k8s-security

Pass

Audited by Gen Agent Trust Hub on Nov 9, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill implements a read-only auditing and policy enforcement logic for Kubernetes security. It uses standard kubectl toolsets to retrieve cluster information.
  • [COMMAND_EXECUTION]: The skill uses various kubectl_* and cilium_* tools to interact with the Kubernetes API. These are documented for the intended purpose of security auditing and do not show signs of arbitrary command injection.
  • [REMOTE_CODE_EXECUTION]: The SKILL.md file references a project-maintained MCP server (Kyverno) and official documentation. These are reputable sources and do not constitute unsafe remote code execution.
  • [DATA_EXPOSURE]: The skill defines patterns for auditing secrets (kubectl_get(resource_type="secrets")). While it accesses sensitive data, the purpose is clearly defined as security auditing, and there are no network exfiltration patterns directed at external or untrusted domains.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes Kubernetes resource configurations (YAML). It contains standard warnings about dangerous RBAC patterns but lacks explicit sanitization logic for ingested data. However, the risk is inherent to the auditing task and follows the capability tier of the toolset used (LOW severity risk).
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 9, 2026, 06:52 PM