xcpng-aiops
It provides governed MCP access to XCP-ng fleet operations through Xen Orchestra’s REST API, with audited reads, writes, dry-run previews, risk tiers, and undo recording where reversible.
Inspect fleet health and inventory: overview, VMs, hosts, pools, SRs/VDIs, snapshots, backup jobs/logs, and tasks.
Run structured RCA analyses for VM health, SR usage, backup failures, and pool patch/HA posture.
Perform governed writes: start/stop/reboot/migrate VMs, create/delete/revert snapshots, and rescan SRs.
Preview writes with
dry_run; reversible writes return an_undo_id.List and apply recorded undo tokens, with a warning when the original effect was not verified.
Benefit from audit logging, budget/runaway guards, descriptive risk tiers, and a partial self-lockout guard for stopping the Xen Orchestra VM.
XCP-ng AIops
Audit availability: Real writes require a committed
pendingaudit row and stop before contacting the platform if storage is unavailable. Reads and previews may continue with a warning; new audit rows store outcome summaries rather than full responses.
Disclaimer: Community-maintained open-source project. Not affiliated with, endorsed by, or sponsored by Vates, the XCP-ng project, or the Xen Orchestra project. "XCP-ng", "Xen Orchestra", and "Xen" are trademarks of their owners. MIT licensed.
AI-powered XCP-ng operations via Xen Orchestra's REST API with a
built-in governance harness — unified audit log, policy engine,
token/runaway budget guard, undo-token recording, and descriptive risk
tiers. Built for homelabs and small/self-hosted XCP-ng fleets that want an AI
agent to triage VM health, storage pressure, backup failures, and patch
posture — with every write audited, previewable, and (where honest)
reversible. Self-contained: no dependencies beyond httpx and the MCP SDK.
Requires a Xen Orchestra instance (XO from sources or the Xen Orchestra Appliance, 5.x with
/rest/v0). XO is the management plane this tool talks to — direct per-host XAPI access is out of scope for v0.1. Do NOT use for Proxmox VE — use proxmox-aiops.
What works
CLI (
xcpng-aiops ...):init,overview,vm list/get/stats/health-rca/start/stop/reboot/migrate,host list/get/missing-patches,pool list/get/posture,sr list/get/vdis/usage-rca/rescan,snapshot list/create/delete/revert,backup jobs/logs/failure-rca,task list,secret set/list/rm/migrate/rotate-password,doctor,mcp.MCP server (
xcpng-aiops mcporxcpng-aiops-mcp): 29 tools (19 read, 8 write, 2 undo), every one wrapped with the bundled@governed_toolharness.Four flagship RCA analyses (cause + action structured output): VM health, SR usage, backup-job failures, pool patch & HA posture.
Encrypted credentials: the XO authentication token lives in an encrypted store
~/.xcpng-aiops/secrets.enc(Fernet + scrypt) — never plaintext on disk. Unlock with a master password fromXCPNG_AIOPS_MASTER_PASSWORD(MCP/CI) or an interactive prompt (CLI).Reversibility:
vm_start↔vm_stoprecord each other as inverses;vm_migratecaptures the REAL source host before moving and records "migrate back";snapshot_createcaptures the created snapshot's REAL id from the XO response and records "delete THAT snapshot". Irreversible ops (snapshot_delete,snapshot_revert,vm_reboot) capture prior state for the audit record and honestly declare no undo.Safety: destructive CLI ops require double confirmation and support
--dry-run; every write MCP tool takes adry_runpreview (no write call, no undo recorded).Self-lockout guard (partial — read this): Xen Orchestra is commonly a VM on a pool it manages, and stopping that VM kills the API this tool talks to —
vm_startcan then no longer be sent, so recovery needs hypervisor console access (xe vm-start). Setxo_self_vm_uuidon the target (xcpng-aiops initasks) andvm_stoprefuses exactly that uuid — on--dry-runas well, since a preview that green-lights a call the tool will then refuse is reporting the wrong outcome. If you do not set it there is no protection at all: XO's REST API exposes no self endpoint and its token carries no claims, so the tool cannot discover which VM it runs on, and it fails open rather than guess. Thedry_runpreview adds a weakerselfVmHintwhen a VM's reported IP matches the configured XO host — that is a coincidence worth checking, not a finding, and it never blocks (it sees nothing without the guest agent and fires on every VM behind a shared proxy).
