The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses subprocess.run in scripts/fetch_transcript.py to execute the yt-dlp binary for fetching video metadata and as a fallback for transcript extraction.
Evidence: subprocess.run(cmd, ...) is called with a list of arguments, which is a secure way to execute external commands compared to shell strings.
[EXTERNAL_DOWNLOADS]: The skill interacts with YouTube's public interface to retrieve transcripts and metadata.
Evidence: It uses youtube-transcript-api and yt-dlp to communicate with youtube.com domains.
[PROMPT_INJECTION]: The skill is subject to indirect prompt injection as it ingests untrusted transcript data from YouTube videos.
Ingestion points: scripts/fetch_transcript.py (via API or yt-dlp).
Boundary markers: None identified in the prompt logic; the transcript is processed as raw text for analysis.
Capability inventory: File writing (analyze_video.py), Subprocess execution (fetch_transcript.py).
Sanitization: Filenames are sanitized in scripts/utils.py:sanitize_filename.

youtube-analysis