Nemoclaw User Deploy Remote

This handles the deployment of NemoClaw to remote GPU instances through Brev, though the old `nemoclaw deploy` command is deprecated now. The current workflow is straightforward: provision a VM, install NemoClaw on the host, then run `nemoclaw onboard`. You'll need the Brev CLI authenticated, provider credentials for your inference backend, and HF_TOKEN set if you're working with gated models. One thing to watch for is setting CHAT_UI_URL in your Brev launchable environment config before the sandbox image builds. It's a utility skill that saves you from manually configuring GPU instances, which matters if you're spinning these up regularly for inference work.

Install to Claude Code

npx -y skills add nvidia/skills --skill nemoclaw-user-deploy-remote --agent claude-code