.webp)
New standards document is open for public comment as organizations seek shared approaches to securing and verifying autonomous AI systems
MELBOURNE, Fla., October 5, 2026 — OpenMatter Network today announced the release of a proposed standards document, Zero-Knowledge Boundary Compliance for Autonomous Agents, for public comment, the latest milestone in its work with Hashgraph Online (HOL) to advance standards for AI compliance and agentic security.
The draft standards document proposes a shared framework for demonstrating that AI agents follow established rules without exposing the sensitive data they use, helping organizations secure AI agent activity across systems and organizational boundaries.
For example, a company could require an AI assistant to remove private customer details before sending information to an outside service. A system built to the proposed standards could block the request if those details remain and provide proof that permitted requests followed the rule, without showing an auditor the private information.
As AI agents take on more tasks and interact with sensitive information, organizations need ways to establish what agents are permitted to do and verify that they operate within those limits. The standard addresses how organizations can control what AI agents access, share and do; prevent unauthorized actions or disclosure of sensitive information; and create a reliable record of the rules governing an agent’s activity. It sets out proposed requirements for applying those rules when agents interact with users, data or external systems, blocking actions that violate them, and producing independently verifiable evidence of compliance without exposing sensitive information.
OpenMatter Network joined the HOL Partner Program in July 2026 and has since worked alongside HOL, Vera Anchor, and other members of the AI Privacy & Security subcommittee to develop the proposed standards for verifiable compliance and agentic security. The work draws on OpenMatter Network’s expertise in cryptographic verification and secure AI collaboration to establish a common framework for defining how AI agents operate and verifying their compliance with established policies.
“AI agents are beginning to operate across systems and organizational boundaries, but the rules for securing and governing those interactions are still taking shape,” said Renee Davis, Chief Business Officer and Co-Founder of OpenMatter Network. “Our work with HOL is focused on giving organizations a shared framework for defining what agents are allowed to do and verifying that they follow those rules. Releasing this document for public comment is an important step toward making that framework practical and useful across the industry.”
“AI agents will not scale safely if trust stops at the boundary of a single platform or organization. We need open standards that let organizations define what an agent is allowed to do, enforce those boundaries, and produce verifiable evidence that the rules were followed without exposing sensitive data. That is exactly the kind of practical, interoperable infrastructure HOL was created to advance, and our work with OpenMatter is an important step toward making it real,” said Michael Kantor, President of HOL.
“At Vera Anchor, we focus on creating durable evidence of digital actions, data, and computational results that can be verified independently of the systems that produced them. For agent governance, defining what an agent is permitted to do is only part of the problem. Organizations also need reliable evidence of what actually occurred across systems and organizational boundaries. Open standards are essential for keeping that evidence portable and independently verifiable beyond the originating platform,” said Andrew McClure, Founder of Vera Anchor.
OpenMatter Network and HOL invite developers, enterprises, security researchers and others working on AI agent standards to review the document and submit comments by October 31, with the possibility of extending the comment period. The document is available on GitHub here, and feedback can be commented directly on the pull requests tab.
Following the public comment period, the working group will review submitted feedback, document the resolution of substantive comments and publish a revised draft. Proposed changes will be evaluated for technical interoperability, privacy, security and implementability, with updated conformance requirements and test vectors incorporated where appropriate before the specification advances through the applicable HOL governance process.
For more information on OpenMatter Network, click here. For more information on HOL, click here.
About HOL
Hashgraph Online (HOL) is an open-source ecosystem for interoperable AI agents. HOL develops standards, SDKs, registries, and developer tools that help organizations identify, discover, verify, and coordinate agents across web and decentralized environments. Its ecosystem includes HOL Guard, a security platform that helps developers and organizations reduce risk when deploying and operating AI agents. Through the HOL Partner Program, HOL collaborates with organizations advancing open standards and infrastructure for the agent ecosystem. For more information, visit https://hol.org.
About OpenMatter Network
Headquartered in Florida’s Space Coast, OpenMatter Network is building the Verifiable Trust Layer for Secure Collaboration and AI Agents. Guided by the principle "Don't Trust Data. Prove It.," the company's cryptographically verifiable architecture enables secure collaboration, governed AI behavior and mathematically verifiable execution across untrusted environments. For more information, visit www.openmatter.network.
Rate us on G2
