agent-platform-model-registry
Pass
Audited by Gen Agent Trust Hub on Nov 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [Command Execution]: The skill uses
gcloudandpython3to interact with model registries. These are standard administrative actions for the intended purpose of the skill. The instructions include explicit requirements for user confirmation for any action that modifies or deletes resources (Tier M and Tier D safety tiers). - [Usage Attribution]: The skill includes instructions to include attribution headers (e.g.,
CLOUDSDK_METRICS_REQUEST_ATTRIBUTION) ingcloudcommands and HTTP requests. This is a common telemetry practice for internal tracking and does not involve sensitive user data exfiltration. - [Environment Grounding]: The skill implements best practices for environment setup, ensuring that the agent resolves project and region parameters before execution and avoids using dangerous 'global' flags where specific scopes are required.
Audit Metadata