reversa-writer

To generate formal design specifications from legacy knowledge, use a reverse engineering writer that reads extracted context and produces traceable requirements.md, design.md, and tasks.md files in a deterministic folder structure.

Documenting legacy code for AI reimplementation involves producing unit-by-unit operational specifications with formal requirements, design details, and tasks, ensuring the output is traceable and faithful to the original system.

Yes, generating reverse engineering documentation requires reading state.json, config.toml or config.user.toml, and surface.json first to determine output granularity and language rules before creating specification files.

Yes, you control the granularity and output format of generated engineering requirements through config.toml or config.user.toml, which routes the creation of canonical files, contracts, and edge case artifacts per unit.

Legacy documentation generation stops if specs granularity is not decided in the configuration files, applying guardrails to ensure formal specifications are only produced when output rules are explicitly configured.

Creating design specifications for a legacy system unit produces canonical files including requirements.md, design.md, and tasks.md, with optional artifacts for contracts, flows, edge cases, decisions, and questions based on the configured doc level.