Krux
February 22, 2026
Anthropic's Claude Now Hunts Critical Bugs Across Codebases
Published: February 22, 2026 at 3:03 PM
Updated: February 22, 2026 at 3:03 PM
100-word summary
Anthropic launched Claude Code Security on February 20, 2026, a tool powered by Claude Opus 4.6 that autonomously scans entire codebases to find and prioritize high-severity vulnerabilities humans often miss. Unlike traditional pattern-based scanners, it performs end-to-end analysis, traces data flows across components, scores severity automatically, and suggests fixes without applying them. The research preview is available now for Enterprise and Team customers, with expedited access for open-source maintainers. Security teams gain faster triage and developers get guided remediation, addressing a critical gap in enterprise and open-source security workflows as vulnerabilities grow more complex.
What happened
Anthropic launched Claude Code Security on February 20, 2026, a tool powered by Claude Opus 4.6 that autonomously scans entire codebases to find and prioritize high-severity vulnerabilities humans often miss. Unlike traditional pattern-based scanners, it performs end-to-end analysis, traces data flows across components, scores severity automatically, and suggests fixes without applying them. The research preview is available now for Enterprise and Team customers, with expedited access for open-source maintainers.
Why it matters
Security teams gain faster triage and developers get guided remediation, addressing a critical gap in enterprise and open-source security workflows as vulnerabilities grow more complex.