Anthropic Launches Free AI Security Scans for Open-Source Projects With 0 Human Review
Updated
Updated · The Verge · Oct 8
Anthropic Launches Free AI Security Scans for Open-Source Projects With 0 Human Review
3 articles · Updated · The Verge · Oct 8
Summary
Anthropic has opened an opt-in service called OSS Scanner that gives open-source projects free, periodic vulnerability scans generated by its strongest models, including Claude Mythos.
0 human reviewers check the reports before they are sent, a trade-off Anthropic says enables faster, more frequent scanning but can also produce incorrect or invalid findings.
AI bug-hunting tools have recently helped uncover major flaws, including May’s “Copy Fail” bug that affected nearly every Linux distribution.
Open-source maintainers are also facing a flood of AI-generated bug reports, with projects tied to Linus Torvalds and Google among those struggling to keep up.