A versatile framework for building, evaluating, and deploying enterprise-grade AI agents with seamless tool and data integration.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install nat-skill
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install nat-skill using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is NVIDIA NeMo Agent Toolkit (NAT)?
The NVIDIA NeMo Agent Toolkit (NAT) is a robust library designed to bridge the gap between large language models and real-world utility. By leveraging Openclaw Skills like NAT, developers can connect agents to diverse data sources and custom tools across various frameworks, including LangChain, LlamaIndex, and CrewAI. It provides a standardized, YAML-driven approach to defining agent behaviors, making it easier to manage complex reasoning patterns and tool-calling logic in professional environments.
NVIDIA NeMo Agent Toolkit (NAT) Use Cases
- Building ReAct or tool-calling agents for automated research and data analysis
- Deploying AI workflows as A2A or MCP servers for multi-agent interoperability
- Evaluating agent performance and profiling resource usage in production-grade pipelines
- Integrating NVIDIA NIM LLMs with custom enterprise toolsets and private data sources
How NVIDIA NeMo Agent Toolkit (NAT) Works
- Initialize the environment by installing the nvidia-nat package and setting the required NVIDIA_API_KEY.
- Define a workflow.yml configuration file to specify the LLM providers, tools (functions), and the agent's structural logic.
- Execute the agent through the CLI using the nat run command to process inputs and trigger tools.
- Optimize the agent by using the built-in profiling, evaluation, and red-teaming tools available within Openclaw Skills.
NVIDIA NeMo Agent Toolkit (NAT) Setup
To get started with NAT using Openclaw Skills, you can install the core toolkit or specific extras based on your framework requirements:
# Core installation
uv pip install nvidia-nat
# Installation with framework extras
uv pip install "nvidia-nat[langchain,mcp,a2a]"
# Configure the primary environment variable
export NVIDIA_API_KEY=<your_api_key_here>
NVIDIA NeMo Agent Toolkit (NAT) Data Schema & Taxonomy
NAT utilizes a highly structured YAML configuration system to manage its metadata and operational logic. The schema is organized into several key categories within Openclaw Skills:
| Component | Description |
|---|---|
| functions | Defines accessible tools such as search engines, calculators, and custom API integrations. |
| llms | Configures the model providers, including model names, temperatures, and endpoint settings. |
| embedders | Specifies embedding models used for vector-based search and retrieval tasks. |
| workflow | Orchestrates the agent type (e.g., react_agent, reasoning_agent) and wires tools to LLMs. |
NVIDIA NeMo Agent Toolkit (NAT) Advanced Features
- Agent-to-Agent (A2A) publishing to allow workflows to be discovered and called by other AI clients
- Model Context Protocol (MCP) server support for standardized tool and resource sharing
- Built-in red-teaming and evaluation suites to ensure agent safety and response quality
- Automatic memory wrappers to provide persistent conversational context across sessions
- Custom function groups for advanced namespacing and shared configuration across multiple agents
SKILL.md
Loading
Related Openclaw Skills
Alibaba Cloud WAF Quick Showcase
A specialized skill for automating the deployment of Alibaba Cloud Web Application Firewall (WAF) to protect web applications running on ECS instances.
Rumour Buster
A dual-engine fact-checking skill that cross-verifies information across Chinese and English search ecosystems to trace sources and dismantle misinformation.
Link-All
An intelligent integration manager that researches, connects, and transforms external platforms into reusable skills.
Message Summary
An intelligent AI skill designed to extract key information, decisions, and action items from Lark conversations into structured summaries.
Sedentary Reminder
An AI skill for designing intelligent sedentary reminder systems that prioritize helpful timing and context-awareness over mechanical repetition.
PySpark to Excel
Effortlessly converts ASCII-style PySpark table outputs into clean Tab-separated text for direct pasting into spreadsheets.
Featured*
