skills/timsonner/agent-skills/kubectl/Gen Agent Trust Hub

kubectl

Fail

Audited by Gen Agent Trust Hub on Sep 3, 2026

Risk Level: HIGHCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [COMMAND_EXECUTION]: The skill includes instructions to use kubectl exec to spawn interactive shells within running pods (kubectl exec -it <pod-name> -n default -- /bin/sh). This allows for arbitrary command execution inside the container environment, which can bypass host-level monitoring and controls.
  • [DATA_EXFILTRATION]: The skill guides the agent to identify and retrieve sensitive data stored within the Kubernetes cluster. Specifically, it provides syntax for listing secrets (kubectl get secrets) and extracting their encoded content using JSONPath queries (-o jsonpath='{.data}'). It also instructs the agent to inspect pod manifests for 'sensitive env vars', posing a risk of credential exposure.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Sep 3, 2026, 02:39 AM
Security Audit — agent-trust-hub — kubectl