sensight

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

• Third-party content exposure detected (high risk: 0.90). The skill directly fetches and ingests user-generated public content from social media and public sites (e.g., social_search / daily_social / search_author_posts in SKILL.md and scripts/sensight.py call APIs for Weibo/X/小红书/微信公众号, blogs, arXiv and return posts which are then read and summarized), so untrusted third‑party content is read and can materially influence agent outputs.