internal-linking

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

• Third-party content exposure detected (high risk: 0.85). The required workflow fetches external web pages at runtime—specifically https://<domain>/query-index.json and each <path>.plain.html (plus /nav.plain.html and /footer.plain.html)—and then ingests the fetched .plain.html body text (including <a> anchor text) into the agent’s LLM context for link extraction and recommendations, which is outsider-authored free text from the target site.