The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes ansible.builtin.command and ansible.builtin.shell to interact with Proxmox VE CLI utilities (pvecm, pveceph, pveum, qm, pvesm). These executions are limited to operations that lack native Ansible modules, such as cluster creation, CEPH initialization, and OSD management. The skill provides clear instructions on making these commands idempotent to prevent unintended state changes.
[EXTERNAL_DOWNLOADS]: The documentation includes instructions for installing the community.proxmox collection using ansible-galaxy. This is the standard delivery mechanism for Ansible extensions and targets the official Ansible registry.
[INDIRECT_PROMPT_INJECTION]: As an automation guide, the skill contains multiple ingestion points where user-provided variables (e.g., cluster_name, vm_name, ceph_network) are interpolated into shell commands.
Ingestion points: Variables defined in group_vars or passed at runtime (SKILL.md, ceph-automation.md, cluster-automation.md).
Boundary markers: The skill relies on standard Ansible variable interpolation syntax; no specific boundary markers for sanitizing variable content are implemented within the provided examples.
Capability inventory: Extensive capabilities including cluster management, user/ACL creation, disk wiping (wipefs), and network configuration.
Sanitization: The skill assumes a trusted administrative context but mitigates risk by prioritizing native modules and encouraging the use of no_log: true for sensitive tasks.
[CREDENTIALS_UNSAFE]: The skill demonstrates safe credential handling by recommending external secret lookups (e.g., using Infisical) and using placeholders instead of hardcoded secrets. It specifically instructs users to use the no_log attribute for tasks involving passwords or tokens.

ansible-proxmox