reversa

Orchestrate phased reverse-engineering sessions with checkpoint-based state files.

Updated May 16, 2026
One-click install
npx skills add https://github.com/ricardops34/rcg-crm --skill reversa
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: reversa
Source: https://github.com/ricardops34/rcg-crm/tree/main/.agents/skills/reversa
Command: npx skills add https://github.com/ricardops34/rcg-crm --skill reversa

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

It coordinates a full reverse-engineering workflow that turns a legacy system into actionable, executable specifications for downstream AI agents.

Core Features & Use Cases

  • Plan orchestration: Runs the Reversa pipeline sequentially and advances phases with checkpoints.
  • State-driven execution: Reads .reversa/state.json to resume from where the process stopped, or to start first-run instructions.
  • User-gated documentation depth: Uses the Scout results to ask for the documentation level (Essencial/Completo/Detalhado) and persists that choice.
  • Dynamic plan update after Scout: Updates .reversa/plan.md by splitting a generic step into per-module tasks once surface mapping is available.
  • Guardrails and safety rules: Enforces non-destructive behavior (writes only to .reversa/, _reversa_sdd/, and regression-watch history) and checkpointing to prevent context loss.

Quick Start

Activate the reversa skill by asking your AI to run /reversa to start the end-to-end reverse-engineering and spec generation workflow for your legacy project.

Frequently Asked Questions about reversa

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate executable specs from a legacy codebase?

Reverse engineering legacy systems into executable specs requires orchestrating a phased workflow that sequentially maps codebases and outputs AI-ready specifications. This process coordinates sequential agents to analyze modules and generate actionable documentation.

Can I resume reverse engineering a legacy system after stopping?

Yes, you can resume reverse engineering a legacy system after stopping because the workflow reads a state.json file to checkpoint progress. This state-driven execution allows you to restart exactly where the process previously halted without losing context.

How do I control documentation depth when reverse engineering code?

You control documentation depth when reverse engineering code through a user-gated decision step that asks you to select an Essential, Complete, or Detailed level. This choice is persisted in configuration files to guide the final spec generation.

What is the best way to map legacy modules for AI agent execution?

The best way to map legacy modules for AI agent execution is to run an initial Scout phase that performs surface mapping, then dynamically updates the plan. This splits generic steps into per-module tasks to guide deeper sequential analysis.

Are there limitations on where reverse engineering files are written?

Limitations on where reverse engineering files are written enforce non-destructive behavior by restricting outputs to the .reversa directory and _reversa_sdd folders. These guardrails prevent the workflow from modifying your original source code.