Adjust autogen to new structure

Author: Wojciech-Rebisz

CONTRIBUTING.md

@@ -34,7 +34,7 @@ You can optionally add a scope (e.g. the agent or module name) in parentheses af
 ### Examples
 
 ```
-feat: add health check endpoint to autogen_agent
+feat: add health check endpoint to autogen mcp_agent
 fix: correct env var name in deployment in langgraph_react_agent
 docs: update README with OpenShift deploy steps
 test: add tests for tool registration

README.md

@@ -22,6 +22,7 @@ Agents are organized by framework. Pick one and follow its README:
 | **LlamaIndex** | [WebSearch Agent](./agents/llamaindex/websearch_agent/) | Agent built on LlamaIndex that uses a web search tool to query the internet and use the results in its answers. |
 | **CrewAI** | [WebSearch Agent](./agents/crewai/websearch_agent/) | CrewAI-based agent with a web search tool to query the internet and answer user questions. |
 | **Vanilla Python** | [OpenAI Responses Agent](./agents/vanilla_python/openai_responses_agent/) | Minimal agent with no framework: only the OpenAI Python client and an Action/Observation loop with tools. Use with OpenAI or any compatible API. |
+| **AutoGen** | [MCP Agent](./agents/autogen/mcp_agent/) | AutoGen AssistantAgent with MCP tools over SSE (e.g. churn prediction, math tools), FastAPI `/chat/completions`. |
 
 ## Deployment Options
 
@@ -54,8 +55,10 @@ agentic-starter-kits/
 │   │   └── websearch_agent/         # CrewAI web search agent
 │   ├── llamaindex/
 │   │   └── websearch_agent/         # LlamaIndex web search agent
-│   └── vanilla_python/
-│       └── openai_responses_agent/  # OpenAI Responses API (no framework)
+│   ├── vanilla_python/
+│   │   └── openai_responses_agent/  # OpenAI Responses API (no framework)
+│   └── autogen/
+│       └── mcp_agent/               # AutoGen + MCP (SSE)
 └── README.md
 ```
 

agents/autogen/README.md

@@ -0,0 +1,14 @@
+# AutoGen Agents
+
+Agent templates built with [AutoGen](https://microsoft.github.io/autogen/) (Microsoft).
+
+## Available Agents
+
+| Agent | Description |
+|-------|-------------|
+| [mcp_agent](mcp_agent/) | AssistantAgent with MCP tools over SSE (e.g. churn deployment, `add`/`sub`), FastAPI `/chat/completions` |
+
+## Resources
+
+- [AutoGen Documentation](https://microsoft.github.io/autogen/)
+- [Model Context Protocol (MCP)](https://modelcontextprotocol.io/)

agents/base/autogen_agent/Dockerfile agents/autogen/mcp_agent/Dockerfileagents/base/autogen_agent/Dockerfile renamed to agents/autogen/mcp_agent/Dockerfile

@@ -18,7 +18,7 @@ The agent uses AutoGen (AssistantAgent) and `autogen_ext.tools.mcp`: it connects
 Go to the agent directory:
 
 ```bash
-cd agents/base/autogen_agent
+cd agents/autogen/mcp_agent
 ```
 
 #### The .env file
@@ -93,15 +93,15 @@ uv pip install -e .
 1. **Terminal 1 — MCP server** (must be running before the agent):
 
    ```bash
-   cd agents/base/autogen_agent/mcp_automl_template
+   cd agents/autogen/mcp_agent/mcp_automl_template
    uv pip install -e .
    PORT=8000 uv run python mcp_server.py
    ```
 
 2. **Terminal 2 — Agent:**
 
    ```bash
-   cd agents/base/autogen_agent
+   cd agents/autogen/mcp_agent
    uv run uvicorn main:app --host 0.0.0.0 --port 8080
    ```
 
@@ -126,7 +126,7 @@ Optional: interactive chat with MCP tools (LangGraph) — from the `mcp_automl_t
 
 ```bash
 cd mcp_automl_template
-# Set MCP_SERVER_URL in .env (or in agents/base/autogen_agent/.env)
+# Set MCP_SERVER_URL in .env (or in agents/autogen/mcp_agent/.env)
 uv run python interact_with_mcp.py
 ```