From oracle-ai-data-platform-workbench-engineer-agent
Lists models available/installed in the AIDP DataLake, inspects their parameters, and browses the MLOps model registry. Essential for resolving model names before using ai_generate() or agent flows.
How this skill is triggered — by the user, by Claude, or both
Slash command
/oracle-ai-data-platform-workbench-engineer-agent:aidp-models-catalogThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Two distinct surfaces:
aidp-models-catalog — available models & parametersTwo distinct surfaces:
ai_generate() / agent flows (LA Models API).aidp-mlops group; see also aidp-mlops).CLI (preferred) — registry browse: aidp mlops <command> --instance-id <DATALAKE_OCID> --auth api_key --profile DEFAULT --region <r>
aidp mlops list-registered-models | get-registered-model | list-model-versionsmlops group — there is no separate models group.)Fallback (no CLI) / platform catalog: the LA Models REST API via oci raw-request (identical
endpoint + auth) — this is the verified path for resolving an ai_generate() model name.
Verify-first (no-fabrication): the platform
ModelsAPI is Limited Availability on20240831; the?modelType=list call is LIVE-VERIFIED 200 below. The MLOps registry side is Preview (not yet probed). Confirm a name with a live read before relying on it; record inreferences/rest-endpoint-map.md.
ai_generate?", "show model parameters",
"list registered models / model versions".20240831)Base: https://aidp.<region>.oci.oraclecloud.com/20240831/dataLakes/<DATALAKE_OCID>/…
GET /models?modelType=<GENERATIVE_AI|BASE|EMBEDDING|LLM> — list installed models (LIVE-VERIFIED 200)GET /models/{id} — model detailGET /models/{id}/modelParameters — parametersGET /models?modelType=GENERATIVE_AI (auth ladder; on 401/403 → session refresh)
→ present available model names/ids. Repeat with BASE / EMBEDDING / LLM to enumerate other catalogs.
GET /models/{id} / …/modelParameters for detail.aidp mlops list-registered-models → list-model-versions (CLI preferred).aidp-ai-sql (ai_generate('<model>', …)) or aidp-agent-flows.oci raw-request --http-method GET \
--target-uri "https://aidp.us-ashburn-1.oci.oraclecloud.com/20240831/dataLakes/<DATALAKE_OCID>/models?modelType=GENERATIVE_AI" \
--profile DEFAULT
ai_generate rather than guessing one.GET /models may 400 on a required param — pass modelType (verified above) to enumerate.ai_generate availability is independent of this catalog. GET /models?modelType=GENERATIVE_AI
can return items: [] on a fresh instance while ai_generate('openai.gpt-5.4', …) still executes fine —
the model is resolved at the Spark engine level, not via this REST catalog. An empty list here does
not mean ai_generate is unavailable. Confirm AI-in-SQL with the smoke test in aidp-ai-sql
(a trivial SELECT ai_generate('<model>', 'hello') cell), not with this endpoint.aidp-ai-sql, aidp-agent-flows, aidp-mlopsnpx claudepluginhub anthropics/claude-plugins-official --plugin oracle-ai-data-platform-workbench-engineer-agent2plugins reuse this skill
First indexed Jun 12, 2026
Tracks ML experiments, runs, metrics, parameters, and model versions in AIDP's MLflow-compatible MLOps. Use when logging/tracking experiments, registering models, or transitioning model-version stages. CLI-driven; verify live endpoints first.
Deploys MLflow models, custom pyfunc, and GenAI agents to Databricks Model Serving endpoints. Queries endpoints, checks status, integrates UC Functions and Vector Search tools.
Manages platform LLM model access and API keys for AI agents. Browse models, configure access, check rate limits, and get OpenAI-compatible endpoint.