youtube-to-viral-posts

Pass

Audited by Gen Agent Trust Hub on Apr 8, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes external, untrusted content from YouTube videos to generate output. \n
  • Ingestion points: YouTube video data processed via the Gemini API in scripts/youtube_to_viral_posts.py. \n
  • Boundary markers: None. The instructions do not include markers to separate video content from the system prompt. \n
  • Capability inventory: The skill writes to a PostgreSQL database (store_in_database in scripts/youtube_to_viral_posts.py). \n
  • Sanitization: The skill does not validate or sanitize AI-generated content before storing it in the database. \n- [PROMPT_INJECTION]: The VIRAL_POST_PROMPT in scripts/youtube_to_viral_posts.py explicitly directs the AI to fabricate statistics (e.g., 'even if its made up'), facilitating the creation of deceptive content. \n- [SAFE]: The skill documentation in SKILL.md and references/youtube_to_viral_posts.md contains discrepancies regarding the use of yt-dlp and specific Google SDK versions compared to the implementation in scripts/youtube_to_viral_posts.py.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 8, 2026, 02:44 PM
Security Audit — agent-trust-hub — youtube-to-viral-posts