As enterprises expand their use of AI agents linked to internal systems, Netskope has introduced new security features to monitor and control Model Context Protocol (MCP) traffic within the Netskope One platform.
This update is for organizations using MCP to connect AI agents to corporate data and tools. This protocol is emerging as the standard for agent-based AI workflows in the enterprise.
MCP allows AI agents to request data, trigger actions, and coordinate tasks across different applications. The same functionality also introduces new security risks when agents manipulate sensitive data or trigger automated actions.
Netskope said the new features allow it to visualize MCP communications and apply access policies to them. The company says this applies to both MCP servers and MCP clients running within an organization.
The platform can now identify MCP servers and clients in use. Attributes such as name, ID, URL, version, host, data source, and protocol are listed.
Netskope is also extending its Cloud Confidence Index (CCI) risk scoring system to MCP servers. CCI is the scoring method used by Netskope for its cloud and SaaS services.
Enhanced CCI scoring gives risk ratings to the MCP servers that AI agents connect to within your organization. According to Netskope, this is intended to help security teams prioritize which AI tools, agents, or integrations pose the greatest risk.
Security teams can then use policy controls within Netskope One for those MCP endpoints. You can allow or block MCP traffic based on assessed risk and context.
The company said the platform supports default blocking options for MCP traffic. Administrators can grant access in certain approved cases.
Netskope said its controls follow a least privilege approach. This means you can limit AI agents and tools to the minimum access needed for their tasks.
This feature also detects and monitors non-human traffic between MCP servers, clients, tools, hosts, data sources, and development tools. The platform treats this as a separate traffic category within the monitoring interface.
Netskope said it can record MCP events in detail. Logs include sessions, initializations, tool requests, responses, and deployments.
This system also inspects data passing through the MCP tool. You can identify sensitive information such as intellectual property and passwords in that traffic.
This inspection enables data loss prevention rules for MCP traffic. These rules can block or quarantine submissions containing specific types of content.
The company says these features will work in parallel with existing Netskope One features. This includes real-time traffic inspection, access control, and data protection across cloud, web, and private applications.
John Martin, chief product officer at Netskope, said AI adoption is now a central focus for most enterprise teams. He said MCP is currently at the center of many of these efforts.
“Every team wants to accelerate AI adoption with confidence, and emerging protocols like MCP are now a cornerstone of that discussion,” said John Martin, chief product officer at Netskope. “MCP also creates new security risks that traditional tools can't resolve. That's why we're further extending the capabilities of Netskope One to help teams review and create policies for MCP traffic and instantly assess the risk of MCP tools. This is critical for organizations to use AI securely as they develop agents that drive business productivity.”
MCP has gained recognition as vendors and developers create databases, SaaS platforms, and servers that expose internal tools to AI agents. Many of these servers are publicly available and can be connected to a wide range of models and agent frameworks.
Security leaders have expressed concern that unmanaged use of MCPs can expose sensitive corporate assets. They also warn that unconstrained autonomous commands from AI systems could impact production systems.
Vendors are responding with products that treat AI agent traffic as its own category. They are building controls similar to those used in the cloud and SaaS, but applied to AI interactions.
Netskope says the new MCP security features are currently in preview for customers. The company plans to make it generally available in the first half of 2026.
