Manifold Scores 7,700 MCP Servers in Manifest Expansion Aimed at Agent Security Teams

Why It Matters

By quantifying provenance and safety of MCP servers, Manifold gives enterprise security teams a verifiable way to vet AI‑agent integrations, reducing hidden data‑exfiltration and execution risks.

Key Takeaways

• Manifold indexed 7,700 MCP servers in Manifest expansion.
• Composite Manifest Score combines Lineage and Safety signals.
• Free tier offers searchable index; enterprise tier adds automatic enrichment.
• Platform now covers 206,000 assets across skills, plugins, and servers.
• Manifold raised $8 million to accelerate development.

Pulse Analysis

The rapid adoption of Model Context Protocol (MCP) as the connective tissue for AI agents has outpaced traditional security reviews. MCP servers act as gateways, allowing agents to invoke external tools and data sources, yet many expose only minimal HTTP endpoints with sparse metadata. This creates a blind spot for security teams that must rely on trust rather than concrete verification, heightening the risk of data leakage or malicious command injection in enterprise workflows.

Manifold’s Manifest expansion addresses this gap by assigning each MCP server a composite score derived from two distinct signal families. The Lineage Score evaluates the publisher’s history, repository age, commit patterns, and identity consistency, while the Safety Score scrutinizes declared interfaces for contradictions, coercive instructions, or embedded prompt‑injection attempts. By surfacing these scores alongside existing skill and plugin indexes, security analysts can quickly prioritize high‑risk servers, enforce policy controls, and integrate automated risk assessments into their CI/CD pipelines.

The move also signals broader market momentum for AI‑focused supply‑chain security. With over 206,000 assets now cataloged across 31,472 unique publishers, Manifold demonstrates the scalability of its model and the appetite for granular risk intelligence. The recent $8 million capital infusion underscores investor confidence in the need for such tooling as enterprises accelerate agentic deployments. Looking ahead, expanded coverage of additional registries and GitHub‑hosted servers could cement Manifest as a de‑facto standard for AI‑agent security governance.