The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Fetches Kubernetes deployment manifests from the official Ingress NGINX and cert-manager GitHub repositories. These resources originate from well-known and established projects in the container orchestration ecosystem.
[COMMAND_EXECUTION]: Provides instructions for using various command-line utilities such as kubectl for resource management, apt for package installation, and wg for VPN configuration. These commands are standard for the technical tasks described in the skill.
[PRIVILEGE_ESCALATION]: Includes administrative steps requiring elevated privileges, such as installing system packages, modifying configuration files in /etc/, and enabling systemd services. These actions are necessary for the stated purpose of configuring cluster-wide networking and VPN access.
[INDIRECT_PROMPT_INJECTION]: The skill provides templates for Kubernetes resources that an agent might populate using user-provided data. This represents a potential surface for indirect injection if untrusted inputs are used without validation.
Ingestion points: User-supplied domain names, labels, and metadata used to populate Ingress and NetworkPolicy templates.
Boundary markers: No explicit delimiters or validation warnings are included in the configuration templates.
Capability inventory: Includes the ability to apply configurations via kubectl, install software via apt, and modify system networking parameters via sysctl.
Sanitization: The skill provides raw templates and does not include logic for sanitizing or escaping the interpolated values.

k8s-networking