Skip to content
Get packs

Agent Platform Model Registry Manager

Lets an agent list, describe, upload, update, and delete ML models in Google's Agent Platform Model Registry with safety confirmations.

Source Download skill
Repository
google/skills
Stars
21,782
Licence
Apache-2.0
Files
1 file

What it does

Once installed, the agent knows how to manage machine learning models stored in the Agent Platform Model Registry on Google Cloud. It can list models, look up details, upload new models or versions, update metadata, and delete models, using gcloud and Python SDK commands. The skill imposes strict safety rules: read-only actions run immediately, uploads and updates require a literal command preview and a 'Yes' reply in a later turn, and deletions require a typed confirmation in a separate turn.

Great for

  • ML engineers who maintain a model catalog in Google Cloud and want help with routine registry tasks
  • Teams that want an agent to avoid accidental model uploads or deletions through enforced confirmation steps
  • Developers who already use gcloud and have access to an Agent Platform project

Skip it if

  • Model training or deploying models to endpoints, which the skill explicitly excludes (use a different skill for those)
  • Users without a Google Cloud project and gcloud access, since every command needs a project and region
  • Anyone expecting the agent to delete models without review: deletion is irreversible and the skill is built to stop and ask first

Try asking

  • “List the latest models in my registry in us-central1 and show their names”
  • “Describe model 1234567890 and tell me its version and creation date”
  • “Upload this model file as a new version with display name 'churn-model-v2' after showing me the exact command to approve”
  • “Delete the model with ID 9876543210 from my registry (I will type 'I confirm' to approve)”
Step-by-step guide: how to use Agent Platform Model Registry ManagerInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 1 original files, the Apache-2.0 licence and an attribution note.

    agent-platform-model-registry.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip agent-platform-model-registry.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip agent-platform-model-registry.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Always state the project and region up front so the agent does not need to look them up
  • Ask the agent to list models first if you do not know the numeric model ID, since update and describe need it
  • Read the exact gcloud command in each confirmation card before replying 'Yes', because changed parameters require a new confirmation

What's inside

  • SKILL.md

Redistributed unmodified from google/skills under the Apache-2.0 licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.