The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/video_to_observation_pack.py uses the subprocess.run module to execute ffmpeg and ffprobe commands. These tools are used to extract keyframes, metadata, and audio from video files provided by the user. The script follows security best practices by passing arguments as lists rather than strings, which mitigates shell injection risks, and the functionality is central to the skill's purpose of behavioral observation.
[SAFE]: The skill processes user-provided video files and optional transcripts. While this represents a surface for indirect prompt injection (where instructions could be hidden in metadata or transcripts), the skill provides structured analysis prompts for the agent to ensure it focuses on factual observation within the Montessori framework.
[SAFE]: The skill references external URLs for the developer's repository and social media contact information. These are presented in plain text and are consistent with the vendor's identity ('beijita-lab').

eeailab-montessori