What problem does it solve? When evolving a feature through a multi-stage pipeline (requirements, clarify, plan, to-do, coding), it is easy to lose track of which step comes next. This Skill inspects the actual artifacts on disk and tells you exactly where the active feature stands and which agent to invoke next, without ever writing feature artifacts itself. ## Core Features & Use Cases - Physical stage detection: Reads active-requirements.json and the feature directory to classify the stage as vazio, requirements, plan, coding-em-progresso, or done based on real files, not self-declared metadata. - Routing matrix: Maps each detected stage to the correct next skill (/reversa-requirements, /reversa-clarify, /reversa-plan, /reversa-to-do, /reversa-coding), including handling of [DÚVIDA] markers and paused features. - Greenfield and legacy support: Works whether or not a reverse-engineering extraction exists in _reversa_sdd/, creating empty folders when needed and persisting spec granularity decisions in .reversa/config.toml via atomic writes. - Use Case: You return to a project after a week, type /reversa-forward, and it reports that actions.md has 3 of 8 tasks checked off, then suggests /reversa-coding and waits for your CONTINUAR confirmation. ## Quick Start Type /reversa-forward in your AI agent to detect the current feature stage and get a suggestion for the next pipeline step.
