Manifold Security launches Manifest AI supply chain intelligence platform

Per Silicon Angle, Manifold Security Inc. has launched Manifest, a new supply chain intelligence platform designed to map and analyze the complex interactions between AI agent components and external systems. The platform aims to help enterprises understand the dependencies that underpin AI agent behavior, including connections to tools and infrastructure that could introduce risk.

The growing AI agent ecosystem presents significant supply chain complexity, with each component carrying its own trust assumptions and potential vulnerabilities. Manifest employs a graph-based approach to model component behavior and their interconnections. It builds execution graphs detailing skill functions and dependencies, alongside environment graphs tracking authorship and registry relationships. This dual perspective aims to reveal risks not easily detected by traditional file-level analysis. A recent report by Manifold highlights over 238,000 unique skills across registries, with confirmed malicious examples leading to data exfiltration. Current security tools often show inconsistent detection rates due to a lack of contextual understanding, as risks frequently arise from component relationships rather than individual assets.

Manifest addresses these inconsistencies by integrating ecosystem-level data, such as author activity and dependency chains, to enhance signal quality and reduce false positives. The platform is available as a free, open-access service with extensive search and analysis capabilities. Enterprise features offer broader coverage by integrating into Manifold's existing platform for securing AI-driven operations.

Source: Silicon Angle