lizard

https://github.com/terryyin/lizard has been useful to track when functions get too convoluted or long, or when there's too much duplication -- in code generated by agents. Still have to see how well it works long term but it's caught things here and there, I have it in the build steps in my scripts so the agent sees its output.

I'm sure the top comment here will be something like "this is invalid because no way can you assign a numerical value to code quality! wtf?!" I'm withholding my own judgement on that. For anyone curious, the authors are coming up with a code quality score using an open-source tool called SoftWipe[0]. From the paper: > SoftWipe is an open source tool and benchmark to assess, rate, and review scientific software written in C or C++ with respect to coding standard adherence. The coding standard adherence is assessed using a set of static and dynamic code analysers such as Lizard (https://github.com/terryyin/lizard) or the Clang address sanitiser (https: //clang.llvm.org/). It returns a score between 0 (low adherence) and 10 (good adherence). In order to simplify our experimental setup, we excluded the compilation warnings, which require a di