Anthropic Releases Claude Security Plugin for Claude Code in Beta: A Multi-Agent Vulnerability Scanner That Runs in Your Terminal

Anthropic has introduced a new security-focused plugin for Claude Code, marking a significant step forward in AI-assisted software development security. The Claude Security plugin, currently available in beta, enables developers to conduct comprehensive vulnerability scans directly within their Claude Code environment. This tool leverages a multi-agent approach to identify potential security issues in code repositories while maintaining developer control over the patching process.

The Claude Security plugin functions as an integrated vulnerability scanner that operates from within an active Claude Code session. Upon activation, the plugin deploys multiple AI agents to thoroughly examine a repository's codebase. Once the scanning process completes, the system presents discovered vulnerabilities to developers, who can then select which issues to address. The plugin subsequently generates patch files for the selected vulnerabilities, which developers review and apply at their discretion. This human-in-the-loop approach ensures that security improvements align with each organization's specific requirements and risk tolerance.

The beta release represents Anthropic's commitment to bridging the gap between development velocity and security assurance, addressing a critical concern in modern software development practices.

• Enhanced security posture for development teams using Claude Code without requiring separate security tools
• Reduced context-switching by integrating vulnerability scanning into existing developer workflows
• Improved code review processes through AI-identified potential security gaps before deployment
• Developer-controlled patching maintains oversight and prevents unwanted automated changes
• Potential acceleration of security-focused development practices across organizations adopting Claude Code
• Sets precedent for other AI coding assistants to integrate comprehensive security features

As artificial intelligence increasingly influences software development practices, integrating security measures directly into AI coding assistants becomes essential. The Claude Security plugin addresses a critical vulnerability in the development pipeline—the potential for AI-generated code to introduce security risks. By providing developers with an accessible, integrated security scanning tool, Anthropic demonstrates that AI assistants can enhance rather than compromise code security. This plugin release signals a broader industry movement toward responsible AI development tools that prioritize both productivity and security, ultimately supporting organizations in maintaining robust security standards while accelerating development timelines.