DUBAI, UAE — May 11, 2026 — As the digital landscape evolves from chatbot experimentation to a network of independent AI agents, the identity of these agents becomes a primary concern. The focus has shifted from capabilities to the question of identity management within AI.
Today marks a significant development with the launch of the Agent Trust Protocol (ATP) by OTT Cybersecurity LLC. This initiative, now recognized by Anthropic’s Cyber Verification Program (CVP), serves as a foundational security layer for the emerging agentic economy.
The Challenge of Anonymity
Currently, AI agents are entrusted with critical tasks such as coding, financial transactions, and contractual agreements, often operating without identifiable credentials. Guy Sheetrit, CEO of OTT Cybersecurity LLC, highlights the risks of this anonymity, likening it to allowing unknown individuals unrestricted access to sensitive areas.
Sheetrit emphasizes that ATP provides AI agents with a verifiable identity and clearly defined responsibilities, ensuring they cannot operate under false pretenses.
Understanding the Agent Trust Protocol
Developed by the Lyrie research team, ATP is an open-source cryptographic standard akin to the security provided by HTTPS for internet communications. It is currently under review by the IETF for global application.
The protocol incorporates five fundamental components: identity verification, scope limitation, attestation, delegation, and revocation. These elements collectively ensure that AI agents operate within secure and authorized parameters.
Collaboration with Anthropic
The inclusion of ATP in Anthropic’s Cyber Verification Program signifies an important collaboration for Lyrie. This alliance enables comprehensive vulnerability assessments and red-teaming within Claude’s infrastructure, advancing the responsible use of dual-purpose cybersecurity tools.
Through this partnership, Lyrie is actively involved in shaping the safe deployment of cybersecurity solutions that can both defend and challenge AI systems.
The Lyrie Ecosystem: Comprehensive Security Solutions
Lyrie.ai extends beyond protocol development, offering a robust security platform to combat and simulate potential AI threats. Key features include Lyrie Hack, a pentesting engine, GPU-accelerated red teaming, and the Omega-Suite for advanced binary analysis.
These innovations align with the 2026 Agentic Security Initiative, ensuring comprehensive protection against evolving AI threats.
Promoting Global Innovation
In alignment with its mission for a secure internet, OTT Cybersecurity LLC has made the ATP reference implementation publicly available under the MIT License. Developers worldwide can access it on GitHub to foster innovation and collaboration.
OTT Cybersecurity LLC, based in Dubai, is the parent entity behind Lyrie.ai. Their team of expert security researchers is committed to establishing a secure framework for autonomous AI in enterprise environments.
For further information, media inquiries can be directed to Guy Sheetrit, CEO, at [email protected] or through their official GitHub and website.
