video-expert-analyzer

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly downloads and ingests content from public third‑party sites (e.g., SKILL.md and pipeline_enhanced.py show commands to fetch videos from Bilibili/YouTube/Douyin/Xiaohongshu, scripts/download_douyin.py and fetch_bilibili_subtitle.py scrape page HTML/APIs), and ai_analyzer.py (Agent mode / API mode) requires the agent to read frames/subtitles and uses those inputs to make scoring/selection decisions, so untrusted user‑generated web content can materially influence tool behavior.