wechat-article-writer

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

• Third-party content exposure detected (high risk: 1.00). The skill explicitly calls web_fetch to retrieve user-provided website content and exa:web_search_exa to search and ingest open web results (including news, reviews and other public/user-generated sites), and the agent is expected to read and integrate that content into the article, exposing it to untrusted third‑party input that could carry indirect prompt injection.