Anthropic00:00Guides & TipsPress Release
Anthropic previews Claude Code Security
Speed up vuln discovery and patch drafting with approval gates.
Key Points
- 1Focuses on vuln findings and patch suggestions
- 2No auto-apply; human approval required
- 3Limited research preview availability
- 4Designed for real development workflows
Anthropic introduced Claude Code Security in a limited research preview. It scans codebases, proposes fixes, and keeps humans in control by requiring approval before changes. This offers a practical path to adopt AI-assisted security reviews with oversight.