VMware AIops

AI-powered VMware vCenter and ESXi operations tool. Manage your entire VMware infrastructure using natural language through any AI coding assistant.

Need read-only monitoring only? Use VMware-Monitor — an independent repository with code-level safety (zero destructive code in the codebase). Install: clawhub install vmware-monitor

When to Use This Skill

• Query VM, host, datastore, cluster, and network inventory
• Check health status, active alarms, hardware sensors, and event logs
• Perform VM lifecycle operations: power on/off, create, delete, snapshot, clone, migrate
• Deploy VMs from OVA, templates, linked clones, or batch specs
• Browse datastores and discover ISO/OVA/VMDK images
• Monitor vSAN health, capacity, disk groups, and performance
• Access Aria Operations (VCF Operations) for historical metrics, anomaly detection, and capacity planning
• Manage vSphere Kubernetes Service (VKS) clusters
• Run scheduled scanning with webhook notifications (Slack, Discord)

Quick Install

All install methods fetch from the same source: github.com/zw008/VMware-AIops (MIT licensed). We recommend reviewing the source code before installing.

# Via Skills.sh (fetches from GitHub)
npx skills add zw008/VMware-AIops

# Via ClawHub (fetches from ClawHub registry snapshot of GitHub)
clawhub install vmware-aiops

# Via PyPI (recommended for version pinning)
uv tool install vmware-aiops==0.6.0

Claude Code

/plugin marketplace add zw008/VMware-AIops
/plugin install vmware-ops
/vmware-ops:vmware-aiops

Usage Mode

Choose the best mode based on your environment:

Calling Priority

• MCP-native tools (Claude Code, Cursor): MCP first, CLI fallback
• Local models / Token-sensitive: CLI first (MCP not needed)
• All other tools: CLI first

Tip: For token-sensitive scenarios, use CLI mode — the AI reads SKILL.md (~2K tokens) and calls commands via Bash. MCP mode loads all 31 tool schemas (~10K tokens) into context on every conversation turn.

CLI Examples

# Activate venv first
source /path/to/VMware-AIops/.venv/bin/activate

# Inventory
vmware-aiops inventory vms --target home-esxi
vmware-aiops inventory hosts --target home-esxi

# Health
vmware-aiops health alarms --target home-esxi

# VM operations
vmware-aiops vm info my-vm --target home-esxi
vmware-aiops vm power-on my-vm --target home-esxi

MCP Mode (Optional)

For Claude Code / Cursor users who prefer structured tool calls, add to ~/.claude/settings.json:

{
  "mcpServers": {
    "vmware-aiops": {
      "command": "/path/to/VMware-AIops/.venv/bin/python",
      "args": ["-m", "mcp_server"],
      "cwd": "/path/to/VMware-AIops",
      "env": {
        "VMWARE_AIOPS_CONFIG": "~/.vmware-aiops/config.yaml"
      }
    }
  }
}

MCP exposes 31 tools across 6 categories. All accept optional target parameter.

list_virtual_machines auto-compacts when inventory exceeds 50 VMs (returns compact fields only). Use limit or fields to override.

Architecture

User (Natural Language)
  ↓
AI Tool (Claude Code / Aider / Gemini / Codex / Cursor / Trae / Kimi)
  ↓
  ├─ CLI mode (default): vmware-aiops CLI ──→ pyVmomi ──→ vSphere API
  │
  └─ MCP mode (optional): MCP Server (stdio) ──→ pyVmomi ──→ vSphere API
  ↓
vCenter Server ──→ ESXi Clusters ──→ VMs
    or
ESXi Standalone ──→ VMs

Capabilities

1. Inventory

2. Health & Monitoring

Monitored Event Types:

3. VM Lifecycle

Guest Operations require VMware Tools running inside the guest OS.

Plan → Apply (Multi-step Operations)

For complex operations involving 2+ steps or 2+ VMs, use the plan/apply workflow:

Plans are stored in ~/.vmware-aiops/plans/, deleted on success, auto-cleaned after 24h.

