NemoClaw

NemoClaw is an open-source agent orchestration harness and execution framework developed by NVIDIA. It provides structured runtimes, modular skill dispatching, and policy guardrails designed to safely run autonomous coding and system agents across local environments and enterprise compute infrastructure. At runtime, NemoClaw manages task execution across sandboxes, handles messaging integrations, and evaluates tool calls against strict security policies. It organizes agent capabilities into modular skills stored within the repository (such as git operations and system interactions) while routing inference and managing sensitive credential access across external services. Unlike standalone coding assistants, NemoClaw functions as an orchestration layer with built-in hardware profiling (including validation for NVIDIA DGX Station setups) and multi-layered policy checks. It prevents unauthorized credential leakage, enforces runner sandboxing, and validates environment preflights before executing agent actions. NemoClaw is open-source and hosted on GitHub, with development managed via Node.js/TypeScript toolchains and structured quality verification gates.

Tags: agent harness, AI, ai agent, cli, developer tools, OpenSource

Visit NemoClaw

Key features

• Sandboxed Agent Runner — Executes agent tool invocations and commands within isolated sandbox environments to safeguard host systems. - Modular Skill System — Organizes executable agent abilities in structured directories (e.g., .agents/skills) for extensible automation. - Policy & Credential Routing — Enforces runtime security rules and disambiguates credential routing for third-party tools and messaging channels. - Hardware & Environment Profiling — Features targeted verification and testing profiles for specialized hardware such as DGX Station environments. - Preflight & Quality Gates — Integrates pre-commit hooks, linting checks, and automated validation scripts (npm run check, npm run validate:pr) to ensure reliable operation. - Multi-Interface Messaging — Routes agent input and output across communication platforms (including chat integrations like Slack) and CLI runners.

Use cases

Use Case: Sandboxed Task Automation Run autonomous agents to perform file modifications, execute shell commands, and run tests within a restricted sandbox that isolates sensitive system paths and secrets.

Use Case: Policy-Governed Git Operations Deploy agent skills to manage Git branches, handle remote protocol fallbacks, and stage changes while enforcing organizational signing and branch access policies.

Use Case: Specialized Hardware Orchestration Run agentic workflows validated against enterprise hardware configurations (such as DGX workstations) with explicit hardware profiling and preflight checks.

Who it is for

NemoClaw is designed for AI systems engineers, DevOps teams, and enterprise developers building governed agent workflows requiring sandboxing, explicit policy checks, and DGX hardware compatibility. It is not intended for individual developers looking for a simple, zero-config terminal coding assistant.

Install guide

Installation Clone the repository and install dependencies using Node.js: bash git clone https://github.com/NVIDIA/NemoClaw.git cd NemoClaw npm install

First Run Run repository preflight checks and validate your environment: bash # Verify your setup and test harness npm test # Run repository-wide validation npm run check Configure required environment variables and API credentials in your environment configuration file before launching the runner.

Example Session Execute agent validation gates and dispatch tasks through the orchestrator harness: bash # Run PR validation checks on the current agent workspace npm run validate:pr

Part of CLI Coding Agents

Categories: AI Coding Agents, Chatbot Development, Developer Tools

Related CLI agents

• ActPlane — ActPlane is an open-source agent execution and policy enforcement harness developed by the Eunomia-bpf organization. It functions as a security…
• hol-guard — hol-guard is a security scanner and agent guardrail harness developed by Hashgraph Online. Built to safeguard agentic runtime environments and…
• Skillreaper — Skillreaper is an open-source skill lifecycle management and auditing tool for AI agent ecosystems, developed by thousandflowers. It analyzes agent…
• Agent Island — Agent Island is an open-source agent orchestration harness created by developer tristan666666 on GitHub. Designed as a framework for running…
• Stratless — Stratless is an open-source agent harness and execution orchestrator developed by the Stratless AI team. It provides a modular infrastructure…
• Google AX — Google AX is an open-source agent harness and execution orchestrator developed by Google. It provides runtime infrastructure and execution sandboxing…