youtube-to-heygen-longform

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

• Third-party content exposure detected (high risk: 1.00). The skill explicitly sends arbitrary YouTube URLs to Gemini for full-content analysis (see SKILL.md and scripts/youtube_to_heygen_longform.py's extract_script_from_youtube), meaning it ingests untrusted, user-generated third-party video content which the agent reads and uses to generate scripts that directly drive downstream actions (HeyGen video creation).

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

• Potentially malicious external URL detected (high risk: 0.90). The skill at runtime passes arbitrary YouTube URLs (e.g., "https://www.youtube.com/watch?v=VIDEO_ID") into Gemini via file_uri to extract video content which directly becomes the generated script/instructions sent to HeyGen, so the external YouTube URL content controls the agent's prompts and is a required runtime dependency.