We stopped AI bot spam in our GitHub repo using Git's –author flag
2 days ago
- #AI spam
- #GitHub contributions
- #Open source sustainability
- GitHub's reporting of AI contributions masked a decline in quality, with bots flooding issues and PRs.
- AI accounts generated up to 253 comments on a single issue, burying legitimate contributions and creating noise.
- The team developed reputation tracking and 'AI sheriff' tools, but these measures sometimes affected real contributors.
- To combat AI spam, they implemented an onboarding process requiring prior contributions for repo access.
- They used a hack involving Git's author flag to whitelist contributors via a manual commit attribution method.
- The article argues AI slop demotivates real contributors, increases security risks, and necessitates a discussion on AI's impact on open source.