Related MCP server: firewall-aiops
Capability matrix (29 MCP tools)
Domain | Tools | Count | R/W |
Overview |
| 1 | read |
VMs |
| 4 | read |
| 4 | write (medium) | |
Hosts |
| 2 | read |
Pools |
| 3 | read |
SRs / VDIs |
| 4 | read |
| 1 | write (medium) | |
Snapshots |
| 1 | read |
| 3 | write | |
Backups |
| 3 | read |
Tasks |
| 1 | read |
Undo |
| 2 | read + replay |
Flagship RCAs
vm_health_rca— VMs halted unexpectedly (auto-poweron / HA restart priority set), paused/suspended VMs, running VMs without guest tools, CPU/memory pressure from RRD stats → cause + action per finding.sr_usage_rca— SRs ranked by physical fullness (near-full ≥ 85%, critical ≥ 95%), thin-provision overcommit (virtual allocation > capacity), orphaned VDIs (attached to no VM) with reclaimable bytes per SR.backup_failure_rca— failed/skipped/interrupted XO backup runs classified: vdi-chain (coalesce not finished), quiesce (guest VSS), transport (remote unreachable), storage-full, unknown — with per-job counts and sample messages.pool_patch_ha_posture— hosts missing patches, hosts pending reboot, version skew across a pool's hosts (breaks live migration / rolling updates), multi-host pools without HA.
What this tool does, and does not, decide
It delivers XCP-ng operations — reads and writes — accurately and efficiently, and records every one of them. It does not decide whether a write is allowed to happen. That is the agent's judgement, or the permission of the Xen Orchestra account whose token you connect it with: give that XO user a read-only ACL, or scope its token down, and the writes fail at Xen Orchestra — the place that actually owns the permission.
So there is no read-only switch, no policy file, no approval gate to configure.
The one thing the tool guarantees is that nothing is silent: every call, over
MCP and over the CLI alike, lands an audit row in ~/.xcpng-aiops/audit.db,
and reversible writes still capture their before-state and record an inverse.
Each tool declares a
risk_level, kept in agreement with its[READ]/[WRITE]documentation tag by a test, and carried into the audit row as a descriptive tier — so a reviewer can see at a glance that a row was a high-risk snapshot delete. It is a label, not a gate.
Running a smaller / local model? See agent-guardrails.md — it lists the guardrails this tool now enforces for you (so you don't spend prompt budget restating them) and gives a ready-made system prompt for what's left.
Quick start
As a Claude Code plugin
One install gives an agent both the skill and the MCP server:
/plugin marketplace add AIops-tools/marketplace
/plugin install xcpng-aiops@aiops-toolsThe MCP server is fetched with uv and pinned to the
package version this plugin declares, so an audit row can be traced back to the
code that wrote it. Credentials are still configured with xcpng-aiops init — see below.
As an OpenClaw plugin
The same bundle is published on ClawHub, where one install delivers the skill and its MCP server together:
openclaw plugins install clawhub:@zw008/xcpng-aiops
openclaw skills info xcpng-aiops # expect: Visible to model: yesRestart the OpenClaw gateway afterwards so it loads the plugin. The MCP server is
fetched with uv, pinned to this exact release, so
uvx has to be on PATH — without it the skill still installs but reports
Visible to model: no. Credentials are configured exactly as below.
