The Agent Skills Directory

[COMMAND_EXECUTION]: The script executes external command-line utilities including ffmpeg, yt-dlp, whisper, and whisperx to perform media processing. These calls are implemented using list-based arguments in subprocess.run, which effectively mitigates standard shell injection vulnerabilities despite the complexity of the commands.
[EXTERNAL_DOWNLOADS]: The skill utilizes yt-dlp to download video files, thumbnails, and metadata from remote URLs provided by the user at runtime.
[PROMPT_INJECTION]: The skill is subject to indirect prompt injection risks due to its processing of external data.
Ingestion points: Media files and subtitles from external URLs or local paths are processed into text context (scripts/watch_video.py).
Boundary markers: None. The skill does not use specific delimiters or instructions to the agent to ignore potentially malicious text within the transcripts.
Capability inventory: The skill performs file system writes and executes several subprocesses for media manipulation (scripts/watch_video.py).
Sanitization: Filenames are slugified to prevent path traversal, but the textual content of the transcripts and video metadata presented to the agent is not sanitized for instructional content.

watch-video