responses-agent-patterns

What problem does it solve?

This Skill addresses the complexities of building reliable and production-ready GenAI agents on Databricks by providing essential patterns for MLflow ResponsesAgent, ensuring compatibility with the AI Playground and robust authentication.

Core Features & Use Cases

• Mandatory ResponsesAgent: Enforces the use of mlflow.pyfunc.ResponsesAgent for all new agents, guaranteeing AI Playground compatibility and automatic signature inference.
• Streaming Responses: Implements patterns for predict_stream to deliver real-time, chunked responses.
• OBO Authentication: Provides a robust pattern for On-Behalf-Of authentication, crucial for secure data access in Model Serving environments.
• MLflow Tracing: Integrates MLflow tracing for detailed observability of agent operations.
• Use Case: When developing a new customer support chatbot agent on Databricks, use these patterns to ensure it correctly handles user requests, streams responses, and securely accesses user data via OBO authentication, all while being traceable in MLflow.

Quick Start

Use the responses-agent-patterns skill to implement streaming responses for your new Databricks GenAI agent.