nutanix-mcp
markdown
# nutanix-mcp
MCP server for Nutanix Prism Central v4 REST API, the management plane for AOS deployments. 8 tools for AI agents to list and inspect VMs, control VM power state (env-gated), list clusters, read alerts, and list subnets.
## Tools
| Tool | Description |
|------|-------------|
| `list_vms` | List all VMs visible to the Prism Central deployment. Returns names, ext IDs, power state, host, cluster, memory and CPU config. |
| `get_vm` | Get full metadata for a single VM by ext_id. |
| `power_on_vm` | Transition a VM from Off to On. Requires `NUTANIX_ALLOW_POWER=true`. |
| `power_off_vm` | Transition a VM from On to Off (graceful shutdown). Requires `NUTANIX_ALLOW_POWER=true`. |
| `list_clusters` | Inventory all clusters under Prism Central. |
| `get_cluster` | Read full config and health of a single cluster. |
| `list_alerts` | Stream the current alert backlog from Prism Central. Used by an on-call agent to triage what is unhealthy across the fleet. |
| `list_subnets` | Inventory all networking subnets visible to Prism Central. Used by an agent placing a new VM to pick the right network. |
## Install
TDQS
Scored across 8 tools
Each tool targets a distinct resource and action: cluster vs VM vs alert vs subnet, with clear read/list vs power control boundaries. No two tools could be confused for the same purpose.
All tool names follow a consistent verb_noun pattern with underscores (e.g., list_vms, power_on_vm). No mixing of styles or vague verbs.
8 tools cover the core operations for Nutanix infrastructure: inventory (list), detail retrieval (get), alert monitoring, and power management. The scope is well-defined without being excessive or too sparse.
While the tool set covers essential read operations and power control, it lacks create, update, or delete for VMs and clusters, which are common tasks. The surface is incomplete for full lifecycle management.