query-tableau-data

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

• Third-party content exposure detected (high risk: 0.75). The required runtime workflow ingests Tableau site metadata via authenticated HTTP calls (REST inventory and GraphQL/VDS introspection), which is outsider-authored free text from other users/workbook authors (e.g., datasource/workbook/view names/descriptions and field captions/formulas) returned by Tableau APIs and then placed into the agent’s LLM context via printed summaries and structured objects.