reverse-document

To generate documentation from an undocumented codebase, analyze the existing implementation to extract mechanics, dependencies, and intent, then draft structured design PRDs or concept documents with user confirmation before saving.

Reverse-documentation for legacy code is the process of analyzing existing implementations to surface architectural decisions and trade-offs, generating structured ADRs or concept documents to provide post-facto clarity for systems lacking original specifications.

Yes, you can create architecture decision records from a prototype implementation by reading the code to collect trade-offs and mechanics, flagging unclear intent, and waiting for your approval before persisting the finalized ADRs to their target folders.

The best way to document combat logic after prototyping is to mirror the requested path to collect mechanics and dependencies, align the extracted findings with a design PRD template, and draft the document after confirming the elicited intent.

Yes, reverse-documentation works for core systems and APIs without external dependencies, reading the implementation directly to extract architecture decisions and intent before drafting template-driven concept documents or PRDs.

You should avoid automated code analysis for generating documentation when the implementation intent is completely obscured or unconfirmable, as the process requires collaborative checkpoints and user approval to ensure the drafted docs match the actual implementation.