Create AI agents and workflows using Microsoft Agent Framework SDK. Supports single-agent and multi-agent workflow patterns. USE FOR: create agent, build agent, scaffold agent, new agent, agent framework, workflow pattern, multi-agent, MCP tools, create workflow. DO NOT USE FOR: deploying agents (use deploy), evaluating agents (use agent/evaluate), Azure AI Foundry agents without Agent Framework SDK.
Build AI agents, agentic apps, and multi-agent workflows using Microsoft Agent Framework SDK.
| Property | Value |
|---|---|
| SDK | Microsoft Agent Framework (Python) |
| Patterns | Single Agent, Multi-Agent Workflow |
| Server | Azure AI Agent Server SDK (HTTP) |
| Debug | AI Toolkit Agent Inspector + VSCode |
| Best For | Enterprise agents with type safety, checkpointing, orchestration |
Use when the user wants to:
1.0.0b260107)| Topic | File | Description |
|---|---|---|
| Server Pattern | references/agent-as-server.md | HTTP server wrapping (production) |
| Debug Setup | references/debug-setup.md | VS Code configs for Agent Inspector |
| Agent Samples | references/agent-samples.md | Single agent, tools, MCP, threads |
| Workflow Basics | references/workflow-basics.md | Executor types, handler signatures, edges, WorkflowBuilder — start here for any workflow |
| Workflow Agents | references/workflow-agents.md | Agents as executor nodes, linear pipeline, run_stream event consumption |
| Workflow Foundry | references/workflow-foundry.md | Foundry agents with bidirectional edges, loop control, register_executor factories |
💡 Tip: For advanced patterns (Reflection, Switch-Case, Fan-out/Fan-in, Loop, Human-in-Loop), search
microsoft/agent-frameworkon GitHub.
This skill delegates to microsoft-foundry MCP tools for model and project operations:
| Tool | Purpose |
|---|---|
foundry_models_list | Browse model catalog for selection |
foundry_models_deployments_list | List deployed models for selection |
foundry_resource_get | Get project endpoint |
.vscode/ configsRead reference files based on user's request:
Always read these references:
.vscode/ configs)Read the relevant code sample:
Model Selection: Use microsoft-foundry skill's model catalog to help user select and deploy a model.
Recommended: Search microsoft/agent-framework on GitHub for advanced patterns.
Decide on the model BEFORE coding.
If user hasn't specified a model, use microsoft-foundry skill to list deployed models or help deploy one.
ALWAYS create/update .env file:
FOUNDRY_PROJECT_ENDPOINT=<project-endpoint>
FOUNDRY_MODEL_DEPLOYMENT_NAME=<model-deployment-name>
All three are required by default:
agent-as-server.md — this is the default entry point.vscode/launch.json and .vscode/tasks.json using templates from debug-setup.md⚠️ Warning: Only skip server mode or debug configs if the user explicitly requests a "minimal" or "no server" setup.
requirements.txt# pin version to avoid breaking changes
# agent framework
agent-framework-azure-ai==1.0.0b260107
agent-framework-core==1.0.0b260107
# agent server (for HTTP server mode)
azure-ai-agentserver-core==1.0.0b10
azure-ai-agentserver-agentframework==1.0.0b10
# debugging support
debugpy
agent-dev-cli
⚠️ Warning: Never use bare
pythonorpip— always use the venv-activated versions or full paths (e.g.,.venv/bin/pip).
Enter a run-fix loop until no startup errors:
.venv/Scripts/python main.py on Windows, .venv/bin/python main.py on macOS/Linux)Guardrails:
Create/update README.md with setup instructions and usage examples.
| Error | Cause | Resolution |
|---|---|---|
ModuleNotFoundError | Missing SDK | Run pip install agent-framework-azure-ai==1.0.0b260107 in venv |
AgentRunResponseUpdate not found | Wrong SDK version | Pin to 1.0.0b260107 (breaking rename in newer versions) |
| Agent name validation error | Invalid characters | Use alphanumeric + hyphens, start/end with alphanumeric, max 63 chars |
| Async credential error | Wrong import | Use azure.identity.aio.DefaultAzureCredential (not azure.identity) |