Communication Routing

Purpose

Ensure efficient information flow between agents by routing requests, sharing artifacts, broadcasting updates, and maintaining communication records for audit and context.

When to Use

• Agent requests information from another agent
• Artifact handoffs between phases
• Broadcast announcements needed
• Cross-functional coordination required
• Context sharing for related tasks

Prerequisites

• Agent registry loaded
• Communication protocols defined
• Message queue available
• Audit logging enabled

Process

Step 1: Receive Communication Request

Parse incoming message:
- Source agent
- Target agent(s) or broadcast
- Message type (request/response/notification)
- Subject and payload
- Priority level
- Response required (Y/N)

Step 2: Validate Routing

Validation checks:
1. Source agent authorized to send
2. Target agent exists and active
3. Message type appropriate
4. Required fields present
5. Payload format valid

Step 3: Determine Route

Routing rules:
- Direct: Single target agent
- Multicast: Specific agent group
- Broadcast: All agents
- Escalation: Human involvement

Route selection based on:
- Message type
- Target specification
- Urgency level

Step 4: Deliver Message

Delivery process:
1. Add routing metadata
2. Log message in audit trail
3. Deliver to target(s)
4. Confirm receipt
5. Track response if required

Step 5: Handle Response

If response required:
1. Set response timeout
2. Monitor for response
3. Route response to original sender
4. Escalate if timeout exceeded

Inputs

Input	Type	Required	Description
message	JSON	Yes	Communication content
source_agent	String	Yes	Sending agent ID
target	String/Array	Yes	Recipient(s)
priority	String	Optional	urgent/normal/low

Outputs

Output	Type	Description
delivery_confirmation	JSON	Delivery status
audit_record	JSON	Communication log
response	JSON	If response received

StudyAbroad-Specific Considerations

• Security agent notifications are always high priority
• External API changes broadcast to architecture + developer
• GDPR-related communications logged for compliance
• Test manager receives all requirement changes

Integration Points

• All Agents: Send and receive communications
• Audit System: Log all communications
• Human: Escalation endpoint

Examples

# Developer needs API spec from Design Agent
Message:
{
  "from": "developer-agent",
  "to": "design-agent",
  "type": "request",
  "subject": "OpenAPI spec for profile endpoints",
  "payload": {
    "endpoints": ["/api/user/profile", "/api/user/settings"],
    "format": "yaml"
  },
  "priority": "normal",
  "response_required": true
}

Routing:
→ Validated: developer-agent authorized
→ Route: Direct to design-agent
→ Delivered: 2024-01-15T10:30:00Z
→ Response received: 2024-01-15T10:32:15Z
→ Routed response to developer-agent

# Broadcast: Requirement change
Message:
{
  "from": "requirements-agent",
  "to": "broadcast",
  "type": "notification",
  "subject": "REQ-015 modified",
  "payload": {
    "requirement_id": "REQ-015",
    "change_type": "acceptance_criteria_updated",
    "summary": "Added email verification step"
  }
}

Routing:
→ Broadcast to: all agents
→ Special routing: test-manager (impact analysis trigger)

Validation

• All messages logged in audit trail
• Delivery confirmation received
• No messages lost or undelivered
• Response timeouts handled
• Routing rules correctly applied