gke-compute-classes
Pass
Audited by Gen Agent Trust Hub on Oct 19, 2026
Risk Level: SAFE
Full Analysis
- Data Processing Security: The skill is designed to analyze cluster logs and configurations, which are ingestion points for external data. It incorporates a defensive 'Critical Injection Rule' advising the agent to treat such content as untrusted and to ignore any embedded directives, effectively addressing potential indirect prompt injection considerations.
- Trusted Resource Integration: For advanced troubleshooting, the skill directs the agent to the official cluster-autoscaler open-source repository. This reference is to a well-known, vendor-maintained source, facilitating accurate verification within a secure framework.
- Diagnostic Script Safety: The provided monitoring script for GKE autoscaler events utilizes standard CLI utilities (gcloud, jq) within the user's established authentication context. This approach enables necessary diagnostic capabilities while maintaining a consistent security posture.
Audit Metadata