Deploy and clean up Model Garden endpoints safely.
model deployment
Deploys Model Garden models to Agent Platform endpoints, checks deployment status and cost, and cleans up deployed resources.
When to use it
Use for active deployments, catalog listing, deployability and cost checks, deployment troubleshooting or status checks, and resource cleanup.
Give it a deployment or cleanup request; it confirms mutating actions, changes Agent Platform resources, and reports the result.
This skill
Agent Platform
Deploys and deletes cloud resources (irreversible)
deployed Agent Platform endpoint
Sends your test prompt
Agent Platform publisher endpoint
Sends a model availability probe
Google Cloud CLI configuration
Sets your default cloud project
Requires gcloud for authentication, Model Garden discovery, operation status checks, and endpoint management.
Requires authenticated Google Cloud access.
Requires a destination project and region.
Requires Python 3 to run the mandatory first-choice hourly cost estimator.