azure-ai-document-intelligence-ts

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [Indirect Prompt Injection Surface]: The skill is designed to process external document data from both local files and remote URLs, which is a standard feature for document analysis tools.
  • Ingestion points: Untrusted data enters the context via the urlSource parameter in REST calls and the readFile operation for local documents in SKILL.md.
  • Boundary markers: The skill does not currently implement specific boundary markers or instructions to ignore embedded content within the analyzed documents.
  • Capability inventory: The extracted data is logged to the console; the skill does not show patterns of executing system commands or performing further network operations based on the document's content.
  • Sanitization: Standard SDK processing is used, though additional validation of extracted text is recommended before using it in further automated logic.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 01:19 PM
Security Audit — agent-trust-hub — azure-ai-document-intelligence-ts