Proxmox AIops

> Disclaimer: This is a community-maintained open-source project and is not affiliated with, endorsed by, or sponsored by Proxmox Server Solutions GmbH. "Proxmox" is a trademark of its owner. Source code is publicly auditable at github.com/AIops-tools/Proxmox-AIops under the MIT license.

Governed VM and container lifecycle operations for Proxmox VE — 43 MCP tools, every one wrapped with the bundled @governed_tool harness: a local unified audit log under ~/.proxmox-aiops/, token/runaway budget guard, undo-token recording, and descriptive risk-tier labels.

> Standalone: the governance harness is bundled in the package (proxmox_aiops.governance) — proxmox-aiops has no external skill-family dependency. Coverage focuses on common Proxmox operations and is not yet exhaustive.

What This Skill Does

Quick Install

uv tool install proxmox-aiops
proxmox-aiops doctor

Or as an OpenClaw plugin, which installs this skill and its MCP server together:

openclaw plugins install clawhub:@zw008/proxmox-aiops
openclaw skills info proxmox-aiops          # expect: Visible to model: yes

Needs uvx on PATH: the MCP server is fetched with uv, pinned to this release.

When to Use This Skill

• List/inspect Proxmox QEMU VMs and their config
• Power ops: start, hard-stop, graceful shutdown, reboot
• Reconfigure (cores/memory), clone, delete, or migrate a VM between nodes
• Grow a VM disk (grow-only — shrink is refused) or move it to another storage
• Create, list, and restore vzdump backups
• Create / delete / list / roll back VM snapshots
• Manage LXC containers (list, start, stop)
• Inspect cluster nodes, quorum, the /cluster/resources inventory, node load/mem, and poll async tasks + fetch their logs by UPID; get a free VMID
• Check HA status / HA-managed resources (handles "HA not configured" gracefully)
• List resource pools and their members
• Inspect VM firewall rules and the cluster firewall enable state (read-only)
• Ping a VM's QEMU guest agent
• List storage pools and their content (ISOs, disk images, backups)
• Diagnose cluster health: rank nodes by CPU/memory/root-fs pressure, and scan VMs/containers for stopped guests, memory saturation, and disks near full — each finding cites the measured number and a concrete action (read-only RCA)

Do NOT use when the target is not Proxmox VE (other hypervisors, Kubernetes, or cloud providers are out of scope for this skill).

Common Workflows

Triage a "cluster feels slow" complaint (read-only)

1. proxmox-aiops diagnose node-pressure → worst-first table of nodes over the CPU/mem/disk thresholds, each row citing the measured % and the fix
2. proxmox-aiops diagnose guest-health → guests near their memory ceiling or with disks near full, plus the list of stopped guests
3. Act on the top finding — e.g. proxmox-aiops vm migrate <vmid> --to-node <n> to shed load off a hot node, or vm reconfigure <vmid> --memory ... for a RAM-starved guest. All of these route through the governed path (audited, undo recorded).

Snapshot, then reconfigure a VM

1. proxmox-aiops vm list → find the vmid and confirm it is the right VM/node
2. proxmox-aiops vm snapshot-create <vmid> --name pre-change → baseline before any risky change
3. proxmox-aiops vm reconfigure <vmid> --cores 8 --memory 16384 → the harness captures the prior cores/memory as the undo descriptor
4. Failure branch: if the change goes wrong, the write recorded an _undo_id — reverse it with proxmox-aiops undo apply <id>, or roll back with proxmox-aiops vm snapshot-rollback <vmid> --name pre-change.

Free a node that is out of memory

1. proxmox-aiops diagnose node-pressure → identify the node flagged high memory
2. proxmox-aiops cluster resources --type vm → find a movable guest on that node
3. proxmox-aiops vm migrate <vmid> --to-node <other> --dry-run → preview, then run without --dry-run (migrate is high risk; the inverse migrate-back is recorded)
4. Re-run diagnose node-pressure to confirm the pressure cleared.

Stop a VM safely

1. proxmox-aiops vm get <vmid> → confirm current status is running
2. proxmox-aiops vm stop <vmid> --dry-run → preview the exact API call
3. proxmox-aiops vm stop <vmid> → double confirmation required; vm_stop records an inverse vm_start undo descriptor
4. Failure branch: if doctor shows the node unreachable or the secret env var is missing, fix credentials with proxmox-aiops secret set <target> before retrying — the stop is never issued against an unauthenticated session.

Usage Mode

MCP Tools (43 — 25 read, 18 write)

Harness features that light up: write tools with a clean inverse (vm_start/vm_stop/vm_shutdown/vm_reconfigure/vm_clone/vm_migrate/vm_snapshot_create/vm_move_disk/ct_start/ct_stop) pass an undo= lambda so the harness records an inverse descriptor (with _undo_id) to the undo store — vm_reconfigure captures the prior cores/memory, vm_clone's inverse is vm_delete(newid), vm_migrate's is migrate-back, vm_move_disk's is move-back to the captured source storage. backup_restore records a vm_delete inverse only when it restored into a free VMID (a forced overwrite is destructive and declares none). Irreversible writes (vm_delete, vm_snapshot_rollback, backup_restore with force) declare no undo and are tagged risk_level=high; vm_resize_disk is grow-only and refuses shrink before any API call. All 43 tools are audit-logged under ~/.proxmox-aiops/ and pass through the budget/runaway guard + risk-tier tagging. Proxmox writes are async (return a task UPID) — poll with task_status (and read lines with task_log) instead of re-issuing (the runaway breaker backs this up).

