I open sourced it too.

It has a bit to go—like automatically monitoring reasoning traces and security—but I thought this would be easier to use than trying to maintain your own skills.md collection if the server handled more of it for you.

I’d be curious if anyone thinks I should keep working on it. I’d love someone to collaborate with on it!

The tradeoff is that the prompt is non-deterministic, but also more flexible. It can handle unknown situations. The generated code executes much faster and reliable for known situations.

If auto-learn would reduce the non-deterministic nature, i.e. chance of failing, that would speed up software development.

Even if the agent has mcp tools avaible, you can't guarantee that it would necessarily use them right? I imagine that you'd have a lot of new tools that do nearly the same things.

Also how does this scale? If the agent has like 300 tools.