The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The launcher script green dynamically fetches the agent-network-k8s and green libraries from the author's GitHub repositories (github.com/getcolors/*) at runtime using Babashka's dependency management system. The external logic is pinned to a specific Git SHA to ensure code integrity.
[COMMAND_EXECUTION]: The skill performs infrastructure management operations, including executing kubectl commands for cluster interaction and calling lifecycle management functions that interact with cloud provider APIs (Vultr) to provision and destroy Kubernetes resources.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it processes external configuration data to drive deployment actions.
Ingestion points: The script reads and processes the colors.yml file, which is resolved by walking up the directory tree from the execution path.
Boundary markers: No explicit boundary markers or instructions to ignore embedded commands are present in the launcher script's ingestion logic.
Capability inventory: The skill has significant capabilities, including shell command execution (kubectl), cloud resource provisioning, and the ability to modify Kubernetes secrets and network policies.
Sanitization: The launcher script lacks explicit sanitization or schema validation for the configuration values before they are passed to the deployment logic.

package-agent-network-k8s-green