video-download
Pass
Audited by Gen Agent Trust Hub on Nov 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external websites through yt-dlp. This data, such as video descriptions or metadata, could contain instructions designed to influence the agent's behavior.
- Ingestion points: External video URLs provided to yt-dlp in SKILL.md (e.g.,
yt-dlp "URL" --dump-json). - Boundary markers: No explicit markers are used to separate untrusted web content from agent instructions.
- Capability inventory: The skill enables subprocess execution of yt-dlp and ffmpeg, and allows writing files to the local disk.
- Sanitization: No sanitization or validation of the content downloaded from external sites is performed before the agent processes the metadata.
- [EXTERNAL_DOWNLOADS]: The skill references installing the yt-dlp Python package and the ffmpeg system utility from standard repositories, and includes instructions for updating the tool using its built-in update command.
- [COMMAND_EXECUTION]: The skill provides numerous commands for executing the yt-dlp CLI tool to download media and metadata.
Audit Metadata