9. References

Continue Your AI Learning Journey

Microsoft Agent Framework SDK Docs — Official API reference and guides
Azure AI Foundry — Model catalog, evaluation, and deployment
Microsoft Learn: AI Agents — Free self-paced learning paths
Semantic Kernel — Open-source AI orchestration SDK
AutoGen — Multi-agent conversation framework from Microsoft

Quick Reference Guide

Task	Command / Code Snippet
Create agent	`new Agent({ name, instructions, model })`
Add memory	`agent.memory = new VolatileMemory()`
Add tool	`agent.addTool(new Tool({ name, handler }))`
Create orchestrator	`new Orchestrator({ agents, defaultAgent })`
Hand off	`orchestrator.handoff(from, to, { context })`
Add guardrail	`new ContentSafetyPolicy({ blockHateSpeech: true })`
Send message	`client.sendMessage("Hello")`

Glossary

Term	Definition
Agent	AI entity with instructions, memory, and tools
Orchestrator	Coordinates multi-agent workflows and handoffs
Tool	Function an agent can call to perform actions
Memory	Persistent storage for agent context
Guardrail	Safety/compliance policy for inputs/outputs
Handoff	Transfer conversation from one agent to another

Bonus Lecture: What's Next?

After completing this guide, you're ready to explore:

Copilot Extensions — Build agents that extend Microsoft 365 Copilot
Custom Model Fine-tuning — Tailor LLMs to your domain
Agent-to-Agent Communication — Decentralized agent networks
Human-in-the-Loop — Escalation workflows for critical decisions

Resources & Discussion

MAF GitHub Repository: https://github.com/microsoft/agents
Stack Overflow: Tag your questions with microsoft-agents
Azure AI Community: Discussion forums and events

📘 Tip: Join the Microsoft Agent Framework community to share your projects, ask questions, and stay updated on the latest features.

✏️ Final Challenge: Build an end-to-end agent that solves a real problem. It should use memory, at least 3 tools, multi-agent orchestration, and be deployed with observability. Share your project with the community!