4. VM Deployment & Provisioning

5. Datastore Browser

6. vSAN Management

Requires pyVmomi 8.0.3+ (vSAN SDK merged). For older versions, install the standalone vSAN Management SDK.

7. Aria Operations (VCF Operations)

REST API at /suite-api/. Auth: vRealizeOpsToken. Rebranded as VCF Operations in VCF 9.0.

8. vSphere Kubernetes Service (VKS)

Kubernetes-native API via kubectl/kubeconfig. VKS 3.6+ uses Cluster API specification.

9. Scheduled Scanning & Notifications

Safety Features

Version Compatibility

pyVmomi auto-negotiates the API version during SOAP handshake — no manual configuration needed.

Supported AI Platforms

MCP Server — Local Agent Compatibility

The MCP server works with any MCP-compatible agent via stdio transport. Config templates in examples/mcp-configs/:

# Example: Aider + Ollama (fully local, no cloud API)
aider --conventions codex-skill/AGENTS.md --model ollama/qwen2.5-coder:32b

CLI Reference

# Diagnostics
vmware-aiops doctor [--skip-auth]

# MCP Config Generator
vmware-aiops mcp-config generate --agent <goose|cursor|claude-code|continue|vscode-copilot|localcowork|mcp-agent>
vmware-aiops mcp-config list

# Inventory
vmware-aiops inventory vms [--target <name>] [--limit <n>] [--sort-by name|cpu|memory_mb|power_state] [--power-state poweredOn|poweredOff]
vmware-aiops inventory hosts [--target <name>]
vmware-aiops inventory datastores [--target <name>]
vmware-aiops inventory clusters [--target <name>]

# Health
vmware-aiops health alarms [--target <name>]
vmware-aiops health events [--hours 24] [--severity warning]

# VM Operations
vmware-aiops vm info <vm-name>
vmware-aiops vm power-on <vm-name>
vmware-aiops vm power-off <vm-name> [--force]
vmware-aiops vm create <name> [--cpu <n>] [--memory <mb>] [--disk <gb>]
vmware-aiops vm delete <vm-name> [--confirm]
vmware-aiops vm reconfigure <vm-name> [--cpu <n>] [--memory <mb>]
vmware-aiops vm snapshot-create <vm-name> --name <snap-name>
vmware-aiops vm snapshot-list <vm-name>
vmware-aiops vm snapshot-revert <vm-name> --name <snap-name>
vmware-aiops vm snapshot-delete <vm-name> --name <snap-name>
vmware-aiops vm clone <vm-name> --new-name <name>
vmware-aiops vm migrate <vm-name> --to-host <host>
vmware-aiops vm set-ttl <vm-name> --minutes <n>
vmware-aiops vm cancel-ttl <vm-name>
vmware-aiops vm list-ttl
vmware-aiops vm clean-slate <vm-name> [--snapshot baseline]

# Guest Operations (requires VMware Tools)
vmware-aiops vm guest-exec <vm-name> --cmd /bin/bash --args "-c 'ls -la /tmp'" --user root
vmware-aiops vm guest-upload <vm-name> --local ./script.sh --guest /tmp/script.sh --user root
vmware-aiops vm guest-download <vm-name> --guest /var/log/syslog --local ./syslog.txt --user root

# Plan → Apply (multi-step operations)
vmware-aiops plan list

# Deploy
vmware-aiops deploy ova <path> --name <vm-name> [--datastore <ds>] [--network <net>]
vmware-aiops deploy template <template-name> --name <vm-name> [--datastore <ds>]
vmware-aiops deploy linked-clone --source <vm> --snapshot <snap> --name <new-name>
vmware-aiops deploy iso <vm-name> --iso "[datastore] path/file.iso"
vmware-aiops deploy mark-template <vm-name>
vmware-aiops deploy batch-clone --source <vm> --count <n> [--prefix <prefix>]
vmware-aiops deploy batch <spec.yaml>