As a CLI or standalone MCP server
uv tool install xcpng-aiops
xcpng-aiops init # interactive wizard: XO URL + encrypted token
xcpng-aiops doctor # verify config, encrypted store, XO reachability + pool count
xcpng-aiops overview # one-shot fleet health summaryinit writes ~/.xcpng-aiops/config.yaml (non-secret connection details) and
stores the XO token encrypted in ~/.xcpng-aiops/secrets.enc. Example
config it produces:
targets:
- name: xo1
url: https://xo.example.com # the XO web origin (management plane)
verify_ssl: true # set false only for self-signed lab certs
api_path: /rest/v0Create the token in the XO UI (user menu → Personal tokens) or with
xo-cli --createToken. For non-interactive use (MCP server, CI, cron) export
the master password so the store can be unlocked without a prompt:
export XCPNG_AIOPS_MASTER_PASSWORD='your-master-password'Where that password then lives: an exported variable is readable by every process this shell starts and is recorded by shell history. On a shared or long-lived host, prefer the interactive prompt, or inject it from a secret manager for the life of the one command that needs it.
MCP client config
{
"mcpServers": {
"xcpng-aiops": {
"command": "uvx",
"args": ["--from", "xcpng-aiops", "xcpng-aiops-mcp"],
"env": { "XCPNG_AIOPS_MASTER_PASSWORD": "your-master-password" }
}
}
}Env-block caveat: MCP clients launch the server with a minimal environment — your shell profile's exports are not inherited. Put
XCPNG_AIOPS_MASTER_PASSWORD(and, if you use them,XCPNG_AIOPS_HOME/XCPNG_AIOPS_CONFIG/XCPNG_AUDIT_APPROVED_BY) in theenvblock above, or the encrypted store cannot be unlocked and every tool returns a teaching error.
Managing secrets
xcpng-aiops secret set xo1 # prompts hidden for the XO token
xcpng-aiops secret list # names only, values never shown
xcpng-aiops secret rm xo1
xcpng-aiops secret rotate-password # re-encrypt under a new master password
xcpng-aiops secret migrate # import a legacy plaintext .env, then retires itA legacy plaintext env var XCPNG_<TARGET_NAME_UPPER>_TOKEN is still honoured
as a fallback with a deprecation warning (migrate with xcpng-aiops secret migrate).
Governance
Every MCP tool — and every CLI write, which routes through the same governed
functions — passes through @governed_tool. It records; it does not authorize
(see above).
Audit — every call (tool, params with secrets redacted, result, status, duration, risk tier, and any operator-supplied approver/rationale) lands in
~/.xcpng-aiops/audit.db(relocate withXCPNG_AIOPS_HOME). The CLI writes the same row the MCP path does — there is no unaudited entry point.Budget / runaway guard — a safety backstop, not an authorization gate: cumulative call and wall-time caps plus a tight-loop circuit breaker (
XCPNG_MAX_TOOL_CALLS,XCPNG_MAX_TOOL_SECONDS,XCPNG_RUNAWAY_MAX) stop a stuck agent from burning unbounded calls/time.Undo recording — reversible writes record a replayable inverse descriptor to
~/.xcpng-aiops/undo.dband return an_undo_id; irreversible writes record prior state only.Risk tier — a descriptive label on the audit row derived from
risk_level; it gates nothing.Output hygiene — all XO-returned text is sanitized and bounded before it reaches the agent.
支持范围 / Supported scope
Area | Read | Write (governed) |
VMs | list / get / RRD stats / health RCA | start, stop (clean/hard), reboot (clean/hard), migrate |
Hosts | list / get / missing patches | — |
Pools | list / get / patch & HA posture RCA | — |
SRs / VDIs | list / get / VDI list (orphan filter) / usage RCA | rescan |
Snapshots | list | create, delete, revert |
Backups | jobs / logs / failure RCA | — |
Tasks | list | — |
缺功能?(Missing something?) Coverage is intentionally focused. Open an issue or PR at github.com/AIops-tools/XCPng-AIops — feature requests, contributions, and comments are all welcome.
Scope & caveats
Verification status: all behaviour is validated against mocked REST responses; there is no recorded end-to-end run against a live Xen Orchestra instance yet.
xcpng-aiops doctoris the fastest live check — seedocs/VERIFICATION.mdfor the full checklist.Endpoint paths (e.g.
