The announced capabilities of Anthropic’s Claude Mythos Preview and the formation of the Project Glasswing coalition represent a structural shift in the vulnerability discovery threat model. The reported ability to autonomously discover thousands of previously unknown vulnerabilities across OS and browser codebases signals that the timeline for adversarial exploitation of frontier AI capability is compressing. Enterprise teams should treat this as a forcing function to accelerate patching on memory safety flaw classes, audit AI agent integrations, and establish EU AI Act governance readiness before August 2, 2026.