managing-k8s-prometheus-operator

Pass

Audited by Gen Agent Trust Hub on Apr 12, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses kubectl to discover monitoring resources and fetch logs from the Prometheus operator.
  • [COMMAND_EXECUTION]: It uses kubectl exec to access internal diagnostic endpoints for Prometheus and Alertmanager on localhost.
  • [PROMPT_INJECTION]: Indirect Prompt Injection Surface: 1. Ingestion points: Kubernetes resource names, log outputs, and JSON responses from monitoring APIs (SKILL.md). 2. Boundary markers: Absent. No delimiters are used to isolate cluster data from instructions. 3. Capability inventory: Use of kubectl exec and kubectl logs across the cluster. 4. Sanitization: Absent. The skill does not filter or escape content retrieved from the Kubernetes environment.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 12, 2026, 04:21 PM
Security Audit — agent-trust-hub — managing-k8s-prometheus-operator