Anthropic Launches Claude Security Public Beta for Enterprise Codebase Analysis
Anthropic has released the public beta of Claude Security, an enterprise-focused tool that uses the Opus 4.7 model to scan codebases for vulnerabilities and generate automated patches.
Anthropic has launched the public beta of Claude Security, a new capability designed to assist enterprise organizations in identifying and remediating vulnerabilities within their codebases. Previously known as Claude Code Security, the tool is now accessible to Claude Enterprise customers directly through the Claude.ai interface Help Net Security.
The system leverages the Claude Opus 4.7 model to perform deep analysis of source code. By tracing data flows and examining how various components interact across different files and modules, the tool identifies potential security weaknesses. When a vulnerability is detected, the system provides a comprehensive report that includes the severity level, confidence score, potential impact, and detailed steps to reproduce the issue Help Net Security.
Beyond detection, Claude Security offers automated remediation support by generating targeted patch instructions. Users can review these suggestions in context before applying them. The platform also includes administrative features such as the ability to scope scans to specific directories or branches, dismiss findings with documented justifications for audit trails, and export data in CSV or Markdown formats. Furthermore, organizations can integrate the tool with external systems like Slack or Jira via webhooks Help Net Security.
The launch of Claude Security follows Anthropic’s recent introduction of the Claude Mythos Preview, a model capable of identifying zero-day vulnerabilities and constructing functional exploits for major operating systems and web browsers. That preview is currently limited to a select group of partners and developers Help Net Security.
In response to these developments, OpenAI has announced plans to roll out its own specialized model, GPT-5.5-Cyber, to a select group of cyber defenders. OpenAI CEO Sam Altman stated that the company intends to collaborate with the broader ecosystem and government entities to establish "trusted access" for the model, with the stated goal of accelerating the security of corporate and critical infrastructure Help Net Security.
The introduction of these AI-driven security tools marks a significant shift in how organizations approach vulnerability management. As major AI labs compete to provide both offensive and defensive capabilities, the industry is moving toward a landscape where automated code analysis and rapid patching become standard components of the software development lifecycle. Future developments will likely focus on how these models integrate with existing security operations and the extent to which they can reliably reduce the burden on human security teams Help Net Security.