authoritative-data-harvester

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

• Third-party content exposure detected (high risk: 0.85). Runtime path scripts/run.py loads data_requirements.json and for type: scrape_tables uses pd.read_html(url) to fetch and parse an arbitrary outsider-provided URL (e.g., the bundled example points to Wikipedia), converting scraped HTML tables into LLM-readable CSV text in crawled_data/processed/.