Netskope advances security for AI agent Model Context Protocol communications

Netskope introduced enhanced capabilities on its Netskope One platform aimed at securing communications involving the Model Context Protocol (MCP), which connects Artificial Intelligence (AI) agents to enterprise data and tools.

The adoption of MCP is increasing as it facilitates AI agents' access to enterprise resources. However, the protocol's ability to enable autonomous commands across numerous public MCP servers introduces various security concerns. The new platform functions aim to address these risks by providing comprehensive oversight and control over MCP traffic.

Technically, the platform now offers continuous and real-time identification of MCP servers and clients, including details such as name, ID, URL, version, host, data source, and protocol. It extends Netskope's Cloud Confidence Index risk scoring to assess MCP servers, supports granular and context-aware policy enforcement with default blocking options for MCP traffic, detects non-human interactions among MCP components, logs relevant MCP events, and identifies sensitive data usage related to MCP tools.

This update allows organizations using Netskope One to manage security for their MCP-enabled AI use cases by maintaining visibility, enforcing least-privilege access, protecting sensitive information, and supporting compliance requirements.

John Martin, Chief Product Officer at Netskope, said, “Every team wants to confidently accelerate AI adoption, and emerging protocols such as MCP are now fundamental to that discussion. MCP also creates new security risks that legacy tools can't solve. That's why we're further extending the market-leading capabilities of Netskope One to enable teams to see and create policies for MCP traffic and immediately assess how risky MCP tools are. This is critical to the secure use of AI as organizations develop agents to drive business productivity.”

The new MCP security features are currently available in a preview state for Netskope customers, with General Availability (GA) planned in the first half of 2026.