nemoclaw-user-reference

Document NemoClaw CLI commands, blueprint architecture, network policies, and troubleshooting procedures.

22.1k|3.0k|Updated Mar 15, 2026
One-click install
npx skills add https://github.com/NVIDIA/NemoClaw --skill nemoclaw-user-reference
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: nemoclaw-user-reference
Source: https://github.com/NVIDIA/NemoClaw/tree/main/.agents/skills/nemoclaw-user-reference
Command: npx skills add https://github.com/NVIDIA/NemoClaw --skill nemoclaw-user-reference

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Provides a single, discoverable reference that explains how NemoClaw combines a CLI plugin and a versioned blueprint to run OpenClaw inside a controlled OpenShell sandbox, and helps operators find command syntax, policy definitions, and resolution steps when onboarding or running sandboxes.

Core Features & Use Cases

  • Architecture Reference: Describes the plugin/blueprint split, sandbox layout, inference routing, and host-side state locations for operators and integrators.
  • CLI & Command Reference: Comprehensive listing of slash commands and standalone host commands with flags and behaviors (onboard, connect, status, logs, rebuild, policy management, skill install).
  • Network & Security Policies: Documents baseline filesystem rules, deny-by-default network policy, policy tiers and presets, and the operator approval flow for dynamic egress.
  • Troubleshooting Guide: Step-by-step resolutions for common failures (Docker, ports, cgroups, OpenShell version mismatches, messaging token conflicts, OOM during image push, local inference issues).
  • Use Case: Use this reference when preparing an onboarding run, auditing sandbox policy presets, or diagnosing why a sandbox cannot reach an inference backend.

Quick Start

Use the nemoclaw-user-reference to look up NemoClaw CLI commands, blueprint details, policy presets, and troubleshooting steps when onboarding or debugging a sandbox.

Frequently Asked Questions about nemoclaw-user-reference

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I troubleshoot NemoClaw sandbox onboarding failures?

NemoClaw network policies use a deny-by-default approach with baseline filesystem rules and tiered presets. Operators must follow a specific approval flow to authorize any dynamic egress, ensuring the OpenClaw sandbox remains secure.

How does the NemoClaw CLI plugin and blueprint architecture work?

The NemoClaw architecture combines a CLI plugin and a versioned blueprint to run OpenClaw inside a controlled OpenShell sandbox. The reference details the sandbox layout, inference routing, and host-side state locations for operators and integrators.

What CLI commands are available for managing OpenClaw sandboxes?

Available CLI commands for managing OpenClaw sandboxes include onboard, connect, status, logs, rebuild, policy management, and skill install. The reference provides a comprehensive listing of these slash commands and standalone host commands with their flags and behaviors.

Why does my OpenShell sandbox fail during image push or runtime?

Your OpenShell sandbox may fail due to OOM errors during image push, messaging token conflicts, or local inference issues. The troubleshooting reference provides targeted diagnostic steps and resolutions for these specific runtime failures.

Can I use NemoClaw to configure baseline filesystem and network rules?

Yes, you can use NemoClaw to configure baseline filesystem and network rules. The reference documents the baseline filesystem rules, deny-by-default network policies, and policy tiers needed to properly secure and configure your OpenClaw environment.