The Agent Skills Directory

[SAFE]: The skill utilizes reputable open-source libraries, specifically yt-dlp and youtube-transcript-api, for all core operations. It also correctly uses public RSS feeds for monitoring channel updates.
[SAFE]: Security best practices for secret management are enforced; the instructions explicitly direct the agent to use environment variables for third-party API keys and forbid requesting secrets from the user in the chat interface.
[INDIRECT_PROMPT_INJECTION]: The skill interacts with external content from YouTube, which represents a potential injection surface.
Ingestion points: User-supplied URLs and search queries are processed by CLI tools as described in SKILL.md.
Boundary markers: The shell command templates do not explicitly specify sanitization or rigorous quoting for the input variables.
Capability inventory: The skill uses subprocess execution of yt-dlp and youtube_transcript_api for data retrieval.
Sanitization: No specific input sanitization or output escaping logic is included in the instructions.

youtube-full