best-practices

Pass

Audited by Gen Agent Trust Hub on Mar 11, 2026

Risk Level: SAFE
Full Analysis
  • Context Provisioning Patterns: The skill outlines methods for loading user attributes (like country and department) and organizational glossaries into an agent's runtime context. This is achieved using the Microsoft 365 Users connector and SharePoint search sources.
  • Data Integration and Ingestion: It demonstrates how to fetch data from trusted enterprise services. The patterns utilize the agent's built-in capabilities to retrieve information from authenticated sources, ensuring data is accessed according to the user's permissions.
  • Dynamic Instruction Management: The templates show how to incorporate retrieved variables into the agent's instructions. It provides guidance on managing token consumption and handling fallback scenarios when data (such as a user's country) is unavailable.
  • Security Best Practices: The documentation correctly identifies that specific connectors require user authentication (OAuth) to function, ensuring that sensitive profile data is only accessed for authenticated sessions.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 11, 2026, 02:39 AM