Anthropic launched OSS Scanner on October 8, a free opt-in service that periodically scans eligible open-source projects for security vulnerabilities using its strongest models, including Claude Mythos.

Anthropic said its models found more than 29,000 candidate vulnerabilities over six months, while people manually reviewed about 6,000. Nearly 5,000 unverified reports were also sent to maintainers who specifically asked to receive everything the models found.

To secure open-source software, we’re launching OSS Scanner.

We’ll use our frontier models to periodically scan opted-in open-source projects for vulnerabilities, at no cost. Our reports will provide a proof-of-concept, explanation, and suggested fix.https://t.co/UFFrLUd25U

— Anthropic (@AnthropicAI) October 8, 2026

Anthropic’s official OSS Scanner announcement.

That speed comes with a tradeoff. Reports are fully model-generated and arrive without human review. Anthropic says an early validation test found that 85 of 97 high- or critical-severity findings across 48 projects met its coordinated-disclosure bar. The Verge noted the same tension: AI bug hunting can uncover real flaws while creating more reports for already stretched maintainers.

The launch fits a broader push toward controlled AI security tooling. BitcoinVersus.Tech recently covered Anthropic restricting internet access during internal evaluations, Microsoft sandboxing AI agents on Windows 11, and Palo Alto Networks expanding continuous AI security testing.

Anthropic explains how Claude is used in cybersecurity workflows.

Core maintainers of eligible projects can enroll through Anthropic’s OSS Scanner program.