Claude Enterprise Introduces New Security and Admin Features

Anthropic has announced a suite of new security and administrative features for Claude Enterprise, aimed at providing security teams with greater visibility and control over organizational settings, data handling, and user access. These updates include a security center, generally available inference hooks, centralized management, and additional access controls.

Security Center for Configuration Oversight

The new security center consolidates Claude’s security settings for chat, Cowork, and Claude Code into a single page. It compares an organization’s settings against Anthropic’s recommended configuration, highlighting discrepancies and providing links to adjust settings. The center refreshes automatically upon each page load and integrates with the Compliance API for external reporting. Administrators can also download the page as a report for audits, marking which controls were automatically checked and which were manually confirmed. While the security center offers guidance, it does not guarantee compliance with legal or regulatory standards.

Inference Hooks for Policy Enforcement

Inference hooks, now generally available, connect Claude to existing data security tools, enabling organizations to enforce data loss prevention (DLP) and other security policies. Two types of hooks are supported: prompt hooks, which send user prompts to security tools for review, and pre-tool-call hooks, which validate tool calls before execution. These hooks operate through signed HTTPS POST requests, allowing organizations to maintain their existing security policies while extending them to Claude’s environment.

Centralized Management Across Organizations

Centralized management introduces a parent-organization console at admin.claude.com, enabling administrators to manage company-wide settings across multiple Claude Enterprise and Claude Platform organizations. Features include single sign-on (SSO), SCIM directory synchronization, domain verification, group management, and Compliance API key oversight. The console also supports multiple identity providers, allowing acquired companies to retain their existing systems or transition gradually. Parent-level roles do not grant access to individual organizations by default but provide visibility and control over overarching settings.

Enhanced Access Controls and Shadow IT Prevention

New access controls allow administrators to sign out members via the Admin API, manage IP allowlists directly, and block personal or external Claude accounts on managed devices. These measures aim to reduce unauthorized usage and improve security oversight. Additionally, workspace folder restrictions can limit data access within Claude.

Endpoint Detection and Response Integration

To address visibility gaps in local virtual machines, Anthropic has introduced an Endpoint Detection and Response (EDR) integration for Cowork and unified Claude. This feature enables administrators to deploy vendor-signed EDR sensors within the isolated virtual machine on macOS and Windows devices. The sensors operate in detection-only mode, reporting activity to the organization’s existing security console without blocking actions. This integration does not extend to cloud-based Cowork sessions.

These updates collectively enhance Claude Enterprise’s administrative and security capabilities, offering organizations tools to better manage configurations, enforce policies, and monitor activity across devices.