Anthropic's new cybersecurity tool, Claude Security, is available for enterprise customers in public beta for Claude Enterprise customers across the world. The tool is designed to defensively scan your code for vulnerabilities and can also suggest and implement patches. Service partners including Accenture, BCG, Deloitte, Infosys, and PwC are already deploying Claude to increase enterprise security. The launch follows reports that Claude Mythos, Anthropic's most guarded model, has demonstrated its capabilities in identifying and exploiting software vulnerabilities. Claude Security uses a human researcher to identify vulnerabilities in code, giving results in a confidence rating and detailing of the severity and likely impact of a finding. The company also introduced real-time safeguards to identify and block risky or prohibited requests and launched Project Glasswing, the official name for the initiative that select partners can begin working with Mythos.