video-to-subtitle-summary

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly fetches and downloads user-generated videos from public short-video platforms (see SKILL.md Step 1 using the TikHub API to get Douyin/Xiaohongshu/Bilibili info and Step 2 downloading the videos), then performs speech-to-text and has Claude read and summarize that untrusted content (Steps 4–7), so third-party content is ingested and can directly influence the agent's outputs and actions.