Claude Security adds continuous scanning and fix tracking
The Enterprise public beta now watches repositories as code changes, tracks whether findings actually get fixed, and exposes an API and MCP connector for Claude.ai and Claude Code.
Anthropic expanded Claude Security’s public beta on 21 October with four additions for Claude Enterprise customers: continuous repository monitoring, automatic fix tracking, team-informed threat models, and an API with an MCP connector.
Scanning used to mean running a check on demand. Now Claude Security watches a repository as code changes and rescans automatically. It also tracks each finding against the codebase, marking it resolved when the fix lands and reopening it if the vulnerable pattern comes back. A new project-level knowledge base lets a team record deployment details and existing controls, which Claude Security uses to prioritise findings and attach an explanation of how an attacker could actually reach the vulnerable code. The new API lists projects and findings and can resolve issues programmatically, and an MCP connector in Claude.ai and Claude Code lets you fix a flagged issue without leaving the session.
Anthropic says the tool has found over a million vulnerabilities since the beta opened in April, and is offering Enterprise customers $25,000 a month in free usage credits for two months, through 31 December.
Why it matters: continuous scanning plus fix tracking turns this from a one-off report you triage into something closer to a running dashboard of what’s actually still broken, which is the part most AI-assisted scanners skip.
The caveat: this tier is public beta and limited to Claude Enterprise customers, not something a solo developer or small team can turn on today.