# iamnotacoder: AI code generation and improvement

iamnotacoder is a toolkit of LLM-assisted scripts that discovers Python repositories on GitHub, then improves their code style, fixes issues and adds tests, integrating with GitHub to fork and commit the results.

The toolkit has three main pieces. A scraper discovers and filters Python repositories on GitHub by criteria such as code quality, size and recency. An optimizer combines static analysis with LLM-driven refactoring and automated testing to improve an existing repository's style, catch potential bugs, security issues and performance problems, and increase test coverage, then integrates with GitHub, including safe forking, to apply the changes. An orchestrator runs the optimizer across every repository the scraper found.

A bonus script generates a basic application from a natural-language description using a team of specialized LLM roles, useful for prototyping. The project describes itself as experimental, LLM-assisted tooling whose output should always be reviewed before use.

It is written in Python and released under the MIT license.

Website: https://github.com/fabriziosalmi/iamnotacoder