later clarification.

Yes, you can generate architecture or API documentation from implementation details by identifying routes and controllers, tracing downstream logic, and mapping data models. This produces structured documentation that reflects the actual system behavior rather than missing or outdated original specifications.

Spec mining works for modernization planning of legacy systems by extracting evidence-grounded specifications that reveal actual system behavior. It maps end-to-end flows and outputs EARS-format requirements, providing the clear baseline needed to plan migrations and align new implementations with existing functionality.