profile

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

• Third-party content exposure detected (high risk: 0.70). The required workflow reads user-provided files from input/<dataset>/ (via pd.read_csv/pd.read_excel/pd.read_json) and then prints their contents/statistics into data-profile.md, so any outsider-authored dataset files become readable free text/values inside the LLM context during summarization.