Linux Foundation to Govern TRACE for AI Security

The Linux Foundation will now govern TRACE, a new open specification for AI agent security. TRACE creates a verifiable record of how AI agents and other confidential workloads operate. This record is hardware-backed and cryptographically verifiable. It links the runtime environment, software, policies, data classification, and invoked tools. OPAQUE developed TRACE with AMD, Intel, Microsoft, and the Technology Innovation Institute. It aims to be portable across different cloud providers and confidential computing platforms. This is important as AI agents handle sensitive data in production environments. TRACE integrates existing standards like RATS, EAT, and SLSA into a unified evidence layer. AMD and Intel note their hardware technologies, like SEV, provide silicon-level protection that TRACE transforms into verifiable evidence. The TRACE reference library has seen about 135,000 downloads on PyPI in 10 weeks. This initiative provides crucial, independent verification for AI operations, especially following reports of AI agents escaping testing environments.