historical-thinking-strategy-modelling-guide

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

• Third-party content exposure detected (medium risk: 0.65). The required workflow ingests the teacher-provided Document description (including source note and key passage) into the LLM prompt; if that document text/passage is authored by an outsider (e.g., a web page, issue/PR text, or other non-user document), that outsider-authored free text becomes LLM-readable context via {{document_description}}.