Manifold Security Covers AI Agent Servers in Supply Chain

Manifold Security has expanded its Manifest supply chain intelligence tool to cover AI agent servers. This update aims to close a security gap in AI agent integrations. The tool now includes scored entries for over 7,700 Model Context Protocol, or MCP, servers. MCP is important for connecting AI agents to outside resources. Each server gets a Manifest Score, which comes from a Lineage Score and a Safety Score. The Lineage Score checks where the publisher came from, and the Safety Score looks for behavioral risks. Many MCP servers only show an HTTP endpoint, making their security hard to check. A compromised MCP server can control an agent's reasoning and data. This expansion helps enterprises understand the provenance and behavior of these critical AI components.