kubernetes-ops

Pass

Audited by Gen Agent Trust Hub on Sep 20, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to use powerful Kubernetes utilities including kubectl exec, kubectl debug, and kubectl port-forward. These commands allow for interactive shell access and network tunneling into the cluster environment, which are standard for the skill's purpose but represent a significant capability tier.
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection because it reads external, potentially untrusted data from the cluster and has the ability to execute further commands based on that data.
  • Ingestion points: Container logs (kubectl logs), resource event streams (kubectl get events), and detailed resource configurations (kubectl describe) referenced in SKILL.md and references/details.md.
  • Boundary markers: Absent. There are no instructions to the agent to treat data from the cluster as untrusted or to ignore instructions contained within logs.
  • Capability inventory: Full resource lifecycle management (kubectl apply/edit/delete), scaling, and interactive container execution (kubectl exec).
  • Sanitization: None. The skill does not provide mechanisms to filter or sanitize output from the cluster before the agent processes it.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 20, 2026, 03:59 PM
Security Audit — agent-trust-hub — kubernetes-ops