JFrog Launches AI Trust Layer with NVIDIA

At NVIDIA GTC, JFrog Ltd. announced the launch of its new JFrog Agent Skills Registry, a platform designed to bring governance, security, and trust to enterprise AI agents. Developed in collaboration with NVIDIA, the solution introduces a verified control layer to help organizations safely deploy and manage autonomous AI systems at scale.

The registry is built to support the NVIDIA Agent Toolkit, including NVIDIA OpenShell, an open-source runtime for developing and running long-lived AI agents. It also integrates with the NVIDIA AI-Q Blueprint, enabling JFrog Artifactory to act as a centralized repository for AI models, agent skills, and related software assets.

As AI agents become a core part of enterprise workflows, they increasingly rely on “skills” to perform tasks. However, unverified or malicious skills can introduce serious risks. JFrog’s Agent Skills Registry addresses this challenge by acting as a system of record that stores, scans, and governs all agent-related assets across the software supply chain.

According to JFrog, the platform ensures that every AI skill is validated before use. It automatically scans uploaded assets for vulnerabilities, malicious code, and compliance risks, helping organizations prevent compromised components from being deployed. This approach mirrors traditional software security practices but extends them to the emerging world of agentic AI.

The integration with NVIDIA provides enterprises with a single, governed endpoint for distributing trusted AI skills across different agent platforms. It also introduces a promotion model, where assets pass through increasing levels of security checks before being approved for wider organizational use. Key features of the solution include centralized visibility into AI assets, policy-driven access controls, and secure execution environments. Through its integration with OpenShell, the platform can run agents in isolated environments, reducing the risk of broader system compromise.

NVIDIA highlighted that governance and security are essential for scaling AI agents in enterprise environments. By combining JFrog’s software supply chain expertise with NVIDIA’s AI infrastructure, the partnership aims to enable organizations to adopt autonomous agents without sacrificing control or compliance. The launch reflects a broader shift in enterprise IT, where AI agents are becoming standard tools across development and operations. With this evolution comes the need for structured governance frameworks. JFrog’s Agent Skills Registry positions itself as a foundational layer to ensure that AI innovation is supported by strong security, transparency, and trust.