Anthropic Launches Free, Opt-In AI Security Scans for Open Source

What changed

Anthropic has launched OSS Scanner, a free, opt-in service for open-source projects. It promises periodic security scans using Anthropic’s strongest models, but its reports will not receive human review.

Why it matters

For open-source maintainers, the useful change is access: projects can add recurring AI checks without paying for them. That could surface possible vulnerabilities earlier, though the scan is a signal to investigate, not a human security assessment.

The missing review matters. AI can flag suspicious code, but maintainers still have to verify the finding, judge its severity and decide what to change. Anthropic has not specified which repositories qualify, how often scans run, how disclosures are handled or how accurate detection is. Until those details are clear, OSS Scanner looks best treated as an extra pair of automated eyes, not a security team in a box.