io.github.shigechika/aruba-central-mcp
A read-only MCP server that exposes Aruba Central (GreenLake New Central, optionally Classic Central) network monitoring data to AI assistants over STDIO.
Access points: list APs (filter by site/status), per-AP status, radios/RF details (channel, utilization, noise floor, TX power), BSSIDs, WLANs, and swarms; AP throughput time series; top APs by bandwidth (wireless/wired/total).
Clients: list connected wireless clients (filter by SSID/band), look up one client by MAC, roaming/mobility trail, client count trends grouped by type/role/VLAN/WLAN/radio/security/protocol, and top clients by usage.
Infrastructure: list switches, aggregated per-site summary (AP counts, online/offline, client counts), and a session-start
health_checkreporting version, base URL, auth state, and optional Classic Central token status.Firmware: per-swarm running versions, recommended version, upgrade status, group/site, plus Classic group compliance version, scheduled time, upgrade state/reason and stale swarms; available firmware versions per device type or swarm (Classic only).
Operations:
daily_briefmorning check flagging sites whose offline-AP ratio exceeds a threshold.Query features: server-side OData filtering, automatic pagination, OAuth2 client-credentials auth with token auto-refresh, STDIO transport, and
--check/--versionCLI options. Nothing is configured or upgraded — every tool only reads.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@io.github.shigechika/aruba-central-mcp list all access points in site HQ"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
aruba-central-mcp
English | 日本語
MCP server for Aruba Central (GreenLake New Central API).
Exposes access point, switch, and wireless client status to MCP-compatible AI assistants (Claude Code, Claude Desktop, OpenAI Codex, etc.) via STDIO transport.
Documentation: https://shigechika.github.io/aruba-central-mcp/
Features
Access Points
Tool | Description |
| List all access points (with optional site/status filter) |
| List AP radios (channel, utilization, noise floor, TX power) |
| List all BSSIDs |
| List WLANs (SSID, security, VLAN) |
| List AP swarms/clusters |
| Get detailed status of a specific AP |
| Get AP throughput trend (TX/RX over time) |
| Top APs by bandwidth usage (wireless/wired/total) |
Clients
Tool | Description |
| List connected wireless clients (with optional SSID/band filter) |
| Find a client by MAC address (direct API lookup) |
| Client count trend over time |
| Top clients by bandwidth usage |
| Client roaming history |
Infrastructure
Tool | Description |
| List all switches |
| Aggregated site-level summary (AP counts, client counts) |
| Report server version and verify Aruba Central authentication (no data fetch) |
| One-shot morning AP health check: sites whose offline-AP ratio exceeds a threshold |
Firmware
Tool | Description |
| Per swarm: group, site, running versions, recommended version, upgrade status; filter by group, site and/or swarm. With Classic Central configured, also the Classic group, the group's compliance version (what Central upgrades the swarm to next) with its scheduled time, Classic's upgrade state and reason, and Classic-only (stale) swarms |
| Firmware versions Classic Central offers for a device type or one swarm (registered only when Classic Central is configured) |
Every tool reads; nothing is configured or upgraded. See docs/firmware-api-notes.md for which API reports what.
Highlights
Server-side OData filtering for efficient queries
OAuth2 Client Credentials authentication (GreenLake SSO)
Automatic pagination for large result sets
Token auto-refresh before expiration
Lightweight: only
mcpSDK +httpx(no pandas)
Related MCP server: wlanpi-mcp
Prerequisites
Python 3.11+
Aruba Central account with API access (GreenLake New Central API)
OAuth2 client credentials (client ID and secret)
Setup
# uv
uv pip install aruba-central-mcp
# pip
pip install aruba-central-mcpOr run without installing:
uvx aruba-central-mcpFrom source:
git clone https://github.com/shigechika/aruba-central-mcp.git
cd aruba-central-mcp
# uv
uv sync
# pip
pip install -e .Configuration
Set the following environment variables:
Variable | Description | Example |
| API gateway URL |
|
| OAuth2 client ID |
|
| OAuth2 client secret |
|
Optional: Classic Central (firmware details)
AOS-8 Instant APs in Classic Central groups are upgraded from Classic Central, and only its API reports why a swarm's upgrade failed, the group compliance version and the versions on offer. Set all four to enable that:
Variable | Description |
| Classic API gateway host, |
| Client id of the API-gateway application |
| Its client secret |
| JSON file holding the token pair; this server rewrites it on every refresh |
Bootstrap the token file once with the access and refresh token Classic Central's API Gateway issues for that application, readable only by the user the server runs as:
umask 077
printf '{"access_token": "%s", "refresh_token": "%s"}\n' "$ACCESS" "$REFRESH" > /path/to/classic-token.jsonThe refresh token is replaced on every refresh (the old one stops working), so
let only this server refresh it, and point every process that runs it at the
same file: refreshes are serialised with a lock file next to it. When the
refresh token is rejected, health_check reports classic: error and asks for
a new pair; when only some ARUBA_CLASSIC_* variables are set it reports
classic: misconfigured and names the missing ones. The directory holding the
file must be writable by that user (the new pair is written next to it, then
renamed); without that, the server refuses to refresh rather than lose the
rotated token. Windows has no cross-process lock here, so run one process there.
