multi-video-summarizer

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

• Third-party content exposure detected (high risk: 0.90). The skill (SKILL.md and video_subtitle.py) automatically fetches and parses public, user-generated content from video platforms (YouTube, Bilibili, Douyin, Xiaohongshu and "any URL supported by yt-dlp"), including scraping page HTML (e.g., extracting window.SETUP_SERVER_STATE in Xiaohongshu) and ingesting subtitles/transcripts which the agent then reads and uses to generate summaries, so untrusted third‑party content can directly influence agent behavior.