The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external sources, which constitutes a vulnerability surface for potential instruction injection.
Ingestion points: Data is ingested from kubectl command outputs and the api.cast.ai API responses in SKILL.md.
Boundary markers: The skill does not employ specific delimiters or boundary markers to isolate data from instructions.
Capability inventory: The skill has access to the cluster and external networks through kubectl, helm, and curl.
Sanitization: Data is processed structurally using jq, but no content sanitization or verification is performed.
[COMMAND_EXECUTION]: Utilizes shell commands for managing and validating Kubernetes cluster state.
Evidence: Inclusion of kubectl, helm, and curl commands for agent validation and cluster rollback in SKILL.md.
[EXTERNAL_DOWNLOADS]: Fetches data from the official CAST AI API for configuration and reporting.
Evidence: curl commands targeting https://api.cast.ai/v1/kubernetes/clusters/ to retrieve cluster policies and savings data in SKILL.md.

castai-prod-checklist