I wrote this, where possible I use non llm tooling to assess and llm just writes up the findings.

https://github.com/bjcoombs/ai-native-toolkit/tree/main#what...