/vms/<id>/actions/snapshot,/vm-snapshots/<id>,/srs/<id>/actions/rescan,/hosts/<id>/missing_patches,/backup/logs) are modelled against the documented XO REST/rest/v0API and need live verification — action names may differ across XO releases.Management plane only: everything goes through XO. Per-host XAPI, XO server management (adding servers, users), and backup job execution (run/restore) are out of scope for v0.1.
Out of scope by design: anything that destroys bulk data (VM/VDI deletion) — only
snapshot_delete/snapshot_revertdiscard state, and both arehighrisk + double-confirmed.
Not for
Other hypervisors or VM platforms (use their own ops tools — e.g. Proxmox VE → proxmox-aiops), NAS/storage appliances, backup software suites, container clusters, or network devices — those are out of scope for this tool.
License
Available Tools
29 toolsbackup_failure_rcaARead-only
[READ][RCA] Classify failed / skipped backup runs — cause + action per job.
Groups recent run failures by job and classifies them: vdi-chain (coalesce
not finished), quiesce (guest VSS), transport (remote unreachable),
storage-full, or unknown (with sample messages for triage).
Args:
limit: Max recent log entries to examine (default 50).
target: Xen Orchestra target name from config; omit for the default.
Inputs: limit, target.
backup_job_listARead-only
[READ] List VM backup jobs (id, name, mode).
Returns {"jobs": [...], "returned": N, "limit": L, "truncated": bool}.
When truncated is true there are more jobs than were returned.
Args:
limit: Max jobs to return (default 200).
target: Xen Orchestra target name from config; omit for the default.
Inputs: limit, target.
backup_log_listARead-only
[READ] Recent backup run logs: status + failed-task messages.
Returns {"logs": [...], "returned": N, "limit": L, "truncated": bool}.
Truncation is measured (one extra record is requested), so a true value
means there really are older runs — re-run with a higher limit.
Args:
limit: Max recent log entries to return (default 50).
target: Xen Orchestra target name from config; omit for the default.
Inputs: limit, target.
host_getARead-only
[READ] Detail for one host by uuid (version, build, memory, VMs).
Args:
host_id: Host uuid (see host_list).
target: Xen Orchestra target name from config; omit for the default.
Inputs: host_id (required), target.
host_listBRead-only
[READ] List hosts: version, state, memory usage, resident VM count.
Args:
pool: Optional pool uuid to filter by.
target: Xen Orchestra target name from config; omit for the default.
Inputs: pool, target.
overviewARead-only
[READ] One-shot health summary: pools, hosts, VMs, SRs, recent backups.
Start any triage here — it fans out over the other read ops (best-effort,
a failing collection reports an error block instead of blanking the rest).
Args:
target: Xen Orchestra target name from config; omit for the default.
Inputs: target.
pool_getARead-only
[READ] Detail for one pool by uuid.
Args:
pool_id: Pool uuid (see pool_list).
target: Xen Orchestra target name from config; omit for the default.
Inputs: pool_id (required), target.
pool_listARead-only
[READ] List XCP-ng pools with master, HA state, default SR.
Args:
target: Xen Orchestra target name from config; omit for the default.
Inputs: target.
pool_patch_ha_postureARead-only
[READ][RCA] Patch & HA posture per pool — cause + action per finding.
Flags hosts missing patches, hosts pending a reboot, version skew across a
pool's hosts (breaks live migration / rolling updates), and multi-host
pools running without HA.
Args:
pool_id: Optional pool uuid to analyze just one pool.
target: Xen Orchestra target name from config; omit for the default.
Inputs: target, pool_id.
snapshot_createADestructive
[WRITE][risk=medium] Snapshot a VM. Inverse: delete THAT snapshot.
The created snapshot's REAL id is captured from the XO response, so the
recorded undo (snapshot_delete) is replayable.
Args:
vm_id: VM uuid to snapshot (see vm_list).
name: Snapshot name (e.g. 'pre-change-2026-07-17').
dry_run: If True, preview without snapshotting (no undo recorded).
target: Xen Orchestra target name from config; omit for the default.
