sglang-setup

Pass

Audited by Gen Agent Trust Hub on Jul 3, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses nvidia-smi to detect GPU hardware and docker to deploy and manage the inference server container.
  • [EXTERNAL_DOWNLOADS]: Pulls the lmsysorg/sglang Docker image from Docker Hub, which is the official repository for the SGLang project.
  • [DATA_EXFILTRATION]: The skill requests a HuggingFace token (HF_TOKEN) via a placeholder for authenticated model access, which is a standard and secure way to handle user credentials without hardcoding them.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 3, 2026, 11:06 AM
Security Audit — agent-trust-hub — sglang-setup