The AI Agent Protocol Stack is Missing a Layer — Here's the Third One
📰 Dev.to AI
Learn about the missing layer in the AI Agent Protocol Stack and how it addresses agent permissioning
Action Steps
- Read the Cisco Research paper on arXiv:2511.19699 to understand the proposed layered agent protocol architecture
- Identify the limitations of existing protocols like MCP and A2A in terms of agent permissioning
- Design a new layer to address the missing aspect of agent permissioning, building on the concepts of L8 and L9
- Implement a prototype of the new layer using standardized message envelopes and speech-act performatives
- Test the prototype with various agent scenarios to evaluate its effectiveness
Who Needs to Know This
AI engineers and researchers working on agent-based systems can benefit from understanding the proposed layered architecture to improve agent communication and negotiation
Key Insight
💡 The missing layer in the AI Agent Protocol Stack is crucial for agent permissioning, enabling agents to understand what they're allowed to do
Share This
🤖 AI Agent Protocol Stack is missing a layer! 📚 Cisco Research proposes a new architecture to address agent permissioning #AI #AgentProtocol
Key Takeaways
Learn about the missing layer in the AI Agent Protocol Stack and how it addresses agent permissioning
Full Article
MCP connects agents to tools. A2A connects agents to agents. Both are excellent. Neither answers one question: How do agents know what they're ALLOWED to do? Cisco Research proposed a layered agent protocol architecture in 2025 (arXiv:2511.19699): L8 (Agent Communication Layer) → Standardized message envelopes, speech-act performatives L9 (Agent Semantic Negotiation Layer) → "does not exist today"</em
DeepCamp AI