From unifi-network
Runs a UniFi network health check diagnosing device status, connectivity issues, firmware updates, and system health via batch API calls.
How this skill is triggered — by the user, by Claude, or both
Slash command
/unifi-network:network-health-checkThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Before running a health check, verify the MCP server is configured:
Before running a health check, verify the MCP server is configured:
UNIFI_NETWORK_HOST is set in the environment.unifi-network-setup skill to configure the UniFi Network MCP server.unifi_tool_index to confirm available tools. If no UniFi tools are listed, the server is not connected.Use unifi_batch to gather all required data in a single parallel operation:
unifi_batch([
{ "tool": "unifi_get_system_info" },
{ "tool": "unifi_get_network_health" },
{ "tool": "unifi_list_devices" },
{ "tool": "unifi_list_alarms" }
])
This single batch call replaces sequential tool calls and returns all data needed for the report. Do not call these tools one at a time.
If device or alarm issues are found and more detail is needed, a follow-up batch can add:
unifi_batch([
{ "tool": "unifi_list_clients" },
{ "tool": "unifi_get_top_clients" }
])
Use these reference documents to interpret the data returned by the batch call:
references/device-states.md — maps device state integer codes to human-readable status (online, offline, isolated, etc.) and explains what each state means operationally. Do not guess at state codes — consult this reference before classifying device status.references/alarm-types.md — describes known alarm types, their severity levels, and recommended remediation steps. Consult before classifying alarm severity or suggesting actions.references/health-subsystems.md — explains the per-subsystem health fields returned by unifi_get_network_health (WAN, LAN, WLAN, VPN), how to interpret status values, and the recommended diagnostic priority order: WAN → LAN → WLAN → VPN.From the device list, identify:
state != 1. Check references/device-states.md for the full state code table.upgradeable field. Report current vs available firmware version.For each active alarm, classify severity using references/alarm-types.md and provide a plain-language explanation with remediation steps from that reference.
Present findings using this structure:
## Network Health Report
**Overall Status:** [Healthy / Warning / Critical]
**Controller:** [version] — uptime [X days]
### Devices ([online]/[total])
- [List any offline or problematic devices with their state code and meaning]
- [List devices needing firmware updates with current and available versions]
### Active Alarms ([count])
- [Summarize each alarm with severity and recommendation]
### Recommendations
1. [Actionable item]
2. [Actionable item]
A healthy network gets a brief "all clear" summary. Do not manufacture concerns for quiet periods.
unifi_batch for initial data gathering — sequential tool calls are significantly slower.unifi_get_network_health shows WAN health issues, that likely explains many downstream problems — lead with that finding and follow the WAN → LAN → WLAN → VPN diagnostic priority from references/health-subsystems.md.npx claudepluginhub sirkirby/unifi-mcp --plugin unifi-networkManages UniFi network infrastructure: inspect clients/devices/sites/WiFi configs/firewall rules/DPI stats/events; actions like block/reconnect/forget clients, restart/locate devices.
Manages UniFi network infrastructure: devices, clients, firewall, VPN, routing, WLANs, and traffic flows via 177 MCP tools. Lazy-loaded tool discovery with permission-gated mutations and preview-then-confirm safety model.
Manages Ubiquiti UniFi networks via unifly Rust CLI: VLANs, SSIDs, firewalls, NAT, DHCP reservations, clients, devices, events, stats, DPI, backups, raw API.