The Agent Skills Directory

[PERSISTENCE]: The skill recommends modifying shell initialization files to store command aliases.
Evidence: The '快捷别名' (Shortcut Aliases) section suggests adding aliases to ~/.bashrc or ~/.zshrc.
[INDIRECT_PROMPT_INJECTION]: The skill processes external, potentially untrusted data which could contain instructions intended to influence the agent's behavior.
Ingestion points: Kubernetes resource manifests (manifest.yaml), container logs (kubectl logs), and cluster event descriptions (kubectl describe).
Boundary markers: None identified in the prompt templates to separate ingested data from instructions.
Capability inventory: The skill includes high-privilege operations such as container shell access (kubectl exec), resource deletion (kubectl delete), and network tunneling (kubectl port-forward).
Sanitization: No sanitization or validation of the ingested logs or manifests is performed before processing.
[COMMAND_EXECUTION]: The skill facilitates the execution of arbitrary commands within container environments and cluster configuration changes.
Evidence: Commands include kubectl exec -it <pod-name> -- /bin/bash for interactive shell access and kubectl patch/kubectl edit for modifying cluster resources.

kubectl-basics