manage_listFabricsLABSCO SUMMARY
Read-only is the default, and the audit log is not optional.
EDIT_MODE_ENABLED starts false, so a fresh install cannot change your fabric even if the model tries. Around that sits the part most single-binary servers skip: users and roles with admin, operator and viewer levels, Fernet-encrypted credential storage, and a complete audit trail with client IP that you can export as CSV. If several people are pointing assistants at the same network, that plumbing is the reason to pick this over a thin API wrapper.
What it is
A Nexus Dashboard server that exposes the platform's REST operations as MCP tools, fronted by a web management interface for users, roles, clusters and audit logs. It ships as a Docker Compose stack with PostgreSQL behind it.
What you get
- Fabrics:
manage_listFabrics,manage_getFabricDetails,manage_createFabric,manage_updatefabricDetails,manage_deleteFabric - Switches:
manage_listFabricSwitches,manage_getFabricSwitch,manage_getSwitchInterfacesOverview,manage_getSwitchConfigurationDiff, plus actions likemanage_executeRediscoverSwitchesandmanage_executeSwitchConfigurationPreview - Networks, VRFs and interfaces each get list, get, summary, create, replace and delete —
manage_listNetworks,manage_getVrf,manage_listInterfacesand their siblings manage_listAnomaliesand the compliance rule tools cover anomalies and compliance;manage_listTemplatescovers templates- Every GET is available in read-only mode; POST, PUT and DELETE require edit mode to be switched on
- Coverage spans five Nexus Dashboard APIs — Manage, Analyze, Infrastructure, OneManage and Orchestration
Requirements
Docker 20.10+ with Docker Compose 2.0+, and a Cisco Nexus Dashboard 4.1+ with NDFC 12.x. Set CERT_SERVER_IP to your server's address, plus ENCRYPTION_KEY and SESSION_SECRET_KEY for production. Self-signed certificates are generated on first start. The stack listens on 7443 for the web UI, 8444 for the API and MCP SSE endpoint, and 15432 for PostgreSQL. Remote clients connect through mcp-remote at https://<YOUR_CERT_SERVER_IP>:8444/mcp/sse.
Setup effort
Build from source — git clone https://github.com/beye91/nexus-dashboard-mcp.git
USE CASES
01
Getting a picture of a fabric you inherited
List fabrics, then walk the switches, interfaces, networks and VRFs inside the one you care about.
02
Chasing down a configuration drift
Read the configuration diff for a switch and see what is out of sync before touching anything.
03
Reviewing anomalies and compliance
Pull fabric anomalies and the compliance rules behind them in one conversation.
04
Letting a team share one connection safely
Give each person a role, keep writes behind edit mode, and read the audit log when something changes.
66 tools
Listed in project docs. Your client may show a different list.
manage_createFabricmanage_getFabricDetailsmanage_updatefabricDetailsmanage_deleteFabricmanage_listFabricSwitchesmanage_createFabricSwitchesmanage_getFabricSwitchesSummarymanage_getFabricSwitchmanage_deleteFabricSwitchmanage_getSwitchInterfacesOverviewmanage_getSwitchConfigurationDiffmanage_executeRediscoverSwitchesmanage_executeReloadSwitchesmanage_deleteFabricSwitchesmanage_executeChangeCredentialSwitchesmanage_executeChangeSwitchIpCollectionmanage_executeMoveNeighborSwitchmanage_executeSwitchConfigurationPreviewmanage_listInterfacesmanage_createInterfacesmanage_getInterfacesSummarymanage_getInterfacemanage_replaceInterfacemanage_deleteInterfacemanage_listNetworksmanage_createNetworksmanage_getNetworksSummarymanage_getNetworkmanage_replaceNetworkmanage_deleteNetworkmanage_listVrfsmanage_createVrfsmanage_getVrfsSummarymanage_getVrfmanage_replaceVrfmanage_deleteVrfmanage_listComplianceAnomalyRulesmanage_createComplianceAnomalyRulemanage_getComplianceAnomalyRulemanage_listAnomaliesmanage_listTemplatesmanage_createTemplatesmanage_getTemplatemanage_replaceTemplatemanage_deleteTemplatemanage_listPoliciesmanage_createPoliciesmanage_getPolicymanage_replacePolicymanage_deletePolicymanage_listAccessControlListsmanage_createAccessControlListsmanage_getAccessControlListsSummarymanage_listSwitchCredentialsmanage_createSwitchCredentialsmanage_getDefaultSwitchCredentialsmanage_createDefaultSwitchCredentialsmanage_executeDeployConfigurationmanage_executeSaveAndDeployConfigurationmanage_getFabricDeploymentFreezeModemanage_executeRedeployFabricsmanage_getSmartSwitchDpusmanage_getSmartSwitchDpusSummarymanage_executeSmartSwitchOnboardingmanage_executeSmartSwitchDeboardingBEFORE YOU INSTALL
What to know first
Tools you get
66 tools named in the project's documentation.
Requires
- Docker 20.10+ with Docker Compose 2.0+
- Node.js 18+
- Cisco Nexus Dashboard 4.1+ with NDFC 12.x, and an account on it
INSTALLATION
Connect your client
This one is built from source rather than installed with a single command.
Server setup
TERMINAL
git clone https://github.com/beye91/nexus-dashboard-mcp.git
cd nexus-dashboard-mcp
echo "CERT_SERVER_IP=<YOUR_CERT_SERVER_IP>" > .env
docker compose up -d --buildOpen https://<YOUR_CERT_SERVER_IP>:7443, accept the self-signed certificate warning, complete the admin setup, then on the Clusters page add your Nexus Dashboard and click Test Connection.
MCP Clients
Claude Desktop
CLAUDE DESKTOP · JSON
{
"mcpServers": {
"nexus-dashboard": {
"command": "npx",
"args": [
"mcp-remote@latest",
"https://<YOUR_CERT_SERVER_IP>:8444/mcp/sse",
"--transport",
"sse-only"
]
}
}
}<YOUR_CERT_SERVER_IP> is your server's IP address. The server uses a self-signed certificate, so mcp-remote rejects it unless NODE_TLS_REJECT_UNAUTHORIZED=0 is set in the environment Claude Desktop starts from (export it, then launch Claude from that shell) or the certificate is added to your system trust store.
Your Nexus Dashboard username and password are entered in the web UI when you add the cluster; NEXUS_USERNAME and NEXUS_PASSWORD in .env are an optional way to preset them. The client block itself needs no credentials.
