ingredient-research

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

• Third-party content exposure detected (high risk: 0.75). The skill’s runtime workflow performs per-ingredient “search for” and “capture” of clinical studies and other evidence, which typically requires fetching and ingesting outsider-authored free text from public web sources (e.g., articles/abstract pages) or third-party databases into the LLM context for summarization.