"A successful attack could be catastrophic”: Anthropic gives more groups access to Claude Mythos
Summary
Anthropic is expanding Project Glasswing and giving roughly 150 additional organizations access to Mythos Preview for defensive security work. The company says the model can scan codebases, find vulnerabilities, and suggest patches, and it is also releasing Claude Security more broadly for trusted security teams. The initiative targets organizations across critical industries such as power, water, healthcare, communications, and hardware. Anthropic frames the program as a way to help the software industry adapt to stronger AI-driven cyber defense while reducing the risk of misuse. The expansion creates a clear opening for security vendors, open-source maintainers, and infrastructure teams that need vulnerability discovery and patching support.
