| name | ontoforge-runtime-api |
| description | Use when the user needs to build against the OntoForge runtime REST API: curl calls, API clients, request payloads, query parameters, or endpoint selection under /api/runtime. This skill is runtime-only and does not cover the modeling API. |
Goal
Help the user build correct calls against the OntoForge runtime REST API only.
Do not use this skill for schema design or modeling endpoints under /api/model/....
Primary Reference
Read runtime-api.md first. It is the bundled runtime API reference for this skill.
Workflow
- Map the user request to the correct endpoint family in
references/runtime-api.md.
- Use the bundled endpoint definitions there as the source of truth for path shape, request body, and query parameters.
- Treat
/api/runtime/{ontologyKey} as the base path for ontology-scoped calls. The only non-scoped runtime endpoint is GET /api/runtime/features.
- If entity type keys, relation type keys, or property keys are unknown, start with schema introspection endpoints before generating write or query calls.
- Preserve documented filter syntax exactly:
filter.{key} and filter.{key}__{op}.
- Set
Content-Type: application/json on JSON POST and PATCH requests.
Boundaries
- Runtime only: schema introspection, entity CRUD, relation CRUD, neighbors, semantic search, OQL query, feature discovery, AI runtime endpoints, saved queries, and runtime data wipe.
- Not modeling: ontology creation, schema mutation, saved query definition, AI agent configuration, import/export via
/api/model/....
Output Style
When answering:
- name the exact endpoint first
- include required path parameters
- include relevant query parameters
- include the JSON body when applicable
- then produce the requested artifact:
curl, client code, fetch wrapper, SDK helper, or typed interface
