What it does
A local tool that checks code written by AI agents against the patterns found in your repository's git history. It flags foreign dependencies, reinvented functions, misplaced code and tests that look gamed. It uses statistics rather than a second language model and shows the divergences so you decide what to accept.
Use cases
- 01Flag dependencies the team does not use
- 02Catch functions that duplicate existing code
- 03Spot tests altered to pass without checking behavior