The Linux Foundation has announced it will host TRACE (Trust, Runtime Attestation and Compliance Evidence), an open specification designed to provide hardware-backed runtime evidence for AI agents and confidential workloads. The specification, contributed by OPAQUE, was developed collaboratively by AMD, Intel, Microsoft, OPAQUE, and the Technology Innovation Institute (TII).
TRACE aims to establish a standard, open evidence layer that enables reliable governance records for AI agents, particularly as organizations deploy increasingly autonomous systems and open-weight models. The initiative seeks to provide a consistent, trustworthy method for proving that sensitive data is handled in accordance with established policies.
The specification creates a standardized, hardware-enforced governance record. This record binds together elements such as the runtime environment, software, policies, data classifications, and tool usage into a portable, cryptographically verifiable artifact. This artifact is designed to travel with the workload across various cloud and confidential computing environments.
The Linux Foundation emphasized that TRACE provides the open-source community with a unified, hardware-attested specification for compliance and security evidence. Hosting TRACE under neutral governance is intended to ensure that trust in AI remains open, portable, and verifiable across different infrastructures.
Rather than introducing an entirely new security framework, TRACE builds upon existing industry standards. These include RATS, EAT, SLSA, SCITT, SPIFFE, and EAR, composing them into a common evidence layer. This layer is designed to function across enterprise, cloud, and sovereign AI infrastructures. The initial specification focuses on current AI agent architectures while laying a foundation for future multi-agent and confidential computing deployments.
The project has seen significant early interest, with nearly 135,000 PyPI downloads within 10 weeks of its initial introduction at the Confidential Computing Summit in June 2026. Under the Linux Foundation, TRACE will receive vendor-neutral governance to support long-term adoption and sustainability, with its technical workstream managed by the Coalition for Secure AI (CoSAI).
Proponents of TRACE argue that runtime evidence must be portable, independently verifiable, and not controlled by a single vendor to be effective as a standard. They believe TRACE offers enterprises and regulators a common method to verify what executed, which policies were enforced, and how AI systems interacted with sensitive data across diverse clouds and infrastructure. Contributing TRACE to the Linux Foundation provides a vendor-neutral home where the industry can shape an interoperable standard.
Intel noted that as AI agents become more autonomous and interact with other agents, sensitive data, and critical business systems, organizations require cryptographic evidence of an agent's identity, authorization, execution environment, and proof of policy enforcement. Hardware-based attestation and confidential computing are seen as crucial enablers for this at scale, forming a basis for independently verifiable AI where enterprises can make informed decisions before agents access data, invoke tools, or delegate actions.
The open specification, technical documentation, and reference implementations for TRACE are currently available at trace.agentrust-io.com and on GitHub.
