sn-search-social-media

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

• Third-party content exposure detected (high risk: 0.75). Yes—this skill’s required workflow includes using browser-use to open public web pages (e.g., Weibo hotlists/Baidu index pages) and then using the agent/LLM to read that fetched page text, which is outsider-authored free text that can contain prompt-injection content.