Claude now reviews and fixes vulnerabilities as you write code

Anthropic released a security-guidance plugin for Claude Code that automatically reviews code changes for common vulnerabilities across three integrated stages: lightweight pattern checks during edits, model-assisted diffs after model turns, and deeper repository-aware reviews during commits/pushes. The plugin aims to catch issues like injection flaws, unsafe deserialization, and insecure DOM APIs before pull requests, is extendable with custom rules, and is free and available for Claude Code v2.1.144+ (deeper reviews require a git repo and use Claude usage budget).