The Agent Skills Directory

[COMMAND_EXECUTION]: Executes a local shell script ./scripts/extend_vm_storage.sh with user-provided arguments for node identification and size increments.
[COMMAND_EXECUTION]: Invokes kubectl to perform cluster-level administrative actions (drain and uncordon) and grep to inspect local configuration files like main.tf.
[REMOTE_CODE_EXECUTION]: Executes administrative commands on the Proxmox virtualization host via SSH (e.g., qm shutdown, qm resize, qm start) and performs guest-level filesystem operations using SSH to the wizard account.
[PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection via user-supplied parameters. 1. Ingestion points: User-provided <node-name> and <size-increment> parameters are interpolated into shell commands in SKILL.md. 2. Boundary markers: Absent; there are no instructions to validate or escape user input before execution. 3. Capability inventory: The skill has extensive capabilities including local command execution, remote SSH execution on infrastructure hosts, and Kubernetes cluster management. 4. Sanitization: None specified in the skill instructions; security relies entirely on the internal validation logic of the unprovided ./scripts/extend_vm_storage.sh file.

extend-vm-storage