gke-platform-troubleshoot
Pass
Audited by Gen Agent Trust Hub on Jun 22, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides numerous playbooks that execute administrative commands using
kubectlandgcloud. These commands are used for diagnosing application states, patching configurations, and managing GCP disk resources. This behavior is consistent with the skill's primary purpose as a troubleshooting tool for platform operators. - [COMMAND_EXECUTION]: Embedded Python one-liners are used to process and summarize output from diagnostic commands, such as calculating total disk usage or parsing JSON status fields from Kubernetes resources. This local processing of data is a standard administrative practice.
- [PROMPT_INJECTION]: The skill processes potentially untrusted data from pod logs and Kubernetes events. While this creates an indirect prompt injection surface where log content could attempt to influence the agent, the skill is designed for human-directed troubleshooting, and the instructions prioritize manual diagnostic steps over autonomous action.
Audit Metadata