VMware AIops Extension

Add vmware-aiops MCP Server as a goose Extension for natural language VMware vCenter/ESXi operations

This tutorial covers how to add vmware-aiops as a goose extension, enabling natural language management of VMware vCenter and ESXi infrastructure. With this extension, goose can list VMs, check health, power on/off machines, deploy from templates, run commands inside VMs, and orchestrate multi-step operations with automatic rollback.

Quick Install

goose Desktop

Install vmware-aiops first:

uv tool install vmware-aiops
vmware-aiops mcp-config install --agent goose

goose CLI

uv tool install vmware-aiops
vmware-aiops mcp-config install --agent goose

Configuration

Prerequisites

You need uv and a running VMware vCenter Server or ESXi host. VMware Tools must be running inside guest VMs for Guest Operations features.

Step 1: Install and configure credentials

uv tool install vmware-aiops

mkdir -p ~/.vmware-aiops
cat > ~/.vmware-aiops/config.yaml << 'EOF'
targets:
  my-vcenter:
    host: vcenter.example.com
    username: [email protected]
    password_env: VMWARE_PASSWORD
    verify_ssl: false
EOF

echo "VMWARE_PASSWORD=your_password" > ~/.vmware-aiops/.env
chmod 600 ~/.vmware-aiops/.env

Step 2: Verify connectivity

vmware-aiops doctor

Step 3: Add to goose

goose Desktop

goose CLI

Example Usage

In this example, I'll use goose to perform a daily health check across my vCenter environment and remediate a disk space issue on a VM — all via natural language.

goose Prompt

Check the health of my vCenter environment. Show any active alarms and
powered-off VMs that haven't been used recently. If datastore usage is
above 85%, tell me which VMs are using the most space.

goose Output

I'll start with a full health check of your vCenter environment.

--- get_alarms ---

Found 2 active alarms:
- esxi-03: CPU ready > 5% (Warning) — sustained for 2 hours
- datastore02: 87% capacity used (Critical) ← needs attention

--- list_virtual_machines (power_state=poweredOff) ---

4 powered-off VMs found:
- vm-test-old (last powered on: 21 days ago)
- vm-dev-unused (last powered on: 14 days ago)
- vm-staging-v1 (last powered on: 8 days ago)
- vm-backup-runner (last powered on: 3 days ago)

--- browse_datastore (datastore02) ---

Largest consumers on datastore02:
1. vm-db-prod: 320 GB (snapshots not cleaned up)
2. vm-web-cluster: 180 GB
3. vm-test-old: 95 GB (powered off)

Recommendation: Delete vm-test-old (powered off 21 days, 95 GB) to free
space immediately. Also, vm-db-prod has stale snapshots — consolidating
them would recover ~120 GB.

Shall I create a plan to remove vm-test-old and consolidate vm-db-prod snapshots?

Yes, go ahead.

--- vm_create_plan ---

Plan created:
Step 1: Power off check for vm-test-old (already off ✓)
Step 2: Delete vm-test-old
Step 3: Consolidate snapshots on vm-db-prod

⚠️  Step 2 is irreversible. Step 3 is safe (read-only snapshot merge).

--- vm_apply_plan ---

Step 1/3: Verified vm-test-old is powered off ✓
Step 2/3: Deleted vm-test-old — freed 95 GB ✓
Step 3/3: Consolidated vm-db-prod snapshots — freed 118 GB ✓

datastore02 is now at 64% capacity. All alarms should clear within 5 minutes.