azure-ai-document-intelligence-dotnet

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md core workflows show the agent calling AnalyzeDocumentAsync on arbitrary document URLs (e.g., fileUri/documentUri like https://example.com/document.pdf) and using BlobContentSource with a blob container SAS URL to ingest external documents, so it clearly fetches and analyzes untrusted third-party content that can influence actions.