k8s-cluster-review

Pass

Audited by Gen Agent Trust Hub on Nov 12, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill leverages the kubectl CLI to perform a wide range of discovery tasks, including inspecting pod security contexts, RBAC bindings, and resource requests. These operations are strictly focused on auditing and do not perform modifications.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data retrieved from the Kubernetes API (live object manifests). While it uses this data to populate remediation templates ('Proposed fixes'), it mitigates the risk of command injection or malicious output by explicitly forbidding the agent from executing those commands directly, thereby ensuring a human-in-the-loop for any cluster changes.
  • [SAFE]: The instructions contain clear, prohibitive guidelines regarding sensitive data access, specifically stating 'Never read Secrets' and 'never exec into a pod'. These constraints effectively limit the potential for privilege abuse or data exfiltration during the audit process.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 12, 2026, 11:17 PM