記事の概要
Netskope has introduced new security controls designed for the Model Context Protocol (MCP) to safeguard enterprise AI agents.
- These controls aim to secure AI assistants and agents that use MCP for communication and context exchange with tools and resources.
- The new capabilities provide granular visibility and control over data flow between AI assistants (such as those leveraging Anthropic's Claude via MCP) and external applications, APIs, and databases.
- Key features include Data Loss Prevention (DLP), threat protection, access control, and audit/compliance reporting for AI agent activities over MCP.
- This development addresses critical security concerns like data leakage and unauthorized access, facilitating the secure adoption of AI technologies in enterprises.