OPAQUE Introduces OPAQUE 3.0 with Verifiable AI Governance and Post-Quantum Security

OPAQUE has launched OPAQUE 3.0, a new platform designed to bring verifiable trust and governance to AI agents, announced in a press release. The release took place at the Confidential Computing Summit, which OPAQUE co-hosts with the Linux Foundation. The platform introduces open standards for cryptographic identity and confidential execution, allowing enterprises to verify what AI agents ran, where they ran, and whether governance policies were enforced.

OPAQUE 3.0 includes two major open source components. The first is Agent Manifest, an extension of the Agent Governance Toolkit that enables cryptographic verification of AI agent integrity and authorization. The second is Confidential MCP, the first verifiably governed and secure implementation of the Model Context Protocol. Both are available at github.com/agentrust-io.

The Technology Innovation Institute (TII) is a founding partner in the standard, bringing post-quantum cryptography to ensure agent identities and signed records remain secure against future quantum attacks. The platform also integrates with confidential computing technologies from Intel, AMD, and NVIDIA. OPAQUE 3.0 runs across public cloud, private cloud, and on-premises environments.

OPAQUE plans to make the platform generally available in July 2026. It also announced the agentrust-io Fellowship, a program for engineers and researchers to contribute to the open governance standard, with applications opening in August.