I built an AI-native codebase framework–could you evaluate it?

I built this open-source project and would really appreciate technical feedback from people here:

https://github.com/xodn348/ai-native

The goal is to make AI-assisted development more reliable through clearer project structure, explicit contracts, and verification workflow.

I made this because applying these patterns from scratch in every project was repetitive and hard to maintain, so I wanted a reusable framework.

If you have time, I’d love your evaluation on:

1. What is useful in practice?

2. What is unclear or unnecessary?

3. What should be improved first?

4. What evidence/tests would you want to see for stronger credibility?

If you find it useful, I’d appreciate a GitHub star as well. I’m open to direct criticism.