video-understanding
Pass
Audited by Gen Agent Trust Hub on Jul 6, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [EXTERNAL_DOWNLOADS]: The references file suggests installing 'yt-dlp' via pip. This is a well-known and reputable tool for video content extraction.
- [PROMPT_INJECTION]: The skill processes untrusted metadata from external video platforms, presenting a surface for indirect prompt injection. 1. Ingestion points: Fetches video titles, descriptions, and metadata from platforms like YouTube and Bilibili (SKILL.md, references/resources.md). 2. Boundary markers: There are no defined boundary markers to isolate the fetched external content from the agent's system instructions. 3. Capability inventory: The skill focuses on parsing and summarizing content which is then used in the agent's reasoning process. 4. Sanitization: The skill does not describe any specific sanitization or filtering of the external metadata before it is processed.
Audit Metadata