The Agent Skills Directory

[PROMPT_INJECTION]: The skill involves processing user-provided input such as Project IDs, Cluster Names, and Regions to generate infrastructure management commands. While it contains no explicit sanitization logic for these inputs, the interaction model is consistent with the skill's primary purpose of cloud orchestration and represents an inherent, low-level operational risk rather than a malicious pattern.
[EXTERNAL_DOWNLOADS]: The instructions include steps for installing Kubernetes add-ons like Kueue and the MPI Operator by applying manifests directly from their official URLs on the 'kubernetes-sigs' and 'kubeflow' GitHub organizations. These are recognized, trusted entities within the Kubernetes ecosystem, and their use in this context is standard practice.
[COMMAND_EXECUTION]: The skill correctly documents and facilitates necessary administrative operations, such as cluster lifecycle management and pod diagnostics via 'kubectl exec'. These capabilities are essential for the skill's stated function and are presented within a framework of hardening and least-privilege principles.

gke-basics