Schema reference
Every entry in this registry validates against one of the JSON Schema documents below. entry-core.v1.schema.json defines the fields every entry carries regardless of type; each content-type schema then extends it, via allOf, with fields specific to that type. These pages describe the schemas as they exist; each page links to the raw schema file on the repository as the authoritative source.
common entry core (v1)
Fields carried by every entry regardless of type (internal project documentation Section 3.2).
agent entry (v1)
The common entry core plus a pointer to the agent's live agent-info endpoint and a cached snapshot of its registry-relevant fields, mirrored verbatim under the endpoint's own field names (internal project documentation Section 3.5.2).
bundle entry (v1)
The common entry core plus an embedded, verbatim Claude Code marketplace.json document, validated against schemas/vendor/marketplace.schema.json, plus a ga4gh sibling object recording provenance and any cross-references to entries already catalogued elsewhere in this registry.
MCP server entry (v1)
The common entry core plus an embedded, verbatim, valid server.json document, validated against schemas/vendor/server.schema.json (internal project documentation Section 3.5.4).
model entry (v1)
The common entry core (entry-core.v1.schema.json) plus a required pointer to a full GA4GH Genomic AI Model Card and a small index projection mirrored from the Model Card and from Hugging Face's own listing conventions (internal project documentation Section 3.5.1).
coding-harness plugin entry (v1)
The common entry core plus an embedded, verbatim Claude Code plugin.json document, validated against schemas/vendor/plugin.schema.json, plus a ga4gh sibling object recording which coding harnesses the plugin targets and how to install it for each (internal project documentation Section 3.5.5).
skill entry (v1)
The common entry core plus an embedded SKILL.md-style frontmatter object and the fields Project #24 specifies as required for a certifiable skill: defined inputs, outputs, provenance logging, human oversight requirements, and evaluation criteria (internal project documentation Section 3.5.3, internal project documentation Section 6.4).