# Datastore
vmware-aiops datastore browse <ds-name> [--path <subdir>]
vmware-aiops datastore scan-images [--target <name>]
vmware-aiops datastore images [--type ova|iso|vmdk] [--ds <name>]

# vSAN
vmware-aiops vsan health [--target <name>]
vmware-aiops vsan capacity [--target <name>]
vmware-aiops vsan disks [--target <name>]
vmware-aiops vsan performance [--hours 1]

# Aria Operations
vmware-aiops ops alerts [--severity critical]
vmware-aiops ops metrics <resource-name> [--hours 24]
vmware-aiops ops recommendations [--target <name>]
vmware-aiops ops capacity <cluster-name>

# VKS (Kubernetes)
vmware-aiops vks clusters [--namespace default]
vmware-aiops vks health <cluster-name>
vmware-aiops vks scale <machine-deployment> --replicas <n>
vmware-aiops vks nodes <cluster-name>

# Scanning & Daemon
vmware-aiops scan now [--target <name>]
vmware-aiops daemon start
vmware-aiops daemon stop
vmware-aiops daemon status

Setup

# 1. Install from PyPI (source: github.com/zw008/VMware-AIops)
uv tool install vmware-aiops

# 2. Verify installation source
vmware-aiops --version  # confirms installed version

# 3. Configure
mkdir -p ~/.vmware-aiops
vmware-aiops init  # generates config.yaml and .env templates
chmod 600 ~/.vmware-aiops/.env
# Edit ~/.vmware-aiops/config.yaml and .env with your target details

What Gets Installed

The vmware-aiops package installs a Python CLI binary and its dependencies (pyVmomi, Click, Rich, APScheduler, python-dotenv). No background services, daemons, or system-level changes are made during installation. The scheduled scanner (daemon start) only runs when explicitly started by the user.

Development Install

git clone https://github.com/zw008/VMware-AIops.git
cd VMware-AIops
uv venv && source .venv/bin/activate
uv pip install -e .

Security

• Source Code: Fully open source at github.com/zw008/VMware-AIops (MIT). The uv installer fetches the vmware-aiops package from PyPI, which is built from this GitHub repository. We recommend reviewing the source code and commit history before deploying in production.

• TLS Verification: Enabled by default. The disableSslCertValidation option exists solely for ESXi hosts using self-signed certificates in isolated lab/home environments. In production, always use CA-signed certificates with full TLS verification.

• Credentials & Config: This skill requires the following secrets, all stored in ~/.vmware-aiops/.env (chmod 600, loaded via python-dotenv):

  • VSPHERE_USER — vCenter/ESXi service account username
  • VSPHERE_PASSWORD — service account password
  • (Optional) Webhook URLs for Slack/Discord notifications

  The config file ~/.vmware-aiops/config.yaml stores only target hostnames, ports, and a reference to the .env file — it does not contain passwords or tokens. The env var VMWARE_AIOPS_CONFIG points to this YAML file.

• Webhook Data Scope: Webhook notifications are disabled by default. When enabled, they send infrastructure health summaries (alarm counts, event types, host status) to user-configured URLs only (Slack, Discord, or any HTTP endpoint you control). No data is sent to third-party services. Webhook payloads contain no credentials, IPs, or personally identifiable information — only aggregated alert metadata.

• Prompt Injection Protection: All vSphere-sourced content (event messages, host logs) is truncated, stripped of control characters, and wrapped in boundary markers ([VSPHERE_EVENT]/[VSPHERE_HOST_LOG]) before output to prevent prompt injection when consumed by LLM agents.

• Least Privilege: Use a dedicated vCenter service account with minimal permissions. For monitoring-only use cases, prefer the read-only VMware-Monitor skill which has zero destructive code paths.

License

MIT