Cisco Donates AGNTCY Project to Linux Foundation for Open AI Agent Collaboration

Cisco has donated its AGNTCY project to the Linux Foundation, aiming to establish an open standard for AI agent collaboration. The initiative is supported by major tech companies like Google Cloud, Dell, Oracle, and Red Hat, with the goal of creating an ‘Internet of Agents’ where AI agents from different vendors can work together seamlessly. Industry experts argue that such a standard would resolve the fragmented landscape of AI agents, which are currently siloed and unable to interact effectively. The Linux Foundation, a well-known hub for open-source innovation, will oversee the project’s neutral governance, aiming to make AGNTCY a widely adopted industry standard.

AGNTCY provides a comprehensive framework for addressing the core challenges of multi-agent collaboration. The project includes four key components: Agent Discovery through its Open Agent Schema Framework (OASF), which acts like a ‘DNS for agents,’ enabling them to locate and recognize each other. Agent Identity is secured through a cryptographic system that allows agents to verify their identities and perform actions across vendors. The messaging protocol, Secure Low-latency Interactive Messaging (SLIM), is designed for complex, multi-modal interactions between agents, with quantum-safe security built in. Lastly, the observability framework provides visibility into multi-agent workflows, crucial for debugging AI systems that operate probabilistically.

The project is positioned to integrate with other emerging AI agency standards such as the Agent2Agent (A2A) protocol and Anthropic’s Model Context Protocol (MCP). AGNTCY aims to enable interoperability and collaboration among these various standards. For instance, agents using the A2A protocol and servers employing MCP can be registered and discovered through AGNTCY directories. The SLIM protocol can transport communications from A2A and MCP, and observability tools within AGNTCY can monitor interactions between different protocols, enhancing transparency and simplifying the debugging of complex workflows. The project’s code and documentation are available on GitHub for public access and contribution.

This move reflects a broader industry effort to standardize AI agent technologies, which could significantly impact the development and deployment of such systems. By creating a common framework, Cisco and its collaborators hope to reduce redundancy and inefficiencies while fostering innovation and broader adoption. The Linux Foundation’s role in this initiative underscores its importance in shaping the future of open-source technology and its potential influence on the AI landscape.