The model is open and extensible through Linked Data. Credentials, presentations and nested objects should convey their types; new terms must each have a URL with published human-readable documentation and a JSON-LD context mapping them, and reuse of existing terms is recommended. Documents should use contexts defining all their terms rather than relying on the default vocabulary; where undefined terms are used, the https://www.w3.org/ns/credentials/undefined-terms/v2 vocabulary applies.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.