CLI Quick Reference

proxmox-aiops vm list [--target <t>] [--node <n>]
proxmox-aiops vm get <vmid> [--node <n>]
proxmox-aiops vm start <vmid> [--node <n>]
proxmox-aiops vm stop <vmid> [--dry-run]              # double confirm
proxmox-aiops vm resize-disk <vmid> --disk scsi0 --size +10G   # grow-only
proxmox-aiops vm move-disk <vmid> --disk scsi0 --storage ceph [--delete]
proxmox-aiops vm agent-ping <vmid>
proxmox-aiops vm snapshot-create <vmid> --name <snap>
proxmox-aiops vm snapshot-delete <vmid> --name <snap> [--dry-run]   # double confirm
proxmox-aiops vm snapshot-list <vmid>
proxmox-aiops backup create <vmid> --storage <s> [--mode snapshot]
proxmox-aiops backup list <storage> [--vmid <id>]
proxmox-aiops backup restore <vmid> --archive <volid> --storage <s> [--force] [--dry-run]  # double confirm
proxmox-aiops cluster resources [--type vm|node|storage]
proxmox-aiops cluster node-status <node>
proxmox-aiops cluster task-log <upid>
proxmox-aiops cluster next-vmid
proxmox-aiops ha status
proxmox-aiops pool list
proxmox-aiops firewall vm-rules <vmid>
proxmox-aiops storage list [--node <n>]
proxmox-aiops diagnose node-pressure                  # rank nodes by CPU/mem/disk pressure (read-only RCA)
proxmox-aiops diagnose guest-health                   # stopped guests, mem saturation, disks near full
proxmox-aiops undo apply <id>                         # reverse a recorded governed write
proxmox-aiops init                                    # onboarding wizard (encrypted creds)
proxmox-aiops secret set <target>                     # manage encrypted secret store
proxmox-aiops doctor
proxmox-aiops mcp                                      # start MCP server (stdio)

> Credentials are managed by the proxmox-aiops init onboarding wizard and the
> proxmox-aiops secret commands, which back an encrypted secret store (no
> plaintext passwords in config.yaml).

Troubleshooting

"Config file not found"

Create ~/.proxmox-aiops/config.yaml with a targets: list (see README), and put secrets in ~/.proxmox-aiops/.env (chmod 600).

"Secret not found. Set environment variable: PROXMOX_<NAME>_SECRET"

Each target needs a per-target secret env var. For target pve-lab, set PROXMOX_PVE_LAB_SECRET=<token-uuid> in .env.

"Token auth requires user in the form 'user@realm!tokenid'"

For API-token auth (recommended, least privilege), user must include the token id after !, e.g. root@pam!claude. For password auth set auth_kind: password and use user@realm.

"No node specified and no default node configured"

Either pass --node <name> / node=<name>, or set node: on the target in config.yaml. VM operations can auto-locate a vmid across nodes, but storage listing needs an explicit node.

Audit & Safety

The skill delivers reads and writes and records them; it does not decide whether a write is
permitted. That is your agent's judgement, or the permission of the account you connect it with —
use a Proxmox VE user or API token granted only read privileges (no VM./Datastore. write roles),
and writes then fail at the server. There is no read-only switch, policy file, or approval gate.

• Audit is the guarantee, and it is not bypassable. Every operation — MCP and CLI alike — is logged to ~/.proxmox-aiops/audit.db (relocatable via PROXMOX_AIOPS_HOME): params, result, status, duration, and the risk tier. The CLI writes the same row the MCP path does.
• PROXMOX_AUDIT_APPROVED_BY / PROXMOX_AUDIT_RATIONALE are optional annotations recorded on the audit row (who/why); they are never required and never block.
• Runaway guard — a safety backstop, not authorization: the same call looped in a tight window trips a circuit breaker. Disable with PROXMOX_RUNAWAY_MAX=0; optional hard ceilings via PROXMOX_MAX_TOOL_CALLS / PROXMOX_MAX_TOOL_SECONDS.
• Undo store records inverse descriptors for reversible writes (start/stop/shutdown/reconfigure/clone/migrate/snapshot-create, container start/stop).
• Writes support --dry-run / dry_run=True and double confirmation at the CLI.

The harness is bundled in the package — no external dependency, no manual setup.

Contributing & feature requests

Coverage is intentionally focused. Missing a device, action, or feature you need? Open an issue or pull request at github.com/AIops-tools/Proxmox-AIops — feature requests, contributions, and comments are all welcome.

License

MIT — github.com/AIops-tools/Proxmox-AIops