kubernetes
Pass
Audited by Gen Agent Trust Hub on Jul 30, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides structured guidance for Kubernetes debugging and security hardening, emphasizing industry best practices such as least-privilege RBAC, NetworkPolicies, and secure secret management via external providers.\n- [COMMAND_EXECUTION]: Instructions utilize standard kubectl binaries for administrative tasks, including log retrieval, resource monitoring, and pod status diagnostics.\n- [EXTERNAL_DOWNLOADS]: The skill references container images hosted on well-known registries (ghcr.io, gcr.io) and standard Go-based dependencies for the kvm-exporter component, all originating from trusted or well-known development sources.\n- [PROMPT_INJECTION]: The skill is designed to ingest and analyze pod logs and cluster events, which represents an inherent diagnostic attack surface for indirect prompt injection.\n
- Ingestion points: Log streams (kubectl logs), resource events (kubectl get events), and status descriptions (kubectl describe).\n
- Boundary markers: Instructions do not specify delimiters for isolating external log content from the agent's instructions.\n
- Capability inventory: Includes capabilities for pod command execution (kubectl exec), node debugging (kubectl debug), and network port forwarding (kubectl port-forward).\n
- Sanitization: The skill relies on the standard output of Kubernetes administrative tools without additional escaping logic.
Audit Metadata