Thousands of high- and critical-severity vulnerabilities emerged from Anthropic's Claude Mythos Preview. The model identified flaws in every major operating system and web browser without specific training for security research. These findings force a rethink of how developers secure code against AI-driven exploits. Frontier Red Team evaluations confirm the risk.