| name | roboneuron |
| description | Control RoboNeuron through its project-scoped MCP servers via mcporter. |
| metadata | {"openclaw":{"emoji":"🤖","os":["linux"],"requires":{"bins":["mcporter"]}}} |
Use this skill when the task involves controlling RoboNeuron through its MCP servers from OpenClaw.
Scope:
- Robot perception/control/VLA bring-up through RoboNeuron MCP.
- Inspecting available RoboNeuron tools with
mcporter.
- Starting or stopping the RoboNeuron MCP daemon/runtime paths used by OpenClaw.
Required command form:
- Always invoke
mcporter with --config /home/guanweifan/RoboNeuron/configs/openclaw/mcporter.json.
- Never rely on editor-imported MCP configs for RoboNeuron.
Startup order:
- For camera-only workflows, use
roboneuron-perception.
- For motion/control workflows, start
roboneuron-control before sending end-effector commands.
- For instruction-conditioned policy execution, start
roboneuron-perception, then roboneuron-control, then roboneuron-vla.
- Use
roboneuron-twist only for /cmd_vel style base motion publishing.
- Use
roboneuron-eef-delta for direct end-effector delta commands.
Useful commands:
mcporter --config /home/guanweifan/RoboNeuron/configs/openclaw/mcporter.json list
mcporter --config /home/guanweifan/RoboNeuron/configs/openclaw/mcporter.json list roboneuron-control --schema
mcporter daemon status
mcporter daemon restart --log
Rules:
- Treat RoboNeuron MCP servers as stateful and persistent. Do not replace them with ad-hoc
--stdio calls unless explicitly debugging.
- Reuse configured server names (
roboneuron-perception, roboneuron-vla, roboneuron-control, roboneuron-twist, roboneuron-eef-delta).
- If a tool call fails after a crash or stale connection, restart the daemon with
mcporter daemon restart.
References:
references/startup-order.md
references/troubleshooting.md