web-search

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

• Third-party content exposure detected (high risk: 0.95). The required workflow uses belt app run tavily/search-assistant / tavily/extract (and/or exa/*) which fetches and extracts public web page text at runtime, and that outsider-authored content is then passed into the LLM via the <search-results> / <content> prompt.