tabular-optimization-ingestion

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

• Third-party content exposure detected (medium risk: 0.65). Outsider free text can enter the LLM context via the runtime ingestion of user-uploaded tables (e.g., CSV/XLSX/JSON files) whose cell contents are arbitrary text authored by someone other than the operating user; the workflow explicitly inspects “uploaded data” and infers roles from column names/values.