spec-miner

Specification extraction from legacy code uses automated code analysis to identify undocumented functionality by examining actual source files, collecting evidence of behavior, and producing structured EARS-formatted requirements. This grounds observations directly in code locations rather than assumptions.

Yes. Reverse engineering uses code analysis tools like grep and glob to search existing codebases, extract functional patterns, and document them with evidence-based specifications. The output is a reverse-spec document mapping features to their implementation locations.

Generate a reverse-spec document from the codebase itself. Code analysis identifies key components, functions, and dependencies, then structures them as EARS-formatted requirements with code references, accelerating understanding without manual code review.

Code evidence documentation involves analyzing the codebase with read, grep, and glob operations to locate functionality, then producing EARS-formatted requirements in a consistent template. All observations reference specific code locations to ensure accuracy.

Yes. By extracting specifications from the existing codebase with code analysis, you map current functionality, dependencies, and behavior patterns. This evidence-based specification reveals integration points and constraints for planning new features without modifying untested systems.

Code analysis captures implemented behavior but may miss undocumented edge cases, implicit assumptions, or design intent not visible in source. Always document uncertainties discovered during analysis and validate extracted specifications against actual system behavior.