k8s-cost

Pass

Audited by Gen Agent Trust Hub on Nov 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes Kubernetes resource metadata which could potentially harbor malicious instructions.
  • Ingestion points: Pod and namespace metadata retrieved through kubectl_get, get_idle_resources, and get_resource_recommendations tools.
  • Boundary markers: The prompt instructions do not include specific delimiters to isolate external data from the agent's internal reasoning.
  • Capability inventory: The skill utilizes tools that can modify cluster state, including kubectl_scale and resource deletion commands for optimization.
  • Sanitization: There is no evidence of input validation or sanitization of the resource names or metrics fetched from the cluster.
  • [COMMAND_EXECUTION]: The skill executes management commands via the Kubernetes API. While intended for cost optimization, these include high-impact operations like scaling deployments and deleting persistent volume claims.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 9, 2026, 06:52 PM