Claude Security adds continuous scanning and fix tracking in expanded enterprise beta
Anthropic announced on October 21, 2026 that Claude Security now scans continuously as code changes, tracks findings until they are fixed, learns from team-supplied environment context, and connects to other tools through an API and MCP, in an expanded public beta for Claude Enterprise customers.
Instead of scheduled point-in-time scans, teams connect repositories through the Claude GitHub app (including GitHub Enterprise Server) or access tokens from GitLab and other Git providers; Claude runs a deep baseline scan, then rescans changes based on risk and complexity. Projects can span multiple repositories to catch issues that live between services.
Claude Security is powered by Anthropic's full model family, including Claude Mythos, and no longer requires Cyber Verification Program access. Anthropic says thousands of organizations have used it to find over a million vulnerabilities since the beta opened in April.
The commercial hook: Enterprise organizations receive $25,000 in free usage credits per month for two months, applying to usage through December 31, 2026, with terms in the Help Center.
For security leads, continuous scanning closes the gap between release cadence and review cadence, but it also means granting a third-party AI persistent read access to source code - worth routing through existing vendor security review before connecting production repositories.
Fix tracking and environment context are the features to test in a pilot: a scanner that remembers what was fixed and why is only useful if its deduplication and false-positive handling hold up on real repos.