skills/nvidia/skills/launch-nemo-rl/Gen Agent Trust Hub

launch-nemo-rl

Pass

Audited by Gen Agent Trust Hub on Jun 12, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill provides operational guidance for the nrl-k8s command-line tool and Kubernetes environment management. No malicious patterns such as prompt injection, obfuscation, or unauthorized data exfiltration were detected.
  • [COMMAND_EXECUTION]: The instructions guide the agent to use nrl-k8s and kubectl to interact with Kubernetes clusters. This behavior is consistent with the skill's stated purpose of managing training infrastructure.
  • [EXTERNAL_DOWNLOADS]: The skill refers to container images on nvcr.io (NVIDIA Container Registry) and documentation/source code on nvidia.com and github.com/NVIDIA-NeMo. These are legitimate vendor-owned resources.
  • [CREDENTIALS_UNSAFE]: The skill promotes secure secret management by instructing the agent to use Kubernetes secretKeyRef for API keys and image pull secrets, explicitly warning against embedding credentials directly in configuration files.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 12, 2026, 07:42 PM
Security Audit — agent-trust-hub — launch-nemo-rl