The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/fetch_transcript.py uses subprocess.run to call the yt-dlp utility for metadata extraction. The call is implemented safely using a list of arguments, and the video ID is validated against a strict regular expression ([a-zA-Z0-9_-]{11}) to prevent command injection.
[COMMAND_EXECUTION]: The script utilizes the xml.etree.ElementTree library for parsing transcript XML. While this library is generally vulnerable to XML External Entity (XXE) attacks, the risk is mitigated in this context as the input is fetched directly from YouTube's official caption tracks over HTTPS.
[PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it processes untrusted transcript data from YouTube videos.
Ingestion points: Data enters the system via scripts/fetch_transcript.py or the WebFetch and WebSearch tools in fallback tiers.
Boundary markers: The skill uses a structured Markdown template (assets/digest_template.md) to isolate the analysis, which helps define boundaries but does not explicitly instruct the agent to ignore commands within the transcript text.
Capability inventory: The skill has access to Bash, Write, WebFetch, and WebSearch tools.
Sanitization: There is no evidence of explicit sanitization or filtering of the transcript text before it is processed by the agent.

youtube-digest