registryajp:AgentCollaboration
AJP AI AGENT SCHEMA STANDARD

ajp:AgentCollaboration

extends schema:Thing

A structured description of how two or more AI agents work together within a multi-agent system. Defines the orchestration hierarchy, data flow, and human oversight for the collaboration. Enables machine-readable representation of conductor/worker patterns and org-board architectures.

REQUIRED PROPERTIES (6)
ajp:collaborationIdTextrequired

Unique identifier for this collaboration

ajp:collaborationNameTextrequired

Human-readable name

ajp:collaborationTypeTextrequired

hierarchical / parallel / sequential / consensus

ajp:workerAgents[ajp:AIAgent]required

Agents participating in this collaboration

ajp:dataSharedBetweenAgentsTextrequired

Description of data passed between agents

ajp:humanApprovalRequiredBooleanrequired

Does collaboration require human approval before acting?

OPTIONAL PROPERTIES (7)
ajp:orchestratorAgentajp:AIAgent

The orchestrating agent (required for hierarchical)

ajp:maxAgentsActiveInteger

Maximum agents active simultaneously

ajp:humanApprovalGateajp:HumanApprovalGate

Gate definition (required if humanApprovalRequired)

ajp:combinedCapabilityDeclarationajp:CapabilityDeclaration

What the collaboration as a whole can do

ajp:combinedProhibitions[ajp:ProhibitedAction]

What the collaboration cannot do

ajp:spanLimitInteger

Max worker agents per orchestrator (default: 5)

ajp:auditTrailajp:AgentAuditRecord

Combined audit for the collaboration

EXAMPLE JSON-LD
{
  "@context": {
    "@vocab": "https://schema.org/",
    "ajp": "https://schema.anthonyjamespeacock.com/"
  },
  "@type": [
    "ajp:AgentCollaboration"
  ],
  "ajp:collaborationId": "<Text>",
  "ajp:collaborationName": "<Text>",
  "ajp:collaborationType": "<Text>"
}
MACHINE ENDPOINT
GET /api/schema/AgentCollaboration
Accept: application/ld+json
View raw JSON-LD →
OTHER TYPES