youtube-to-viral-posts

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.90). The skill explicitly ingests public, user-generated YouTube content (SKILL.md "Downloads YouTube video via yt-dlp" and scripts/youtube_to_viral_posts.py which passes the YouTube URL as file_uri to Gemini for analysis), so the agent reads and acts on untrusted third‑party content that could carry indirect prompt-injection instructions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill sends user-supplied YouTube URLs (e.g., "https://www.youtube.com/watch?v=VIDEO_ID" or "https://youtu.be/VIDEO_ID") to Gemini at runtime via file_uri, thereby injecting remote video content directly into the model context which controls generation and is required for the skill to operate.