azure-ai-agents-persistent-dotnet
Pass
Audited by Gen Agent Trust Hub on May 14, 2026
Risk Level: SAFE
Full Analysis
- [Indirect Prompt Injection Surface]: The skill is designed to process untrusted data from user messages, files, and search results. While this is the intended purpose of the SDK, it represents a surface where instructions could be embedded in data. Production implementations should consider input validation and the use of boundary markers.
- [Tool Access and Execution]: The skill enables capabilities such as Code Interpreter and Function Calling. These tools allow the agent to execute code or call external APIs, which are standard features for this SDK but require careful configuration of permissions and environment isolation.
- [Resource Management]: The skill demonstrates best practices by using environment variables for configuration and
DefaultAzureCredentialfor secure authentication, avoiding hardcoded secrets.
Audit Metadata