reversa-coding

Execute actions.md into legacy-code changes with phase-aware dependency handling.

1.5k|383|Updated Apr 26, 2026
One-click install
npx skills add https://github.com/sandeco/reversa --skill reversa-coding-sandeco
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: reversa-coding
Source: https://github.com/sandeco/reversa/tree/main/agents/reversa-coding
Command: npx skills add https://github.com/sandeco/reversa --skill reversa-coding-sandeco

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill converts a legacy project’s plan of actions into real code changes while producing traceable artifacts for later auditing and regression checking.

Core Features & Use Cases

  • Executes actions.md phase-by-phase: Transforms all pending actions into implementation work, respecting declared dependencies and parallelizable subsets.
  • Keeps forward progress accountable: Updates actions.md checkboxes and appends execution records to progress.jsonl.
  • Produces audit and regression trails: Generates legacy-impact.md (what the legacy was changed) and regression-watch.md (what must stay true in future extractions).

Quick Start

Run the reversa-coding skill when you want the system to execute the active plan and start coding the feature using the prepared actions.md.

Frequently Asked Questions about reversa-coding

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

FAQPage Schema
How do I generate legacy code changes from a spec while maintaining an audit trail?

Legacy code changes are generated by executing actions.md phase-by-phase, which transforms planned actions into implementation work while producing an append-only progress log for auditability.

How do I automate regression testing for legacy modernization projects?

Regression testing is automated by generating a regression-watch.md file tied to domain rules, specifying what must remain true during future extractions and forward-cycle coding of active features.

Do I need a domain context file before starting phase-aware code generation?

Yes, phase-aware code generation requires a pre-existing domain.md file within the _reversa_sdd/ directory, which provides the necessary domain rules and context extracted before forward-cycle coding begins.

What's the best way to track dependencies when scaffolding project features from an action plan?

Dependency tracking is handled through phase-aware execution that respects declared dependencies and parallelizable subsets, updating actions.md checkboxes and logging execution records to progress.jsonl.

How does safe code generation handle impact analysis for legacy modernization?

Impact analysis is handled by generating a legacy-impact.md file that documents exactly what legacy code was changed during execution, ensuring all modifications remain traceable for later auditing.

Can I execute parallelizable coding tasks without losing forward progress accountability?

Yes, parallelizable subsets are executed while maintaining forward progress accountability through strict preflight checks and an append-only progress.jsonl log that records every execution step.