Anthropic Opens Claude Security Public Beta for Automated Code Vulnerability Scanning

AI-Powered Code Security Now in Public Beta

Anthropic launched the public beta of Claude Security on April 30, 2026, making it available immediately to Claude Enterprise customers, with access for Claude Team and Max customers to follow.

What It Does

Built on Claude Opus 4.7, Claude Security scans codebases for vulnerabilities by tracing data flows and examining how components interact across files and modules. A multi-stage validation pipeline reduces false positives before any finding reaches a reviewer.

• Flexible scanning: Target the entire codebase, specific directories, or branches
• Patch generation: Each finding includes a reviewable patch suggestion with confidence level and severity rating
• Workflow integrations: Send results to Slack, Jira, or other tools via webhook
• Export: Download findings as CSV or Markdown for audit systems
• Triage tracking: Dismiss findings with documented reasons for audit trail

Security Partners

Named integration partners include CrowdStrike, Microsoft Security, Palo Alto Networks, SentinelOne, TrendAI, and Wiz.

Access

Admins enable Claude Security via the admin console. It is accessible from the Claude.ai sidebar or directly at claude.ai/security, with no API integration or custom agent setup required.

Source: Help Net Security, SiliconANGLE