news.volyx.in

We stopped AI bot spam in our GitHub repo using Git's –author flag (archestra.ai)

501 points by ildari · 103 days ago · 237 comments on HN

Article summary

The article discusses how the author's company, Archestra.ai, stopped AI bot spam in their GitHub repository by using Git's --author flag and implementing a system that requires contributors to complete a CAPTCHA on their website before being whitelisted. This approach allowed them to block at least 500 bots in the first week. The system uses a GitHub Action that adds the user's handle to an EXTERNAL_CONTRIBUTORS.md file and pushes a commit to the main branch authored under their account. This solution has been effective in reducing AI-generated spam PRs and issues.

Main themes

  • AI-generated spam
  • GitHub repository management
  • Bot mitigation strategies
  • Open-source community
  • Contributor authentication
  • CAPTCHA implementation

What commenters say

  • Implementing a token system could help mitigate AI-generated spam, but it may be vulnerable to gaming if not implemented well.
  • GitHub's business model is based on ads and may not have an incentive to block AI-generated spam, which could be a major problem for repositories with bounties.
  • A system where contributors are given tokens or scores based on their contributions could help distinguish between genuine and spammy contributions.
  • Blocking accounts that have a high rejection rate of PRs could be an effective way to reduce spam, but it may not be a foolproof solution.
  • The use of CAPTCHA and whitelisting can be an effective way to prevent AI-generated spam, but it may not be suitable for all repositories or communities.
  • The problem of AI-generated spam is not just a technical issue, but also a matter of misaligned incentives and the need for a more nuanced approach to community management.
  • Some commenters suggest that GitHub should implement features such as temporary blocking of accounts or flagging of suspicious activity to help mitigate AI-generated spam.
  • Others propose that a more holistic approach is needed, including changes to the way contributions are evaluated and rewarded, to effectively address the issue of AI-generated spam.