Agentic Skills

Agentic Skills are pre-built capabilities that an AI agent can call when built on the NVIDIA Build platform. Rather than requiring developers to implement every function from scratch, skills provide ready-made building blocks that can be composed into more complex agentic workflows. [NVIDIA Build Overview]

NVIDIA Build organises available skills into four domains: [NVIDIA Build Overview]

Skills are a core primitive in the agentic AI model promoted by NVIDIA Build. An agent — such as one secured by NemoClaw — can invoke individual skills as tools, keeping each capability modular, testable, and independently maintained. [NVIDIA Build Overview]

• NemoClaw — safe agent execution layer that controls access and protects data when agents use skills. [NVIDIA Build Overview]
• NVIDIA Build Inference Endpoints — free inference endpoints for leading models, which skills may invoke. [NVIDIA Build Overview]
• Blueprint Collection — workflows and code samples that demonstrate how to wire skills into full AI applications. [NVIDIA Build Overview]
• DGX Station — hardware platform with playbooks that include setting up agents that make use of skills. [NVIDIA Build Overview]