kubernetes-pentesting

Warn

Audited by Gen Agent Trust Hub on Jul 21, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONDATA_EXFILTRATIONREMOTE_CODE_EXECUTIONCREDENTIALS_UNSAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides instructions for escaping containers to the host node using nsenter and chroot commands within privileged pods, granting full control over the underlying system.
  • [DATA_EXFILTRATION]: Contains methods for harvesting sensitive Kubernetes service account tokens, etcd secrets, and cloud instance metadata (including AWS IAM credentials, GCP OAuth tokens, and Azure identities).
  • [REMOTE_CODE_EXECUTION]: Describes techniques for deploying privileged pods with hostPID, hostNetwork, and host filesystem mounts to execute code with node-level privileges.
  • [CREDENTIALS_UNSAFE]: References paths to high-privilege credential files such as /etc/kubernetes/admin.conf, /etc/kubernetes/kubelet.conf, and ~/.kube/config.
  • [COMMAND_EXECUTION]: Employs Python one-liners to dynamically process, filter, and decode sensitive Kubernetes secrets and container registry credentials from API outputs.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jul 21, 2026, 01:24 PM
Security Audit — agent-trust-hub — kubernetes-pentesting