databricks-ai-functions

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

• Third-party content exposure detected (high risk: 0.85). Runtime LLM context can include outsider-authored free text when ai_query is called with request built from text_blocks/document_text derived from ai_parse_document over files loaded from a Unity Catalog volume (read_files(...) / spark.read.format("binaryFile").load(...)), where the documents’ contents are not authored by the operating user.