Anthropic Launches Claude Security to Fight AI Cyber Threats

Anthropic has launched Claude Security to help defenders keep pace with a surge in AI-powered cyberattacks, saying that rapid tooling like Mythos can drastically shorten the window from discovery to exploitation. Claude Security is currently in public beta for Claude Enterprise customers, offering code-scanning capabilities that identify vulnerabilities and propose fixes using Opus 4.7 on the Claude Platform or via partners building with Claude.

The release also enables targeted scans of specific directories within a repository, and allows findings to be exported in CSV or Markdown formats and sent to Slack, Jira, or other tools through webhooks. Anthropic notes improvements in detection accuracy, reduced false positives, and added confidence scoring based on feedback from hundreds of organisations.

The company is integrating Claude’s technology with major security platforms through partners including CrowdStrike, Microsoft Security, and Palo Alto Networks, alongside consulting firms such as Deloitte and Accenture, to bolster defences as AI accelerates cyber threats. Claude Security also features detailed findings with confidence levels, severity, impact, and reproduction steps to guide remediation.