Inputs: name (required), vm_id (required), target, dry_run.
snapshot_deleteADestructive
[WRITE][risk=high] Delete a VM snapshot by uuid. IRREVERSIBLE.
Captures the snapshot's prior state (name / time / parent VM) for the audit
record; declares no undo.
Args:
snapshot_id: Snapshot uuid (see snapshot_list).
dry_run: If True, preview without deleting.
target: Xen Orchestra target name from config; omit for the default.
Inputs: snapshot_id (required), target, dry_run.
snapshot_listARead-only
[READ] List VM snapshots, optionally filtered to one VM.
Returns {"snapshots": [...], "returned": N, "limit": L, "truncated": bool}.
When truncated is true there are more snapshots than were returned.
Args:
vm_id: Optional VM uuid to filter (see vm_list).
limit: Max snapshots to return after filtering (default 200).
target: Xen Orchestra target name from config; omit for the default.
Inputs: limit, vm_id, target.
snapshot_revertADestructive
[WRITE][risk=high] Revert a VM to a snapshot. IRREVERSIBLE.
The VM's CURRENT state is replaced by the snapshot — take a fresh snapshot
first (snapshot_create) if you may need to come back. Captures the
snapshot's state for the audit record; declares no undo.
Args:
snapshot_id: Snapshot uuid to revert to (see snapshot_list).
dry_run: If True, preview without reverting.
target: Xen Orchestra target name from config; omit for the default.
Inputs: snapshot_id (required), target, dry_run.
sr_getARead-only
[READ] Detail for one SR by uuid.
Args:
sr_id: SR uuid (see sr_list).
target: Xen Orchestra target name from config; omit for the default.
Inputs: sr_id (required), target.
sr_listARead-only
[READ] List SRs with capacity, physical usage, virtual allocation.
Returns {"srs": [...], "returned": N, "limit": L, "truncated": bool}.
When truncated is true there are more SRs than were returned.
Args:
pool: Optional pool uuid to filter by.
limit: Max SRs to return after filtering (default 200).
target: Xen Orchestra target name from config; omit for the default.
Inputs: pool, limit, target.
sr_rescanADestructive
[WRITE][risk=medium] Rescan an SR (metadata refresh — no data change, no undo).
Lowest-impact write in this tool, but a write nonetheless: it refreshes SR
metadata and records no undo (there is no prior state to restore).
Args:
sr_id: SR uuid (see sr_list).
dry_run: If True, preview without rescanning.
target: Xen Orchestra target name from config; omit for the default.
Inputs: sr_id (required), target, dry_run.
sr_usage_rcaARead-only
[READ][RCA] SR usage root-cause analysis — cause + action per finding.
Ranks SRs by physical fullness (near-full / critical), flags thin-provision
overcommit (virtual allocation > capacity), and totals orphaned VDIs with
reclaimable bytes per SR.
Args:
target: Xen Orchestra target name from config; omit for the default.
Inputs: target.
task_listARead-only
[READ] List XO tasks, optionally filtered by status.
Returns {"tasks": [...], "returned": N, "limit": L, "truncated": bool}.
When truncated is true the task feed had more entries than were returned —
re-run with a higher limit or a narrower status filter.
Args:
status: Optional filter: pending / success / failure.
limit: Max tasks to return after filtering (default 200).
target: Xen Orchestra target name from config; omit for the default.
Inputs: limit, status, target.
undo_applyADestructive
[WRITE][risk=medium] Apply a recorded undo by dispatching its inverse tool.
The inverse runs through its own governed tool, so it is audited under its
own risk tier. Pass dry_run=True to preview the inverse call without
executing it. A token can only be applied once.
Args:
undo_id: The undoId from undo_list (or an ``_undo_id`` in a write result).
dry_run: If True, preview the inverse tool + params without running it.
target: Passed through to the inverse tool when it accepts a target.
