azure-ai-projects-dotnet
Pass
Audited by Gen Agent Trust Hub on May 6, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- Installation of Official Packages: The skill includes commands to install standard libraries from a well-known package registry, such as
Azure.AI.ProjectsandAzure.Identity. These are standard dependencies for integrating with cloud services and are managed through the official .NET package manager. - Authentication Best Practices: The documentation correctly guides the use of
DefaultAzureCredential, which is the recommended approach for secure, managed-identity-based authentication, reducing the reliance on hardcoded secrets. - Retrieval of Connection Metadata: The skill demonstrates how to retrieve connection details using the
includeCredentialsparameter. This is a standard administrative feature of the SDK for accessing necessary resource metadata, though it should be used judiciously in secure environments. - Processing of External Inputs: The skill shows how agents ingest data through message creation and dataset uploads (e.g., in
SKILL.md). This ingested data can interact with capabilities like the Code Interpreter or Web Search. While no specific boundary markers are shown in the snippets, implementing them when handling untrusted data is a standard practice to ensure consistent agent behavior.
Audit Metadata