agent-platform-model-registry

You can manage machine learning models in the Agent Platform Model Registry by automating model discovery, version registration, and metadata updates to eliminate manual CLI errors. This skill enforces tiered confirmation safeguards for all mutating operations.

Before managing the model registry, you need pre-configured Google Cloud Application Default Credentials and active project settings. These environment configurations are required to successfully execute registry commands.

Yes, you can securely delete unused models and all their versions from the model registry. The skill enforces explicit confirmation steps for destructive operations to prevent irreversible data loss.

You can perform read-only model discovery to list all registered models and retrieve full metadata for specific versions. This allows quick auditing of your machine learning models without risking accidental changes.

The best way to update metadata like display names and labels for existing model versions is through automated safe model uploads. Built-in confirmation safeguards prevent unintended modifications during the process.

You should use automated model version registration when maintaining multiple ML models for Agent Platform use cases. It prevents human error during frequent uploads, metadata updates, and deletions across multiple versions.