Agent Data Pod Vocabulary

RDF classes and properties for AI agent data in Solid Pods.

Namespace: https://awkronos.github.io/web/vocab# · Version 0.2.0 · Download .ttl

Overview

This vocabulary defines classes and properties for representing AI agent data in RDF. It extends FOAF for agent identity and PROV-O for audit trails.

Prefix: agent: = <https://awkronos.github.io/web/vocab#>

Classes defined in the Agent Data Pod vocabulary
Class	Subclass Of	Description
`agent:AIAgent`	`foaf:Agent`	An AI agent with a Solid Pod profile
`agent:MemoryEpisode`	`prov:Entity`	A unit of agent memory (fact, event, preference)
`agent:Capability`	—	An action the agent can perform
`agent:CapabilitySet`	—	Collection of agent capabilities
`agent:AuditEntry`	`prov:Activity`	Record of a modification to agent data
`agent:AgentDescriptionService`	—	ANP Agent Description service endpoint

Properties for agent profiles
Property	Domain	Range	Description
`agent:mcpEndpoint`	`agent:AIAgent`	`xsd:anyURI`	Model Context Protocol endpoint URL
`agent:capabilities`	`agent:AIAgent`	`agent:CapabilitySet`	Link to capability declarations

Properties for memory episodes
Property	Domain	Range	Description
`agent:content`	`agent:MemoryEpisode`	`xsd:string`	Human-readable episode content
`agent:importance`	`agent:MemoryEpisode`	`xsd:decimal`	Relevance score (0.0-1.0)
`agent:memoryType`	`agent:MemoryEpisode`	`xsd:string`	episodic, semantic, procedural, emotional, reflective
`agent:tag`	`agent:MemoryEpisode`	`xsd:string`	Categorical tags for retrieval
`agent:embedding`	`agent:MemoryEpisode`	`xsd:base64Binary`	Vector embedding (float32 little-endian)
`agent:embeddingModel`	`agent:MemoryEpisode`	`xsd:string`	Embedding model identifier
`agent:embeddingDim`	`agent:MemoryEpisode`	`xsd:integer`	Embedding vector dimensions
`agent:embeddingFormat`	`agent:MemoryEpisode`	`xsd:string`	Format specification (e.g., "float32-le")
`agent:chunkingStrategy`	`agent:MemoryEpisode`	`xsd:string`	Text chunking method used

Properties for capability declarations
Property	Domain	Range	Description
`agent:capability`	`agent:CapabilitySet`	`agent:Capability`	Link to individual capability
`agent:name`	`agent:Capability`	`xsd:string`	Machine-readable capability name
`agent:description`	`agent:Capability`	`xsd:string`	Human-readable description
`agent:requiredPermission`	`agent:Capability`	`xsd:anyURI`	ACL permission required
`agent:protocol`	`agent:Capability`	`xsd:anyURI`	Invocation protocol URI

Properties for audit trail entries
Property	Domain	Range	Description
`agent:contentHash`	—	`xsd:string`	RDFC-1.0 canonicalized content hash
`agent:signature`	—	`xsd:base64Binary`	Cryptographic signature
`agent:signedBy`	—	`xsd:anyURI`	Signer's WebID or DID

This vocabulary imports: