Claude Mythos AI Finds 10,000 High-Severity Flaws in Widely Used Software
Anthropic disclosed on May 23, 2026, that its Project Glasswing has identified more than 10,000 high- or critical-severity vulnerabilities across systemically important software worldwide. Launched last month, the initiative uses Anthropic's Claude Mythos AI to proactively secure critical global software. The vulnerabilities were found in widely used applications, though specific software names were not disclosed. The findings underscore the growing role of AI in cybersecurity, particularly for defensive purposes. Anthropic emphasized that Project Glasswing is a defensive effort aimed at strengthening the security posture of essential software infrastructure. The scale of vulnerabilities discovered—over 10,000 in just one month—highlights both the prevalence of security flaws and the potential for AI-assisted vulnerability discovery. The company has been working with affected vendors to patch the flaws, though no details on exploitation or impact have been released.
Developers must prioritize patching these widespread vulnerabilities to prevent potential exploits.