xiaohongshu

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

• Third-party content exposure detected (high risk: 1.00). The workflow's required Step 2 "竞品调研 + 入库" explicitly calls mcp__xiaohongshu-mcp__search_feeds and mcp__xiaohongshu-mcp__get_feed_detail to fetch user-generated Xiaohongshu posts (public social-media content) and then instructs the agent to read and use those posts' titles/content/analysis to drive design, tags, and generation decisions, which exposes it to untrusted third-party content that could carry indirect prompt injections.