dxpert.ai — UNS MCP server: Sparkplug B lint, no key
Lints Sparkplug B topics and Unified Namespace names from inside your agent. No broker connection, no API key.
claude mcp add --env DXPERT_API_BASE=<DXPERT_API_BASE> --transport stdio dxpertai-free-uns-tools-no-api-key -- npx -y @dxpert/uns-toolsWhat dxpert.ai — free UNS tools (no API key) does
Run it through npx; it speaks JSON-RPC over stdio and ships without dependencies. Two checks stay on your machine: lint_sparkplug_topic matches one MQTT topic string or a whole batch against the Sparkplug 3.0 topic grammar, while check_uns_namespace hunts for uneven hierarchy depth, mixed letter case, spaces inside a segment, siblings that look like one node but are two, repeated paths and stray wildcards. Input can be a single string, newline-separated lines or an array.
Both tools judge text only. Nothing here reaches a broker, decodes a payload or confirms that a device exists. Naming gets a consistency pass, not a namespace design, and your plant stays unknown to the server. The lint logic is copied from separate npm packages, and the test suite checks that verdicts match those originals.
run_readiness_diagnostic is the one tool that goes online. It takes a 16-question self-assessment, then returns scores on ten axes from 0 to 5 plus a maturity stage and the foundation gaps holding back the stated ambition. Treat the output as a preliminary screening, not an audit — that framing comes back in the response and the server expects you to repeat it. The endpoint is rate limited, so it fails without connectivity.
What it helps with
- catch a malformed Sparkplug B topic before a broker does
- spot UNS siblings that differ only by letter case
- enforce one naming style across a topic set
- get a rough read on industrial AI readiness
Example prompts for the agent
- Run
lint_sparkplug_topiconspBv1.0/Plant1/DDATA/Line1/PLC1and show the findings - Check the topic paths in
topics.txtwithcheck_uns_namespaceand list every issue - Take my intake answers through
run_readiness_diagnosticand explain the weak axes
How to install dxpert.ai — free UNS tools (no API key)
Platform CLI
claude mcp add --env DXPERT_API_BASE=<DXPERT_API_BASE> --transport stdio dxpertai-free-uns-tools-no-api-key -- npx -y @dxpert/uns-toolsReplace <…> with your own values
Platform CLI
codex mcp add dxpertai-free-uns-tools-no-api-key --env DXPERT_API_BASE=<DXPERT_API_BASE> -- npx -y @dxpert/uns-toolsReplace <…> with your own values
Config file
{
"mcpServers": {
"dxpertai-free-uns-tools-no-api-key": {
"command": "npx",
"args": [
"-y",
"@dxpert/uns-tools"
],
"env": {
"DXPERT_API_BASE": "<DXPERT_API_BASE>"
}
}
}
}Replace <…> with your own values
Config file
{
"servers": {
"dxpertai-free-uns-tools-no-api-key": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@dxpert/uns-tools"
],
"env": {
"DXPERT_API_BASE": "<DXPERT_API_BASE>"
}
}
}
}Replace <…> with your own values
Platform CLI
gemini mcp add -e DXPERT_API_BASE=<DXPERT_API_BASE> dxpertai-free-uns-tools-no-api-key npx -y @dxpert/uns-toolsReplace <…> with your own values
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about dxpert.ai — free UNS tools (no API key)
How do I install dxpert.ai — free UNS tools (no API key) in Claude Code?
Add it with claude mcp add dxpert-uns-tools -- npx -y @dxpert/uns-tools. Node 18+ is enough; there is no separate install step.
What does dxpert.ai — free UNS tools (no API key) do?
It validates Sparkplug B topic strings, flags inconsistent naming in a Unified Namespace and can run the public industrial AI-readiness diagnostic. None of it asks for credentials.
Does it need a network connection?
The two validators run offline. Only run_readiness_diagnostic makes a network call, and that one is rate limited.