Inputs: undo_id (required), target, dry_run.
undo_listARead-only
[READ] List recorded, not-yet-applied undo tokens (most recent first).
Each entry names the original tool, the inverse tool that ``undo_apply``
would run, and a human note. Use the ``undoId`` with ``undo_apply``.
One extra row is read so ``truncated`` is measured rather than guessed from
the row count matching the limit; when it is true, older tokens exist.
Each entry carries ``effectVerified``. False means the original write
lost its response, so the change it reverses is PROBABLE, not confirmed —
check the live state before applying, and do not report the result as a
restore of a state that may never have been reached.
Args:
limit: Max rows to return (default 50, capped at 500).
target: Unused (undo state is host-local); accepted for CLI uniformity.
Inputs: limit, target.
vdi_listARead-only
[READ] List VDIs (virtual disks), optionally per SR or orphaned-only.
Orphaned = attached to no VM (no VBD) — candidates for reclaiming space.
Returns {"vdis": [...], "returned": N, "limit": L, "truncated": bool};
when truncated is true there are more disks than were returned.
Args:
sr: Optional SR uuid to filter by.
orphaned_only: Only VDIs not attached to any VM.
limit: Max VDIs to return after filtering (default 200).
target: Xen Orchestra target name from config; omit for the default.
Inputs: sr, limit, target, orphaned_only.
vm_getARead-only
[READ] Detail for one VM by uuid (state, host, OS, tools, tags).
Args:
vm_id: VM uuid (see vm_list).
target: Xen Orchestra target name from config; omit for the default.
Inputs: vm_id (required), target.
vm_health_rcaARead-only
[READ][RCA] VM health root-cause analysis — cause + action per finding.
Flags VMs halted unexpectedly (auto-poweron / HA set), paused or suspended
VMs, running VMs without guest tools, and CPU / memory pressure from recent
stats. Analyze one VM (vm_id) or the whole fleet.
Args:
vm_id: Optional VM uuid to analyze just one VM.
target: Xen Orchestra target name from config; omit for the default.
Inputs: vm_id, target.
vm_listARead-only
[READ] List VMs with power state, host, tools status, sizing.
Returns {"vms": [...], "returned": N, "limit": L, "truncated": bool}.
When truncated is true there are more VMs than were returned — re-run with
a higher limit rather than treating this as the whole fleet.
Args:
power_state: Optional filter (Running / Halted / Paused / Suspended).
pool: Optional pool uuid to filter by.
limit: Max VMs to return after filtering (default 200).
target: Xen Orchestra target name from config; omit for the default.
Inputs: pool, limit, target, power_state.
vm_migrateADestructive
[WRITE][risk=medium] Live-migrate a VM to another host. Inverse: migrate back.
The REAL source host is captured BEFORE the move so the recorded undo
(migrate back to it) is replayable.
Args:
vm_id: VM uuid (see vm_list).
host_id: Destination host uuid (see host_list).
dry_run: If True, preview without migrating (no undo recorded).
target: Xen Orchestra target name from config; omit for the default.
Inputs: vm_id (required), host_id (required), target, dry_run.
vm_rebootADestructive
[WRITE][risk=medium] Reboot a VM (clean; hard with force). No undo.
A reboot has no meaningful inverse — the prior power state is captured for
the audit record only.
Args:
vm_id: VM uuid (see vm_list).
force: Hard reboot instead of a clean guest reboot.
dry_run: If True, preview without rebooting.
target: Xen Orchestra target name from config; omit for the default.
Inputs: vm_id (required), force, target, dry_run.
vm_startADestructive
[WRITE][risk=medium] Start a VM. Inverse: vm_stop.
Args:
vm_id: VM uuid (see vm_list).
dry_run: If True, preview without starting (no undo recorded).
target: Xen Orchestra target name from config; omit for the default.
Inputs: vm_id (required), target, dry_run.
vm_statsBRead-only
[READ] Recent CPU / memory stats for a VM (RRD-backed averages).
Args:
vm_id: VM uuid (see vm_list).
granularity: seconds / minutes / hours / days (default seconds).
target: Xen Orchestra target name from config; omit for the default.
