ai-wechat-hotspot-writer

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

• Third-party content exposure detected (high risk: 0.90). The skill's required workflow (SKILL.md "Gather Hotspots" and references/source-playbook.md) explicitly instructs the agent to collect and read content from open/public third‑party sources — e.g., X/Twitter, Weibo, Reddit, Hacker News, Product Hunt, GitHub Trending — and to browse or use social search when sensight is unavailable, so untrusted/user-generated content is ingested and directly used to score, select, and write items (risking indirect prompt injection).