reversa-detective

To extract business logic from legacy code, you analyze Git history and code artifacts to reconstruct domain rules, state transitions, and decision rationale. This process identifies implicit business constraints from conditionals, validations, and enums to produce retroactive ADRs and domain glossaries.

Yes, you can generate retroactive ADRs from Git commit history by performing Git archaeology. This analyzes commit logs, reverts, and refactors to reconstruct past decision records and rationale, outputting traceable ADR markdown files with confidence levels.

Inferring state machines and RBAC permissions from an existing codebase requires analyzing code conditionals and access controls. The process derives valid entity status transitions and RBAC/ACL matrices when those concepts exist, outputting state-machines.md and permissions.md files.

To start reverse engineering a legacy codebase for domain knowledge, you need to read the .reversa/state.json configuration file for output settings and consume existing Scout or Archaeologist artifacts from the output directory and .reversa/context before generating documentation.

The doc_level parameter controls reverse engineering documentation outputs by determining the depth of generated files. It mandates a domain.md file while optionally triggering the generation of state-machines.md, permissions.md, and retroactive ADRs based on the desired documentation level.

Reconstructing decision rationale from missing requirements has limitations because it relies on implicit clues like TODO comments and code intent. If domain concepts like state machines or RBAC matrices do not exist in the codebase, those specific inference outputs cannot be generated.