The Agent Skills Directory

[SAFE]: The skill is authored by a trusted organization and follows security best practices for AI agent tools. It explicitly instructs the agent to avoid exfiltrating secrets, redacting sensitive terms like 'key', 'token', and 'password' from its output. It also restricts command execution to repo-specific tools (uv run nemotron) and forbids broad environment exploration (e.g., printenv, export). All external resources, such as Hugging Face and NVIDIA's own repositories, are well-known or trusted services used for their intended purpose in the machine learning workflow.

nemotron-customize