Anthropic has launched Claude Security in public beta for Claude Enterprise customers, giving organizations a purpose-built tool to scan codebases for vulnerabilities and generate proposed fixes using Claude Opus 4.7, which the company describes as among the strongest models available for finding and patching software vulnerabilities. The tool was previously known as Claude Code Security and had been tested by hundreds of organizations during a limited research preview before today’s broader rollout.
Claude Security can be accessed directly from the Claude.ai sidebar or at claude.ai/security, and requires no API integration or custom agent build for Enterprise customers — a deliberate design choice aimed at lowering the barrier to adoption for security and engineering teams that may lack the resources to stand up bespoke AI tooling. Rather than relying on pattern matching or static rule sets, the platform scans repositories by reasoning about code the way a human security researcher would, understanding how components interact across files, tracing data flows, and reading source code in context. It then surfaces detailed findings that include confidence ratings, severity assessments, and instructions for targeted patches that can be worked through directly in Claude Code.
The public beta release arrives with a feature set built for enterprise security workflows. Anthropic has included scheduled and targeted scans, the ability to dismiss findings with documented reasons for audit purposes, and export options in both CSV and Markdown formats. Webhook integrations with Slack, Jira, and other widely used enterprise tools are also included, allowing security teams to route findings directly into their existing ticketing and alerting infrastructure without additional configuration. The combination of automated scanning and workflow integration is designed to reduce the time between vulnerability discovery and remediation across large, complex codebases.
Anthropic has also assembled a broad coalition of technology and services partners to extend Claude Security’s reach into the platforms and workflows organizations already rely on. Technology partners integrating Opus 4.7 into existing security platforms include CrowdStrike, Microsoft Security, Palo Alto Networks, SentinelOne, TrendAI, and Wiz — a lineup that spans some of the most widely deployed enterprise security tools on the market. On the services side, firms including Accenture, BCG, Deloitte, Infosys, and PwC are deploying Claude-integrated security solutions for their clients, signaling that major consulting practices are beginning to build AI-native security offerings around Anthropic’s models. Access for Claude Team and Max customers is expected to follow.