How to obtain API credentials
Log in to HPE GreenLake Platform
Go to Manage Workspace > Personal API clients
Click Create Personal API client
Enter a nickname and select Aruba Central as the service
Copy the
client_idandclient_secret— the secret is shown only once
For details, see:
Usage
Claude Code (plugin)
This repository doubles as a single-plugin marketplace, so Claude Code can install the server for you:
/plugin marketplace add shigechika/aruba-central-mcp
/plugin install aruba-central-mcp@aruba-central-mcpThe plugin launches uvx aruba-central-mcp and reads the same three environment
variables described in Configuration; export them before
starting Claude Code. It also passes the four optional ARUBA_CLASSIC_*
variables through, so firmware_status gets its Classic details and
firmware_versions is offered when they are exported; left unset they arrive
empty and Classic stays off.
uvx must be on the PATH of the process that runs Claude Code — a login
shell usually has it, but a GUI-launched app may not; install
uv system-wide if the plugin fails to start.
Claude Code (manual)
claude mcp add aruba-central \
-e ARUBA_CENTRAL_BASE_URL=apigw-uswest4.central.arubanetworks.com \
-e ARUBA_CENTRAL_CLIENT_ID=your-client-id \
-e ARUBA_CENTRAL_CLIENT_SECRET=your-client-secret \
-- uvx aruba-central-mcpOr add to .mcp.json:
{
"mcpServers": {
"aruba-central": {
"command": "uvx",
"args": ["aruba-central-mcp"],
"env": {
"ARUBA_CENTRAL_BASE_URL": "apigw-uswest4.central.arubanetworks.com",
"ARUBA_CENTRAL_CLIENT_ID": "your-client-id",
"ARUBA_CENTRAL_CLIENT_SECRET": "your-client-secret"
}
}
}
}Claude Desktop
Add to claude_desktop_config.json:
{
"mcpServers": {
"aruba-central": {
"command": "uvx",
"args": ["aruba-central-mcp"],
"env": {
"ARUBA_CENTRAL_BASE_URL": "apigw-uswest4.central.arubanetworks.com",
"ARUBA_CENTRAL_CLIENT_ID": "your-client-id",
"ARUBA_CENTRAL_CLIENT_SECRET": "your-client-secret"
}
}
}
}Codex
codex mcp add aruba-central \
--env ARUBA_CENTRAL_BASE_URL=apigw-uswest4.central.arubanetworks.com \
--env ARUBA_CENTRAL_CLIENT_ID=your-client-id \
--env ARUBA_CENTRAL_CLIENT_SECRET=your-client-secret \
-- uvx aruba-central-mcpAny other MCP client that launches stdio servers works the same way: give it
uvx aruba-central-mcp as the command and set the three ARUBA_CENTRAL_*
variables in its environment.
Direct execution
export ARUBA_CENTRAL_BASE_URL="apigw-uswest4.central.arubanetworks.com"
export ARUBA_CENTRAL_CLIENT_ID="your-client-id"
export ARUBA_CENTRAL_CLIENT_SECRET="your-client-secret"
python3 -m aruba_central_mcpCLI Options
aruba-central-mcp --version # Print version and exit
aruba-central-mcp --help # Show usage and required environment variables
aruba-central-mcp --check # Verify environment variables and OAuth2 authentication, then exit
aruba-central-mcp # Start MCP server (STDIO, default)With no options, the process runs as an MCP STDIO server (the mode used by MCP clients).
--check exit codes: 0 success, 1 config error, 2 auth error.
