What problem does it solve? Turning a written implementation plan into actual code often loses traceability: nobody records which legacy rules were touched, which behaviors must keep working, or what progress was made. This Skill executes an actions.md plan phase by phase and automatically produces audit artifacts (legacy-impact.md, regression-watch.md, progress.jsonl) so every change stays anchored to the extracted legacy or greenfield specification. ## Core Features & Use Cases - Phased plan execution: Runs pending [ ] actions from actions.md in order (Preparação, Testes, Núcleo, Integração, Polimento), respecting dependencies and parallel markers, and supports scoping to a phase or ID range. - Context anchoring: Requires an anchor in _reversa_sdd/ — either legacy extraction (architecture.md + domain.md) or greenfield artifacts (prd.md + SDD specs) — and aborts cleanly if none exists. - Audit artifact generation: Writes legacy-impact.md (affected components, impact type, severity) and regression-watch.md (stable watch IDs for modified or removed rules), plus append-only progress.jsonl entries per action. - Use Case: After running /reversa-to-do to produce an action plan for a feature, invoke this Skill to implement the tasks, then hand off to /reversa-sync to converge the delivery back into the extraction. ## Quick Start Run the reversa-coding skill to execute all pending actions in my feature's actions.md and generate the legacy impact and regression watch reports.
