lerobot

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). The skill instructs the agent to download and load datasets from the Hugging Face Hub (such as lerobot/pusht or user-specified datasets via LeRobotDataset / StreamingLeRobotDataset) and allows loading external sim environments via EnvHub (make_env(..., trust_remote_code=True)), which ingest outsider-contributed text/code from public repositories without prior user selection of specific items.

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

• Attempt to modify system services in skill instructions detected (medium risk: 0.60). The skill prompt instructs the agent to install ffmpeg using sudo apt install ffmpeg on Linux, which asks the agent to use sudo privileges.