Development
git clone https://github.com/shigechika/aruba-central-mcp.git
cd aruba-central-mcp
# uv
uv sync --dev
uv run pytest -v
# pip
python3 -m venv .venv
.venv/bin/pip install -e ".[test]"
.venv/bin/pytest -vLive smoke test
The unit suite mocks Central at the transport layer, which is what makes it
fast — and also what makes it blind to a tool that has stopped returning real
data. scripts/smoke_test.py runs every registered tool against the
configured tenant and fails on empty, malformed or error answers:
# needs the same ARUBA_CENTRAL_* environment variables as the server
uv run python scripts/smoke_test.py
uv run python scripts/smoke_test.py --only radios --tracebackRead-only. Every tool here reads; nothing in Central is configured. A future tool that writes must be listed as state-changing and skipped, and a test enforces that.
No payloads in the report. Tool names, statuses and row counts only; error text is redacted too, since an error routinely quotes the device, client MAC or site it was asked about.
Nothing network-specific in the specs. The AP, the serial number and the client MAC that the per-device tools need are discovered at run time from the listings, and skipped when the network has none to offer. Two tests keep it that way: one refuses those parameters as literals, the other bans anything address-shaped anywhere in the file, because this repository is public.
Empty answers pass for the listings and the time-series tools — a site with no swarms configured is a real deployment — but a lookup handed a name discovered seconds earlier must not come back empty, and those probes say so.
CI enforces the cheap half: a tool registered without a probe spec fails the build (
tests/test_smoke_probes.py), so adding a tool forces the question "how would we know it works?".scripts/smoke_harness.pyis the engine and holds no Central knowledge: it is kept identical across the servers that share it, so fix engine bugs once and sync the file rather than patching this copy.
Its first run found a real one: get_client_mobility_trail was requesting a
page size the endpoint rejects, so the tool had been failing for every client.
API Reference
This server uses the GreenLake New Central API:
/network-monitoring/v1/aps— Access points/network-monitoring/v1/radios— AP radios/network-monitoring/v1/bssids— BSSIDs/network-monitoring/v1/wlans— WLANs/network-monitoring/v1/swarms— AP swarms/clusters/network-monitoring/v1/switches— Switches/network-monitoring/v1/clients— Clients/network-monitoring/v1/clients-trend— Client count trends/network-monitoring/v1/clients-topn-usage— Top clients by usage/network-monitoring/v1/top-aps-by-usage— Top APs by usage
License
Available Tools
18 toolsdaily_briefA
Run a morning AP health check across all sites.
Fetches all access points and aggregates their online/offline status by
site. Sites whose offline AP ratio is strictly greater than
offline_threshold percent are flagged as WARNING (uses >, so a
site with exactly offline_threshold % offline is still OK).
Args: offline_threshold: Percentage of offline APs that triggers a WARNING (default 10.0). Sites at or below this threshold appear in the OK section. Pass 0.0 to flag any site with at least one offline AP.
Output tiers:
CRITICAL — API connection failure
WARNING — offline AP ratio strictly above threshold
OK — offline ratio at or below threshold
Returns a Markdown summary with site-level AP counts and anomaly details.
Inputs: offline_threshold. Returns: result
find_client_by_macA
Look up a single client by MAC address (wireless or wired).
Returns detailed info including OS, manufacturer, VLAN, and connection status. Uses direct API lookup (no full-list scan).
Args: mac_address: Client MAC address (e.g. "aa:bb:cc:dd:ee:ff").
Inputs: mac_address (required). Returns: result
firmware_statusA
Firmware of every AP swarm (cluster): running versions, upgrade status, group and site.
Use this to see which swarms still need a firmware upgrade and what Central will do next. Read-only. The whole call is bounded (about 48 s); what could not be read in time is named under Notes.
One line per swarm with its group(s) and site(s), AP count and how many are online, the running versions with AP counts, the recommended version and New Central's per-AP upgrade status (/network-services/v1/firmware-details). APs that belong to no swarm are listed per group and site.
When Classic Central is configured (ARUBA_CLASSIC_*), each swarm also gets
its Classic group, the group's compliance version (the version Central
keeps the group's swarms on: an upgrade of a Classic IAP group happens when
the compliance version changes, and scheduled is when that is set to
run; a site-level upgrade in the UI does not change it), marked "matches
running" or "pending", and "older than ..." when it is lower than the
running or recommended version, plus Classic's upgrade state,
reason and upgrade_required. New Central reports no group for
AOS-8 APs, so the group comes from Classic; Classic data is cached per
group for a few minutes. Swarms Classic knows but New Central does not
(e.g. 0 APs, Down) are listed at the end when no site or swarm filter is
given. A Classic failure is reported under Notes; the New Central data is
still shown.
