jetson-llm-serve

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill launches unpinned container images at runtime (e.g., vllm/vllm-openai:latest, ghcr.io/nvidia-ai-iot/vllm:latest-jetson-orin, nvcr.io/nvidia/sglang:26.01-py3) via docker run, which fetches and executes remote code from those URLs/registries and are required dependencies for the skill.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 0.80). The skill explicitly instructs running privileged system-change commands (e.g., "sudo nvpmodel -m 0 && sudo jetson_clocks" and sudo lsof checks) and host-level Docker operations that modify the machine's runtime state, so it pushes the agent to perform privileged system modifications.