Home/Agentic AI/Agent Communication/Message Structure

Agent Communication Protocols

Learn how AI agents exchange information and coordinate actions

Your Progress

0 / 5 completed
1
2
3
4
5

Anatomy of Agent Messages

Well-structured messages are the foundation of effective agent communication. Every message contains several key fields that define its purpose, origin, destination, and payload.

Interactive: Message Field Explorer

Click each field to explore its role in agent messages:

🎯

Performative

The intent or action of the message

Purpose:

Tells the receiver what kind of action to expect

Examples:
REQUEST
INFORM
QUERY
AGREE
REFUSE
CONFIRM

Complete Message Example

// Real agent message structure:
{
"performative": "REQUEST",
"sender": "coordinator-agent",
"receiver": "worker-agent-42",
"content": "{ "action": "process_dataset", "dataset_id": "xyz-123" }",
"language": "JSON",
"reply-to": "coordinator-agent",
"ontology": "task-management",
"protocol": "contract-net"
}

✓ Required Fields:

  • • Performative (intent)
  • • Sender (source agent)
  • • Receiver (target agent)
  • • Content (message data)

⭐ Optional Fields:

  • • Language (content format)
  • • Reply-to (response address)
  • • Ontology (domain context)
  • • Protocol (conversation type)

🎯 Design Principles

  • •Be explicit: Include all necessary fields for clarity
  • •Use standard performatives: Stick to recognized intents (REQUEST, INFORM, etc.)
  • •Validate content: Ensure payload matches the declared language format
  • •Include metadata: Protocol, ontology, and reply-to improve coordination
←Prev