Filters combine (all must match); each is an exact name, case-insensitive.
Empty means no filter. group uses the Classic group when New Central has none.
Args: group: Central group name. site: Site name. swarm_name: Swarm (cluster) name.
Inputs: site, group, swarm_name. Returns: result
get_ap_statusB
Get detailed status of a specific access point by name.
Args: ap_name: AP device name (case-insensitive).
Inputs: ap_name (required). Returns: result
get_ap_throughputA
Get throughput trend for a specific AP as a time-series table.
Returns timestamped TX/RX throughput in bps. Defaults to last 3 hours.
Args: serial_number: AP serial number (use list_aps or get_ap_status to find it). interface_type: "WIRELESS" (Wi-Fi), "WIRED" (Ethernet uplink), or "LTE". start_at: Start time in RFC 3339 format. Defaults to 3 hours ago. end_at: End time in RFC 3339 format. Defaults to current time.
Inputs: serial_number (required), end_at, start_at, interface_type. Returns: result
get_client_mobility_trailA
Get roaming history for a wireless client as a time-series table.
Shows which APs the client connected to, when, and on which SSID. Useful for troubleshooting connectivity issues or tracking movement. Defaults to last 3 hours.
Args: mac_address: Client MAC address (e.g. "aa:bb:cc:dd:ee:ff"). start_at: Start time in RFC 3339 format. Defaults to 3 hours ago. end_at: End time in RFC 3339 format. Defaults to current time.
Inputs: mac_address (required), end_at, start_at. Returns: result
get_clients_trendA
Get client count trend over time, grouped by a chosen dimension.
Useful for capacity planning and usage pattern analysis.
Args: site_id: Filter by site ID. Empty for all. site_name: Filter by site name. Empty for all. start_at: Start time in RFC 3339 format (max 1 month range). Must not be in the future. end_at: End time in RFC 3339 format. Must not be in the future. group_by: Dimension to group by. Supported values: - Any client_type: TYPE, ROLE, VLAN - WIRELESS only: WLAN, RADIO, SECURITY, PROTOCOL Default is TYPE. client_type: Client category: ALL, WIRELESS, or WIRED. Default is ALL. Must be WIRELESS when group_by is WLAN, RADIO, SECURITY, or PROTOCOL (the API returns 400 otherwise).
Inputs: end_at, site_id, group_by, start_at, +2 more. Returns: result
get_site_summaryA
Get a summary of all sites with AP and client counts.
Aggregates data across all APs and clients, grouped by site name. Shows total APs, online/offline counts, and client count per site.
Inputs: no named parameters. Returns: result
get_top_apsA
Get top access points ranked by bandwidth usage over a time range.
Returns a ranked list of APs sorted by descending bandwidth consumption.
Args: usage_type: "wireless" (Wi-Fi traffic), "wired" (uplink traffic), or "total" (wireless + wired combined). Default: "total". site_id: Filter by site ID. Empty for all sites. limit: Maximum number of APs to return (1-25, default 10). start_at: Start time in RFC 3339 format (e.g. "2025-01-01T00:00:00Z"). Max 30 days range. Defaults to 24 hours ago if omitted. end_at: End time in RFC 3339 format. Defaults to current time if omitted.
Inputs: limit, end_at, site_id, start_at, +1 more. Returns: result
get_top_clients_by_usageA
Get top clients ranked by bandwidth usage.
Args: site_id: Filter by site ID. Empty for all. site_name: Filter by site name. Empty for all. start_at: Start time in RFC 3339 format (max 1 month range). end_at: End time in RFC 3339 format. limit: Maximum number of clients to return (1-100, default 5).
Inputs: limit, end_at, site_id, start_at, +1 more. Returns: result
health_checkA
Report server version and Aruba Central backend authentication.
Call this at session start (or after a tool-call timeout) to confirm the MCP is up, see which version is running, and verify the Aruba Central backend can be authenticated. Lightweight: it builds the client and obtains an OAuth2 access token (GreenLake SSO, reusing the cached token) — it does NOT fetch APs, switches, clients, or any other data endpoint.
Always returns the same keys: status (healthy / degraded / error),
service, version, base_url (the configured
ARUBA_CENTRAL_BASE_URL, empty string if unset), auth (ok / error /
missing-env) and classic (not-configured / ok / error / misconfigured:
the optional Classic Central token used by the firmware tools;
misconfigured means some ARUBA_CLASSIC_* variables are set and others
are not). error and misconfigured make the status degraded and add
classic_detail. On a degraded or error result, detail carries the
reason (the New Central one first when both fail).
