reversa-writer

Generates executable specifications from legacy code as requirements, design, and tasks documents.

Updated May 22, 2026
One-click install
npx skills add https://github.com/GleisonOliveira/dockerpilot-mcp --skill reversa-writer-gleisonoliveira
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: reversa-writer
Source: https://github.com/GleisonOliveira/dockerpilot-mcp/tree/main/.agents/skills/reversa-writer
Command: npx skills add https://github.com/GleisonOliveira/dockerpilot-mcp --skill reversa-writer-gleisonoliveira

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Reverse-engineering a legacy system often ends with scattered notes instead of actionable documentation. This Skill turns extracted knowledge about a legacy codebase into formal, traceable specifications organized as one folder per unit, each containing requirements.md, design.md, and tasks.md files detailed enough for an AI agent to reimplement the functionality. ## Core Features & Use Cases - Folder-per-unit spec generation: Creates requirements.md, design.md, and tasks.md per module, endpoint, use case, feature, or custom granularity defined in config.toml. - Confidence-tagged traceability: Marks every statement as confirmed, inferred, or a gap, and maps legacy files to spec units in a code-spec matrix. - Incremental, checkpointed workflow: Generates one file at a time with user confirmation, saves progress to state.json, and supports pause/resume across sessions. - Use Case: After analyzing a legacy monolith, run this Skill to produce per-module specification folders plus OpenAPI definitions, user stories, and a traceability matrix for a planned rewrite. ## Quick Start Ask the agent to generate the specification plan for the legacy project using the reversa writer, then type CONTINUAR to produce each file step by step.

Frequently Asked Questions about reversa-writer

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

FAQPage Schema
How do I generate specifications from a legacy codebase?▼

Run this Skill during the generation phase of a reverse engineering analysis. It reads prior analysis artifacts from .reversa/state.json and surface.json, proposes a file-by-file plan, and creates requirements.md, design.md, and tasks.md per unit after you confirm each step.

What is folder-per-unit spec granularity?▼

Granularity defines what a unit is: a module, endpoint, use case, feature, hybrid module-with-use-cases, or custom folder. It is configured in the [specs] section of config.toml, and each unit gets its own folder with the three canonical spec files.

Can I resume spec generation after clearing the session?▼

Yes. Progress is saved to the redator_progress field in .reversa/state.json after every file. After three or more units, the Skill suggests pausing with /clear and resuming via /reversa in a fresh session without losing work.

Does it overwrite existing specification files?▼

No. The Skill follows a non-destructive directive: it never deletes, moves, or modifies existing files in the output folder. Pre-existing unit folders receive only missing files, and existing canonical files are left untouched.

What do the confidence markers in generated specs mean?▼

Every statement is tagged as confirmed (green, verified in code), inferred (yellow, reasoned from evidence), or a gap (red, requiring human validation). This ensures traceability and prevents fabricated requirements from entering the specification.