The Agent Skills Directory

[SAFE]: The skill provides a secure framework for managing Kubernetes jobs. Operational scripts include extensive validation for input data, such as path normalization and character checking to prevent injection and traversal attacks.
[EXTERNAL_DOWNLOADS]: The skill references downloads for kubectl and minikube from official Google and Kubernetes repositories (dl.k8s.io, storage.googleapis.com). These are well-known, trusted services, and the references are documented neutrally.
[COMMAND_EXECUTION]: Shell commands for job lifecycle management (submit, status, logs, cancel) use standard kubectl calls. The documentation explicitly warns against passing secrets via command-line arguments, recommending secure alternatives like stdin or environment variables.
[INDIRECT_PROMPT_INJECTION]: While the skill processes external 'Action Request' data, the render_action_job.py script employs robust sanitization, including schema version checks, path validation, and JSON-encoding of values before manifest interpolation, effectively mitigating injection risks.

tao-run-on-kubernetes