---

name: compute-notebooks-and-images description: "Route CubeStudio notebook, registry, image catalog, GPU resource, and monitoring questions." metadata: disco-role: operating disable-model-invocation: true license: NOASSERTION

Compute notebooks and images

Use this sub-skill for CubeStudio questions about:

• project, resource group, and cluster placement
• notebook / IDE lifecycle and URLs
• Docker image registry management and online image builds
• notebook, Theia, Jupyter, and GPU image catalogs
• GPU resource strings, selector switching, and monitoring / resource views

Do not use this sub-skill for:

• platform install, Kubernetes bring-up, or offline cluster setup → deploy-and-operate
• pipeline or job-template authoring → pipelines-and-job-templates
• serving / inference image deployment → serving-aihub-and-llm
• generic backend or FAB customization → backend-and-configuration

Read first:

• references/resource-and-kubernetes-api.md
• references/notebook-workflows.md
• references/image-catalog.md
• references/troubleshooting.md
• scripts/parse_resource_gpu.py

Router notes:

• notebook placement is driven by project / org / cluster config, resource_gpu, and selector labels
• resource_gpu strings are authoritative only when they pass the bundled parser
• notebook / IDE lifecycle questions map to create, open, reset, renew, stop, and save behavior
• image registry questions map to Repository, Images, Docker, and NOTEBOOK_IMAGES
• monitoring questions map to Grafana, Prometheus, total_resource, and GPU plugin caveats

When a request needs platform installation or image families outside notebooks, route it to the sibling sub-skill instead of expanding this one. Do not require the original checkout in answers; use the bundled references and helper script only.