Someone sees a help wanted issue, pastes it into an AI, submits without testing, loops through review comments without understanding the code. The PR looks plausible at first glance but falls apart under review. Maintainers waste 30 minutes before realizing it's garbage.

This is happening at scale now. And it's worst in projects with bounty programs or GSoC where there's incentive to “contribute.”

GitHub tells you if someone is a first-time contributor to your repo. It doesn't tell you anything about their history elsewhere.

I built a Chrome extension that shows contributor reputation right on the PR page: merge rates across repos, spray-and-pray patterns, red flags.

Not sure I got the signals right. Would love feedback from maintainers.