An agent artefact is a produced output or result generated through agent execution. The Agent Ontology declares it as Artifact, using American spelling for the identifier.1 visionGraph places it under the AI Research Area hub.

An Agent Action yields an artefact through producesArtifact. The only property of the class itself is description, a human-readable string. Format, storage and integrity are not fixed, so a file, a message and a decision record all qualify.

The artefact is the outcome end of the chain from Agent Intent to action: an intent leads to an action, and the action leads to an artefact. The action itself may be recorded separately as an Agent Trace Event, which describes that the action happened, while the artefact is what it left behind.

Footnotes

  1. Semantic Agent Communication Community Group, “Core Agent Ontology” (Agent Ontology, CC BY 4.0). https://github.com/s-agent-comm/agent-ontology/blob/750e06d276dd568313a6c03c4e78203887d3fa4b/ontologies/core.ttl ↩