agent-platform-tuning-management

Pass

Audited by Gen Agent Trust Hub on Sep 18, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses gcloud auth login and gcloud auth application-default login to establish standard authentication with Google Cloud services.
  • [EXTERNAL_DOWNLOADS]: The skill installs the google-cloud-aiplatform package via pip. This is the official SDK for Google Cloud Vertex AI and is a standard dependency for cloud management tasks.
  • [DYNAMIC_EXECUTION]: Provides static Python code snippets that interact with the google-cloud-aiplatform SDK. These snippets are used to list, get, and cancel tuning jobs via official API endpoints.
  • [INDIRECT_PROMPT_INJECTION]: While the skill ingests job metadata (such as model names and states) from the Google Cloud API, it implements a robust safety framework. It defines 'Tier D' actions (destructive) that mandate explicit typed confirmation from the user before the agent can execute a cancellation command, effectively mitigating risks associated with untrusted data influencing agent actions.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 18, 2026, 05:40 AM
Security Audit — agent-trust-hub — agent-platform-tuning-management