A Single Inspection Point Across Claude Enterprise Surfaces

Claude Enterprise centralizes DLP checks for prompts and tool responses before processing, with real-time blocking across its products.

Compliance teams now have a single inspection point for every prompt and tool response before it reaches the model.

The organization’s DLP server issues an allow-or-block decision that is enforced in real time. The mechanism covers chat, Claude Code, Claude Cowork, and other Enterprise surfaces, including tool calls made through MCP connectors, skills, or plugins. Previously limited to client-side hooks in Claude Code, this control can now be applied centrally across the organization.

Each inference request passes through a signed WebSocket connection to the security server. The prompt and its context are inspected before generation begins, while tool responses are checked before being returned to the model. The open protocol, built on webhooks and a documented schema, works with existing DLP infrastructure from Netskope, Palo Alto Networks, Proofpoint, Zscaler, or an organization’s internal solution. Shadow mode, role-based exclusions, and gradual percentage-based rollouts help streamline deployment.

The feature is available in beta to Claude Enterprise customers.