The Agent Skills Directory

[COMMAND_EXECUTION]: The script executes the yt-dlp command-line tool using subprocess.run with argument lists. This is a secure implementation that avoids shell interpolation and prevents command injection.
[DATA_EXPOSURE]: The skill ingests external content from YouTube (metadata and transcripts) and writes it to the local file system. It includes a markdown_text sanitization function to escape markdown control characters, reducing the risk of structural injection when the resulting files are viewed or processed.
[PROMPT_INJECTION]: As a data ingestion tool, the skill processes untrusted content from YouTube transcripts. While it implements basic sanitization for markdown formatting, the textual content of the transcripts could contain adversarial instructions intended to influence downstream AI agents that read the vault files. This is a characteristic of the tool's primary function rather than a specific vulnerability in the skill's code.

ingest-youtube