Inputs: vm_id (required), target, granularity.
vm_stopADestructive
[WRITE][risk=medium] Stop a VM (clean shutdown; hard with force). Inverse: vm_start.
A clean shutdown needs the guest tools running; force maps to a hard
power-off. Captures the prior power state.
Refuses the VM declared as running Xen Orchestra (xo_self_vm_uuid on the
target) — stopping XO removes the API vm_start would travel over, so
recovery needs hypervisor console access. dry_run refuses it too: a preview
that returns green for a call that will be refused is a wrong preview. That
guard is exact and opt-in: an undeclared target is refused nothing, on
either path. The dry-run adds a weaker IP-based selfVmHint (null when there
is none) that is a coincidence to check, never a verdict and never a block —
XO's API exposes no self endpoint, so nothing here can be certain.
Args:
vm_id: VM uuid (see vm_list).
force: Hard power-off instead of a clean guest shutdown.
dry_run: If True, preview without stopping (no undo recorded).
target: Xen Orchestra target name from config; omit for the default.
Inputs: vm_id (required), force, target, dry_run.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 29 tools
Every tool targets a distinct resource+action: list vs get pairs are clear, the four RCA tools (backup_failure_rca, pool_patch_ha_posture, sr_usage_rca, vm_health_rca) are differentiated by domain, and destructive pairs (snapshot_delete vs snapshot_revert, vm_stop vs vm_reboot) are cleanly separated. The aggregate overview tool is clearly marked as a fan-out, so it does not blur the granular reads.
Consistent snake_case resource_action convention throughout (vm_list, vm_get, vm_start, host_list, sr_get, snapshot_create, undo_apply). The only deviations (overview, the *_rca tools) still follow the resource-first snake_case style and remain predictable.
29 tools is above the comfortable range and pushes into 'heavy' territory for a single server. The breadth (pools, hosts, VMs, SRs, VDIs, snapshots, backups, tasks, undo) justifies much of it, but the surface is at the upper limit of what an agent can navigate reliably.
Snapshot lifecycle (create/delete/revert) and VM power lifecycle (start/stop/reboot/migrate) are well covered, and the undo system is a strong addition. However, there is no VM create/delete/clone, no host maintenance or pool provisioning actions, and backup tools are read-only, leaving notable gaps for real ops workflows.
Maintenance
Related MCP Connectors
Governed MCP workflow automation for MSPs and agencies: approvals, durable state, signed evidence.
Remote MCP for Copilot CLI switch gate MCP, structured receipts, audit logs, and reviewer-ready evid
Remote MCP for Android CLI agent build gate, structured receipts, audit logs, and reviewer-ready evi
Capability-routed CLI/MCP: signed provenance exports, WCAG-validated themes, sovereign deploy modes.
Related MCP Servers
- AlicenseAqualityBmaintenanceGoverned Prometheus + Grafana operations — firing-alert and scrape-target RCA, alert noise/flapping analysis, silences, and dashboards, with unbypassable audit logging (MCP + CLI), budget/runaway guards, dry-run, and undo/rollback.39MIT
- AlicenseAqualityBmaintenanceGoverned OPNsense + pfSense firewall operations — gateway-health, rule-shadow, and blocked-traffic RCA, with guarded rule/alias writes, unbypassable audit logging (MCP + CLI), budget/runaway guards, dry-run, and undo/rollback.352MIT
- AlicenseAqualityBmaintenanceGoverned SSO/IAM operations for Keycloak and Authentik — login-failure, stale-permission, client-config, and MFA RCA, with unbypassable audit logging (MCP + CLI), budget/runaway guards, dry-run, and undo/rollback.29MIT
- AlicenseAqualityBmaintenanceGoverned CI/CD operations for self-hosted GitLab and Gitea — pipeline-failure, runner, artifact-bloat, and stale-branch RCA, with unbypassable audit logging (MCP + CLI), budget/runaway guards, dry-run, and undo/rollback.28MIT