Anthropic has introduced Claude Code Security, an AI-powered service built into Claude Code that scans software codebases for vulnerabilities and recommends fixes, with the company rolling out the capability in a limited research preview for Enterprise and Team customers.
The company said in its announcement that the new capability scans codebases for security flaws and suggests targeted software patches for human review, with the goal of helping teams find and fix issues that traditional methods often miss.
How The Tool Works
Anthropic said Claude Code Security is designed to identify complex, high-severity software vulnerabilities by analyzing code in a way it compares to a human researcher. The system focuses on data flows, understands how components interact, and looks for issues that rule-based scanners can overlook.
Each finding goes through a multi-stage verification process before it is presented to users. The company said Claude re-examines each result, attempts to prove or disprove its own conclusions, and filters out false positives. Findings are assigned severity and confidence ratings and appear in a dashboard for human review, which is intended to help teams prioritize fixes.
Testing And Research Background
Anthropic said the service draws on more than a year of work on Claude’s cybersecurity capabilities. The company said the model was tested in Capture-the-Flag events and in work with Pacific Northwest National Laboratory, where it was refined to detect and patch real vulnerabilities.
Using Claude Opus 4.6, the team said it identified more than 500 previously undetected bugs in open-source projects. The company said Claude Code Security brings those defensive capabilities into Claude Code so teams can review findings and iterate on fixes within their existing tools.
In its announcement, Anthropic said it expects a large share of the world’s code to be scanned by AI in the near future and said attackers will also use AI to find exploitable weaknesses faster. The company said defenders who move quickly can use similar tools to find and patch those weaknesses.
Availability And Access
Claude Code Security is being offered in a limited research preview for Enterprise and Team customers, as well as for open-source maintainers. Anthropic said open-source maintainers can apply for free, expedited access, while Enterprise and Team customers will receive early access and collaboration during the preview period.
Featured image credits: LAv1
For more stories like it, click the +Follow button at the top of this page to follow us.