reversa-reconstructor

Generates a bottom-up reconstruction plan from Reversa specs and executes tasks one at a time.

Updated Aug 14, 2026
One-click install
npx skills add https://github.com/Ruggery/OCR_Cartorio --skill reversa-reconstructor-ruggery
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: reversa-reconstructor
Source: https://github.com/Ruggery/OCR_Cartorio/tree/main/.agents/skills/reversa-reconstructor
Command: npx skills add https://github.com/Ruggery/OCR_Cartorio --skill reversa-reconstructor-ruggery

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Reimplementing a software system from specification documents is error-prone and token-expensive when done ad hoc. This Skill turns the specification artifacts produced by the Reversa framework into an executable, resumable reconstruction plan and implements each task on demand, reading only the files each task needs. ## Core Features & Use Cases - Plan Generation: Reads a minimal set of spec files (architecture, dependencies, gaps, confidence report) and produces _reversa_sdd/reconstruction-plan.md with tasks ordered bottom-up: database schema, domain entities, state machines, leaf units, API layer, and user flows. - Migration-Aware Reconstruction: Detects completed migrations via _reversa_sdd/migration/handoff.md and lets the user choose between rebuilding from original specs or from target-stack migration specs, including cutover and parity-validation tasks. - Token-Preserving Execution: Executes one task at a time, reading only the files listed in each task's Lê: field, marking status in the plan, and pausing for user decisions on spec gaps. - Use Case: After running Reversa on a legacy codebase, copy _reversa_sdd/ into a fresh directory and invoke the Skill to reimplement the system task by task, pausing and resuming at any point. ## Quick Start Ask the agent to run /reversa-reconstructor in a directory containing the _reversa_sdd folder to generate the reconstruction plan and start executing tasks.

Frequently Asked Questions about reversa-reconstructor

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

FAQPage Schema
How do I rebuild a system from Reversa specification documents?▼

Place the _reversa_sdd/ folder in your working directory and invoke /reversa-reconstructor. It reads architecture, dependencies, and gap reports to generate reconstruction-plan.md, then executes tasks one at a time when you say INICIAR or CONTINUAR.

How does the reconstruction plan order implementation tasks?▼

Tasks follow a bottom-up order derived from dependencies.md: database schema first, then domain entities, state machines, leaf units, intermediate units, the API layer, and finally user flows. Infrastructure always precedes domain logic.

Can I reconstruct from migration specs instead of original specs?▼

Yes. If _reversa_sdd/migration/handoff.md exists, the Skill asks whether to rebuild from original specs or migration specs. Migration mode adds tasks for target schema, data migration scripts, cutover, and parity validation.

What happens if a migration is still in progress?▼

If _reversa_sdd/migration/.state.json exists but handoff.md does not, the Skill warns that the migration is unfinished and proceeds with the original specs only after your confirmation. You can decline to stop the process.

Does the reconstructor read all specification files at once?▼

No. Planning reads only a small set of files like architecture.md and dependencies.md. Each execution task reads only the files listed in its own Lê field, which preserves context tokens and allows pausing and resuming.

What happens when a spec has gaps during implementation?▼

When a gap marked as a lacuna is found, the Skill pauses execution and asks the user for a decision before continuing. It never invents behavior that is not documented in the specifications.