Everyone is looking through ClaudeCode's leak, I've been working on a better way to represent codebases to observe how my agents are working on them.

Wanted to dive and see what ClaudeCode is all about, so visualized it. You can run it locally from our fork: https://github.com/CodeBoarding/claude-code

We are open-core: https://github.com/CodeBoarding/CodeBoarding

Interested to hear what is something you didn't expect, for my 2 minute check I saw there are Computer-Use capabilities baked into it, however I've never observed that behavior in my usage. Probably not enabled by default.