data-journalism

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

• Third-party content exposure detected (high risk: 0.80). The skill's documentation and code examples explicitly show fetching and ingesting open/public third-party content (e.g., "scraping public information" under "Building your own dataset", gpd.read_file('https://example.com/data.geojson'), Google Maps geocoding and Census batch geocoder examples, and references to public portals like Data.gov), which are untrusted/user-generated sources the agent would be expected to read and interpret as part of the workflow.