Anthropic's Mythos AI Skips CISA in Federal Cybersecurity Rollout, Raising Coordination Questions

The Cybersecurity and Infrastructure Security Agency (CISA) lacks access to Anthropic's Mythos Preview, the AI model designed to identify and patch security vulnerabilities—despite the agency's mandate as the nation's central coordinator for federal cybersecurity defense. The exclusion, reported by Axios and confirmed by The Verge, creates a conspicuous gap in the federal rollout of advanced AI-driven security tools, with other major agencies reportedly already deploying the model.

Federal agencies including the Commerce Department and the National Security Agency have gained access to Mythos Preview, Anthropic's flagship product marketed specifically for vulnerability detection and remediation. The model represents a new category of AI application aimed at automating parts of the security patching workflow. However, CISA—tasked with coordinating cybersecurity across the federal government—remains outside that pipeline. The Trump administration is currently negotiating expanded access across additional agencies, according to sources cited by Axios.

The situation surfaces questions about coordination gaps within the federal cybersecurity apparatus. A central agency responsible for setting baseline security standards and orchestrating response across departments operates without the same vulnerability-hunting capabilities available to other federal bodies. CISA's role includes overseeing the Trusted Internet Connections program and Continuous Diagnostics and Mitigation initiative, both of which depend on timely awareness of exploitable weaknesses. The implications of incomplete access to next-generation security tooling could complicate those mandates, particularly as AI-assisted vulnerability discovery becomes a more prominent feature of the threat landscape.