Why Use Custom Tools
We explore the trade-offs of providing custom tools to LLMs, discussing how restricting tool freedom improves security (e.g., preventing data exfiltration via curl) while reducing flexibility, and how running the LLM in a virtual environment offers an alternative balance between safety and capability.
Previous Module
Optimizing Generative AI Chat Apps
This Module
Develop AI Agents on Azure
Developing AI Agents on Azure
Why Use Custom Tools
Custom Functions and Azure Functions
Logic Apps and OpenAPI Specifications
MCP vs OpenAPI
Integrating MCP Tools with Azure AI Agents
Retrieval Augmented Generation for Agents
Exploring Foundry IQ
Configuring Data Sources and Retrieval in Foundry IQ