Claude Mythos identifies critical security gaps at scale – Anthropic limits access through Project Glasswing and offers a guardrailed parallel model called Claude Fable 5.
TrendAI leverages Anthropic’s Claude model in Project Glasswing to automate source code analysis, enabling faster identification and coordinated disclosure of vulnerabilities in critical software.
Anthropic expands Mythos access to 150 new organizations; security experts warn of structural changes driven by frontier AI models and the risk of vulnerability chaining.
