From the source
Anthropic is adding security visibility and control features to Claude Enterprise, giving security teams one place to review how Claude is configured across chat, Cowork and Claude Code.
- The security center is a new page in Organization settings that compares current settings with Anthropic's recommended configuration, showing the current value, the recommended value and a link to change it, and it rechecks the configuration each time the page loads.
- Anthropic says the page is guidance based on its recommendations, not a certification or a guarantee of compliance with any law or standard, and organizations with the Compliance API can read the same results through it.
- Two kinds of inference hooks are now generally available: prompt hooks send prompts from supported Claude products to a security tool for a check, and pre-tool-call hooks check tool calls before they run.
- Centralized management gives a parent organization its own admin console at admin.claude.com for company-wide settings such as single sign-on, directory sync, domain verification, groups and Compliance API keys, with new parent owner and parent admin roles.
- Admins can now sign a member out of Claude with a single Admin API call, manage their own IP allowlist, and block personal and other outside Claude accounts in Claude Desktop on managed computers to help prevent shadow IT.
- Anthropic also announced an Endpoint Detection and Response integration that runs an organization's sensor inside the unified Claude and Cowork virtual machine on the user's computer, covering that virtual machine rather than Cowork sessions in the cloud.
