Lyrie.ai Joins Anthropic Cyber Verification Program, Releases ATP

According to DevOps.com and CyberNews press releases, OTT Cybersecurity LLC, the company behind Lyrie.ai, announced two reported milestones: acceptance into Anthropic's Cyber Verification Program (CVP) and the public release of the Agent Trust Protocol (ATP). The releases state the ATP specification is available at lyrie.ai/research and that a reference implementation is published under an MIT license on GitHub at github.com/OTT-Cybersecurity-LLC/lyrie-ai. The announcements add that ATP is slated for submission to the IETF. The press materials include direct quotes from Guy Sheetrit, CEO and Founder of OTT Cybersecurity LLC, saying, "Being among the first companies accepted into Anthropic's Cyber Verification Program validates what we've built. Lyrie isn't a security tool that sits alongside AI. It's the security layer that AI runs on top of."

The ATP specification, as described in the releases, defines five cryptographic primitives:

• •Identity, who the AI agent is
• •Scope, what the agent is authorized to do
• •Attestation, whether the agent or its instructions have been tampered with
• •Delegation, who delegated authority
• •Revocation, whether authority has been revoked

The releases state ATP is open and royalty-free and that a reference implementation is available under an MIT license on GitHub.

DevOps.com and CyberNews frame these milestones as positioning Lyrie.ai as foundational infrastructure for an agentic AI era where autonomous agents perform sensitive actions. The reporting highlights that enterprises and governments are deploying autonomous agents that can read mail, write code, move money, and sign contracts, and that a scalable security model for those agents is claimed to be missing.

Standards for cryptographic identity and scoped authority are a recurring industry response to the operational risks posed by agentic systems. Companies and open groups proposing royalty-free protocols and publishing reference implementations under permissive licenses commonly aim to lower integration friction and encourage third-party audits. Observed patterns in similar standardization efforts show that publication and a permissive license accelerate experimentation but do not guarantee broad interoperability or enterprise adoption without ecosystem buy-in and enforcement points.

For practitioners and security teams, useful indicators will include whether ATP gains implementers beyond the originating project on GitHub, whether the ATP draft is actually submitted and progresses within the IETF, and whether Anthropic's CVP publishes participant lists or verification criteria. Activity on the repository, independent security reviews, and early enterprise integrations are practical signals of adoption.