Anthropic opens Claude Security to enterprise customers in beta mode

Anthropic has launched Claude Security, an AI-powered tool that scans codebases for vulnerabilities and automatically generates patches, in beta mode for its enterprise customers.

The product is powered by Claude Opus 4.7 and has been rolled out to Claude Enterprise customers globally, with access for Team and Max subscribers to be launched soon.

The tool was first released as Claude Code Security in research preview mode in February. Hundreds of organisations have since used it to detect bugs that existing tools had missed for years.


Shares of companies in the cyber security solutions space nosedived after Anthropic unveiled the tool.

Unlike traditional scanners that match against known vulnerability patterns, Claude Security reasons through code the way a human researcher would. It traces data flows and maps how components interact across files and modules. A multi-stage validation pipeline then verifies each finding independently before it reaches an analyst, with every result assigned a confidence and severity rating.

From there, users can move directly into a Claude Code session to review and apply a fix, with full context on the issue's likely impact and reproduction steps.

This is part of Anthropic's wider push to make frontier AI capabilities available to defenders. It sits alongside new cyber safeguards built into Opus 4.7, which automatically detect and block high-risk cybersecurity requests, and Project Glasswing, Anthropic's initiative to protect critical software infrastructure.

Anthropic’s technology partners, including CrowdStrike, Palo Alto Networks, SentinelOne, TrendAI and Wiz, are integrating Opus 4.7 into their platforms.