The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/extract_audio.py executes the yt-dlp and ffmpeg command-line tools using subprocess.run with argument lists. This facilitates media downloading and processing as a core feature of the skill.\n- [COMMAND_EXECUTION]: The skill exposes the yt-dlp feature to access browser cookies for authentication on various platforms. This involves accessing sensitive local data but is presented as a legitimate tool feature for the user.\n- [PROMPT_INJECTION]:\n
Ingestion points: The skill processes external media content and metadata from URLs passed to the extract_audio.py script (e.g., in the url argument).\n
Boundary markers: No explicit markers or instructions are used to distinguish between untrusted data and system instructions during processing.\n
Capability inventory: The skill possesses the capability to execute shell commands (yt-dlp, ffmpeg) and write files to the local system.\n
Sanitization: The script uses secure subprocess calling conventions (argument lists) to prevent shell injection, though it lacks sanitization for metadata strings retrieved from external sources.

audio-extractor