Anthropic launches Claude Security public beta on Claude Opus 4.7
Anthropic just rolled out Claude Security, a new AI-powered tool designed to help spot and fix security flaws in software.
Right now, it's in public beta for enterprise users and runs on the Claude Opus 4.7 model.
This launch follows its earlier Project Glasswing, which also focused on catching code vulnerabilities.
Claude security scans and suggests fixes
Claude Security scans whole code bases or folders, kind of like having an AI security researcher on your team.
It grades each issue by confidence and severity, then suggests fixes so developers know what needs attention first.
You can plug it into existing security setups with scheduled scans or export data as needed.
Anthropic is also teaming up with companies like CrowdStrike and Deloitte to bring this tech into more cybersecurity products for businesses.