The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to execute kubeseal, kubectl, and standard Unix utilities (base64, printf, cat) to manage sensitive resource data.
[EXTERNAL_DOWNLOADS]: The documentation describes the use of kubeseal --cert <URL>, which enables fetching public key certificates from external network locations.
[PROMPT_INJECTION]: The skill processes external data retrieved from a Kubernetes cluster via the kubernetes_resources_get tool, creating a potential surface for indirect prompt injection if the cluster data is untrusted.
Ingestion points: Secret values are read from the cluster environment via kubernetes_resources_get for updates and re-sealing operations.
Boundary markers: The skill recommends using quoted heredocs (<<'EOF') in shell commands to prevent accidental interpolation of secret values.
Capability inventory: The agent is granted capabilities for shell execution (Bash), file operations (Read, Write), and Kubernetes resource management.
Sanitization: The skill mandates a defensive sanitization step where temporary files containing plaintext secrets must be overwritten (e.g., echo "" > /tmp/secret-temp.yaml) immediately after use.

kubeseal