The Agent Skills Directory

Command Execution: The skill utilizes standard tools like kubectl and gcloud to interact with Kubernetes clusters and retrieve diagnostic information. These operations are conducted within a defined "read-only diagnostics boundary," ensuring that the agent gathers evidence without performing unauthorized mutations to the cluster environment.
External References: The skill generates deep links to the Google Cloud Console (console.cloud.google.com) to facilitate log analysis. These links point to official vendor infrastructure and are used to provide the user with direct access to evidence supporting the diagnosis.
Indirect Prompt Injection Surface: The skill analyzes external data such as Kubernetes events and kubelet logs, which is a necessary component of its diagnostic function.
Ingestion points: External data enters the agent context via kubectl get/describe commands and log queries described in SKILL.md.
Boundary markers: The skill explicitly instructs the agent to operate non-interactively and enforces a requirement that a human user must apply any proposed fixes.
Capability inventory: The skill utilizes kubectl and gcloud for information gathering and credential management.
Sanitization: The skill relies on human oversight and manual verification of the diagnostic findings and proposed commands rather than automated filtering of log content.

gke-node-notready