youtube-search
Pass
Audited by Gen Agent Trust Hub on Mar 10, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill processes untrusted external data (YouTube transcripts, titles, and descriptions) which presents an indirect prompt injection surface. Ingestion points: Video metadata and transcripts are fetched in 'SKILL.md' (Methods D and E) and 'references/parsing-examples.md'. Boundary markers: The skill does not implement delimiters or safety instructions to prevent the agent from following commands embedded within the fetched external data. Capability inventory: The skill uses 'requests', 'yt-dlp', and 'youtube-transcript-api' to pull text content into the agent's context. Sanitization: No sanitization or validation of the retrieved text is performed before it is processed.
- [EXTERNAL_DOWNLOADS]: The skill utilizes well-known external libraries and APIs for its functionality. Fetches data from official Google ('googleapis.com') and SerpAPI ('serpapi.com') endpoints. Recommends the installation of standard Python packages: 'requests', 'youtube-transcript-api', and 'yt-dlp'.
Audit Metadata