proxmox-vms

name	proxmox-vms
description	Manage Proxmox QEMU/KVM virtual machines via the REST API — list, inspect, start/stop/reboot, create, clone, migrate, snapshot/rollback, resize disks, and edit VM config. Load this for any virtual-machine task on Proxmox. For GPU/PCIe passthrough or other host-file-level VM config, use proxmox-host-config instead.

Proxmox VMs (QEMU/KVM)

Manage VMs through the REST API (see proxmox-api for auth + the pve-api.sh helper). Obey the homelab-core:homelab-safety skill: reads are free; any start/stop/create/delete/config-change needs a summarized plan + approval first.

Inspect (read-only)

scripts/pve-api.sh GET /nodes/{node}/qemu                              # list VMs on a node
scripts/pve-api.sh GET /nodes/{node}/qemu/{vmid}/status/current        # power state, uptime, HA
scripts/pve-api.sh GET /nodes/{node}/qemu/{vmid}/config                # full config (disks, net, cpu, mem)
scripts/pve-api.sh GET /cluster/resources | jq '.data[]|select(.type=="qemu")'

Lifecycle (write — confirm first)

scripts/pve-api.sh POST /nodes/{node}/qemu/{vmid}/status/start
scripts/pve-api.sh POST /nodes/{node}/qemu/{vmid}/status/shutdown     # graceful (ACPI)
scripts/pve-api.sh POST /nodes/{node}/qemu/{vmid}/status/stop         # hard stop (data risk if mid-write)
scripts/pve-api.sh POST /nodes/{node}/qemu/{vmid}/status/reboot

Prefer shutdown (graceful) over stop (hard power-off). These return a UPID — poll the task for success (see proxmox-api).

Snapshots (do this BEFORE risky changes)