Inputs: no named parameters.
list_apsA
List access points with device-level overview.
Use this for AP inventory and availability checks. For RF details (channel, utilization, noise floor), use list_radios instead.
Args: site: Filter by site name (exact match, server-side). Empty for all. status: Filter by status (e.g. "ONLINE", "OFFLINE"). Empty for all.
Inputs: site, status. Returns: result
list_bssidsA
List all BSSIDs (per-radio, per-SSID broadcast identifiers).
Useful for identifying which radio on which AP is broadcasting a specific SSID. Each AP radio broadcasts one BSSID per WLAN.
Args: site: Filter by site name (exact match, server-side). Empty for all.
Inputs: site. Returns: result
list_clientsA
List connected wireless clients with connection details.
Returns client name, MAC, IP, SSID, band, SNR, connected AP, and auth type. For a single client's full details, use find_client_by_mac instead.
Args: ssid: Filter by SSID name (exact match, server-side). Empty for all. band: Filter by wireless band (exact match, server-side). Empty for all.
Inputs: band, ssid. Returns: result
list_radiosA
List AP radios with RF details.
Use this for RF troubleshooting: channel assignment, channel utilization, noise floor, and TX power. Each AP typically has 2-3 radios (one per band).
Args: site: Filter by site name (exact match, server-side). Empty for all. band: Filter by band (e.g. "2.4 GHz", "5 GHz"). Empty for all.
Inputs: band, site. Returns: result
list_swarmsA
List all AP swarms (clusters) from Aruba Central.
Returns cluster name, conductor AP, site, IP, and firmware version.
Args: site: Filter by site name (exact match, server-side). Empty for all.
Inputs: site. Returns: result
list_switchesB
List all switches from Aruba Central.
Returns switch name, status, model, type, IP, firmware, and MAC address.
Inputs: no named parameters. Returns: result
list_wlansA
List configured WLANs (SSIDs) with security and VLAN settings.
Shows each WLAN's name, band, operational status, security level, and VLAN assignment.
Args: site_id: Filter by site ID. Empty for all. serial_number: Filter by AP serial number. Empty for all.
Inputs: site_id, serial_number. Returns: result
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 18 tools
Most tools have clearly distinct resource/action targets, and several descriptions explicitly cross-reference alternatives such as list_aps vs list_radios. A few pairs still have soft boundaries: list_aps vs get_ap_status, and get_site_summary vs daily_brief.
Names are consistently snake_case and mostly follow list_/get_/find_ verb_noun patterns. The report-style names health_check, firmware_status, and daily_brief are minor deviations but remain readable and clear.
At 18 tools the server is above the typical 3-15 sweet spot, but the broad Aruba Central monitoring domain justifies a fairly wide surface. Most tools cover a distinct resource or analytical view, so the count does not feel bloated.
The toolset covers core read-only monitoring well: AP/client/switch/WLAN inventory, RF details, throughput and client trends, firmware status, health checks, and site summaries. Minor gaps remain around alerts/events, per-switch port detail, and configuration-changing operations if those are in scope.
Maintenance
Related MCP Connectors
MCP server that lets your AI assistant build and deploy apps, hosting each one at a live URL.
An MCP server that provides an API to LLMs to manage their JumpCloud resources.
MCP server that lets AI assistants use all OneSchema features exposed via the public API.
MCP server for OpenAI API (chat completions, image generation, embeddings) via AceDataCloud
Related MCP Servers
- FlicenseNot gradedqualityDmaintenanceA demo MCP server that connects an AI to external systems by exposing Python functions as tools and data as resources, using the stdio transport.-

wlanpi-mcpofficial
AlicenseBqualityBmaintenanceAn MCP server that exposes WLAN Pi capabilities like device info, Wi-Fi scanning, service management, and VLANs to AI assistants via the wlanpi-core REST API.716BSD 3-Clause- FlicenseNot gradedqualityDmaintenanceProduction-grade MCP server that exposes the complete HPE Aruba Networking Central REST API as 90 tools across 19 categories, enabling AI assistants to manage Aruba Central networks.-
- AlicenseNot gradedqualityDmaintenanceMCP server that exposes HPE Aruba EdgeConnect Orchestrator SD-WAN infrastructure as LLM-callable tools for listing appliances, reading interface/tunnel/bandwidth state, and changing HPE SSE tunnel state.1Apache 2.0