@swirly Yes; tools like these will never be 100% accurate. I am currently not sure how I'd approach analyzing the contents of a CLAUDE.md or AGENTS.md file to determine whether it is meant to attract or deter LLMs. But the existence of such a file is, in my opinion, a potential indicator for AI activity. As such, it gets reported; but if there aren't any other signals, you'll still have an A+ rating
P
polyphonychat@fed.amazonawaws.com
@polyphonychat@fed.amazonawaws.com
Posts
-
Are you tired of having to manually traverse commit histories to find out whether a repository has been made using AI/LLMs? -
Are you tired of having to manually traverse commit histories to find out whether a repository has been made using AI/LLMs? -
Are you tired of having to manually traverse commit histories to find out whether a repository has been made using AI/LLMs? -
Are you tired of having to manually traverse commit histories to find out whether a repository has been made using AI/LLMs?Are you tired of having to manually traverse commit histories to find out whether a repository has been made using AI/LLMs? There's now an easier way!
You can access the first version ofrepo-slopscoreat slopscan.ava.pet, paste inhttps://git repo URLs and have it scan repositories for signs of AI activity. The source code is, of course, available at codeberg.org/polyphony/repo-slopscore !
The prototype is most definitely rough around the edges, but also functional! You can read more about how the internals of it work in theREADME.mdin our repository!
RE: fed.amazonawaws.com/notes/alrwb40eqhsxkxpb
