NVIDIAOfficial·36 Agent Skills Included

NemoClaw

Run always-on AI agents safely inside hardened OpenShell sandboxes

Runs always-on AI agents like OpenClaw, Hermes, and Deep Agents inside hardened NVIDIA OpenShell sandboxes with guided onboarding and managed inference. Removes the risk of exposed credentials, open network access, and uncontrolled agent behavior through network policies, SSRF validation, and sandbox isolation. Ships agent skills that automate contributor setup, issue planning, PR publication, security review, triage, and release management.

22,335|3,061

↗

npx skills add NVIDIA/NemoClaw --all -g -y

Available:

Tells the AI agent how the repository is structured, which packaged skill owns each user, contributor, or maintainer task, and what security, testing, and product-scope rules to enforce while working.

All Skills in This Repository (36)

Pure Emerald Level Indicators

📦 In Repo

NVIDIA

nemoclaw-user-get-started

Install NemoClaw and create a sandboxed OpenClaw assistant for local testing.

nemoclaw-contributor-create-pr

Automates GitHub pull request creation with enforced prerequisites and NemoClaw's PR template via gh CLI.

nemoclaw-maintainer-find-review-pr

Find open GitHub PRs labeled security and priority: high and detect duplicates.

nemoclaw-user-reference

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

nemoclaw-maintainer-security-code-review

Review GitHub pull requests for security flaws and generate PASS/WARNING/FAIL reports.

nemoclaw-user-triage-instructions

Assign 1–5 labels and generate tiered triage comments for GitHub issues and PRs.

nemoclaw-user-configure-security

Document NemoClaw security controls and credential storage risks.

nemoclaw-user-skills-coding

Expose NemoClaw documentation as discoverable skills for AI coding assistants.

nemoclaw-maintainer-evening

Check release progress, bump straggling issues, and prepare annotated Git tags.

nemoclaw-user-overview

Explain NemoClaw's integration of OpenClaw and OpenShell for sandbox creation and inference routing.

nemoclaw-user-deploy-remote

Deploys NemoClaw to remote GPU hosts with sandboxing and credential rotation.

nemoclaw-maintainer-morning

Triage the NemoClaw backlog, label targeted items, and output a daily plan.

Official

Intermediate

Frequently Asked Questions

FAQPage Schema

How to install NemoClaw?▼

Run `npx skills add NVIDIA/NemoClaw --all -g -y` in your terminal to install the full skill suite globally.

What does NemoClaw do?▼

It runs always-on AI agents such as OpenClaw, Hermes, and LangChain Deep Agents inside hardened NVIDIA OpenShell sandboxes with network policies, credential protection, and managed inference.

How do NemoClaw agent skills help contributors?▼

Dedicated skills walk your AI agent through checkout setup, issue planning, implementation with tests, and publishing a compliant pull request with DCO sign-off and verified commits.

Can maintainers automate triage and releases with NemoClaw skills?▼

Yes. Maintainer skills cover issue triage, PR comparison, security review, E2E failure repair, release tagging, and drafting release announcements.

Does NemoClaw work with Claude Code and other AI agents?▼

Yes. The skills follow the standard SKILL.md format and work in Claude Code, OpenClaw, Cursor, Codex, and other compatible agents.