gke-workload-troubleshooting

Pass

Audited by Gen Agent Trust Hub on Sep 18, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes data from untrusted sources that could contain malicious instructions.
  • Ingestion points: The skill ingests untrusted data via kubectl logs (Step 3), kubectl get events, and gcloud logging read (Step 2).
  • Boundary markers: There are no explicit delimiters or instructions to the agent to ignore embedded commands within the logs or events being processed.
  • Capability inventory: The skill possesses significant capabilities including shell execution (kubectl, gcloud, git) and the ability to create/update GitHub Pull Requests (Step 5).
  • Sanitization: The instructions do not specify any sanitization, filtering, or validation of the content retrieved from logs or events before the agent processes it.
  • [COMMAND_EXECUTION]: The skill dynamically constructs shell commands using variables extracted from user prompts or environment defaults.
  • Evidence: Commands such as kubectl get deployment {workload_name}, gcloud logging read ... --project="{project_id}", and git log -p -S "{image_name}" -- {manifest_file_path} interpolate parameters directly into shell strings.
  • Risk: If the platform does not sanitize these parameters, an attacker could provide a malicious string (e.g., for {workload_name}) to execute arbitrary commands on the host system.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 18, 2026, 05:40 AM
Security Audit — agent-trust-hub — gke-workload-troubleshooting