Expanding Project Glasswing (anthropic.com) AI

Anthropic says it is expanding Project Glasswing, which lets vetted partners use Claude Mythos Preview to scan codebases for vulnerabilities, from about 50 initial partners to roughly 150 new organizations across more than 15 countries. The company says partners have already found over 10,000 high- or critical-severity flaws and that the program will increasingly emphasize disclosing, fixing, and deploying patches, alongside tools like Claude Security. Anthropic also outlines plans to further broaden access and its Cyber Verification Program as advanced cyber-capability AI becomes more widely available.