I analyzed a lot of repositories when developing Valknut (https://github.com/sibyllinesoft/valknut) in order to eval the algorithms and catch bugs. Since people love leaderboards and I want to support well written open source software, I figured I'd make a little app so people can test their work and optionally show off their score for promotional purposes. TS/JS/Go/Python/Rust are currently supported.

If you disagree with scoring, I'm happy to have a conversation and do evals based on your feedback, but keep in mind that the code that is optimal for humans is not quite the same code that's optimal for agents (though there's a lot of crossover!).

If there's interest I'm happy to put together a Github badge to let you show off your ranking/score.