Monday, December 22, 2025
Netskope has introduced a dedicated security framework for the Model Context Protocol (MCP), addressing one of the most urgent challenges emerging in the AI ecosystem: securing autonomous AI agents as they interact with real business data. The new capabilities, built into the Netskope One platform, aim to give organizations the visibility, control, and protection they need as MCP adoption accelerates across the industry.
First launched by Anthropic in November 2024, MCP has quickly become a critical open standard for connecting AI agents to external systems. By acting as a universal connector, MCP enables AI applications to access company data, integrate with software tools, and automate complex tasks. Its rapid adoption is reshaping enterprise workflows. OpenAI, Google, and Amazon have all integrated the protocol, and more than a thousand publicly accessible MCP servers have surfaced in just a short time.
While this growth highlights MCP’s importance, it also introduces significant security risks. The protocol’s power—granting AI systems access to sensitive datasets and allowing them to issue commands—can become a vulnerability if not properly monitored. Public MCP servers, misconfigurations, and unauthorised agent interactions increase the risk of data exposure and unwanted system activity.
Netskope’s new MCP security features are designed to close those gaps. Through the Netskope One platform, organisations receive continuous, real-time visibility into all MCP servers, clients, and tools operating within their environment. The platform automatically identifies MCP components and logs detailed attributes, including server name, ID, version, host, URL, and connected data sources. This level of insight helps security teams detect shadow tools, unauthorized integrations, or suspicious machine-to-machine communication.
The company is also extending its trusted Cloud Confidence Index (CCI) risk scoring model to MCP servers. This enables enterprises to evaluate the security posture of AI tools and integrations instantly, identifying which MCP-connected agents pose the highest compliance or data protection risk.
Beyond monitoring, Netskope is introducing advanced access controls tailored for MCP workflows. Organizations can enforce granular, context-aware policies that define which agents or tools are allowed to make MCP requests. For added security, a default block mode can halt all MCP activity until it is explicitly authorized. To protect intellectual property, credentials, and proprietary data, Netskope has also embedded real-time data loss prevention (DLP) capabilities into MCP interactions.
“Every team wants to confidently accelerate AI adoption, and emerging protocols such as MCP are now fundamental to that discussion,” said John Martin, Chief Product Officer at Netskope. “But MCP also creates new security risks that legacy tools can’t solve. That’s why we’re extending the market-leading capabilities of Netskope One.”
The new MCP security functions are available now in preview for Netskope customers, with full general availability expected in the first half of 2026.
To join our expert panel discussions, reach out to info@intentamplify.com
Recommended News
Tags:
ByteDance debuts its new AI voice assistant powered by the Doubao LLM on ZTE’s Nubia M153 smartphone, priced at CNY 3,499 and now available for limited pre-orders
ChatBar AI and Anam introduce human-real AI personas combining real-time avatars with ChatBar’s secure RAG engine, offering lifelike, brand-aligned AI interactions deployable in minutes.
