🚀 Azure AI Foundry Agent – Python Test Harness

This project is a lightweight, production-style Python application designed to connect to and interact with Azure AI Foundry agents using the azure.ai.agents SDK.

It provides a clean baseline for:

• Validating connectivity to Foundry-deployed agents
• Sending and receiving messages
• Building reusable agent interaction patterns
• Serving as a foundation for more advanced AI-enabled applications

---

📌 Project Goals

This repository focuses on two primary scripts:

🔹 basic_test.py

A minimal validation script that:

• Authenticates using Azure credentials
• Connects to a Foundry project endpoint
• Retrieves an existing agent
• Sends a simple message
• Prints the agent’s response

👉 Purpose: Quickly confirm that your environment, authentication, and agent connectivity are working correctly.

---

🔹 image_test.py (optional/extended)

An extended test script that:

• Accepts user input (text and/or image)
• Sends structured messages to the agent
• Demonstrates more advanced interaction patterns

👉 Purpose: Serve as a foundation for building user-driven or multimodal agent workflows.

---

🛠️ Prerequisites

Before getting started, ensure you have:

• ✅ Python 3.10+
• ✅ VS Code (recommended)
• ✅ Azure CLI installed
• ✅ Access to an Azure AI Foundry project
• ✅ A deployed agent (created via UI or API)

---

📥 Setup Instructions

1️⃣ Clone the Repository

git clone https://github.com/<your-username>/<your-repo>.git
cd <your-repo>

---

2️⃣ Open in VS Code

code .

---

3️⃣ Create a Virtual Environment

python -m venv .venv

---

4️⃣ Activate the Virtual Environment

Windows (PowerShell):

.venv\Scripts\Activate

Mac/Linux:

source .venv/bin/activate

---

5️⃣ Select the Python Interpreter in VS Code ⚠️

Even after activating the virtual environment, VS Code may still be using a different Python interpreter.

To ensure the correct one is selected:

Press Ctrl + Shift + P Search for: Python: Select Interpreter Choose the interpreter that points to: .venv\Scripts\python.exe (Windows)

or

.venv/bin/python (Mac/Linux)

✅ You should now see .venv in the bottom-right corner of VS Code.

---

5️⃣ Install Dependencies

pip install -r requirements.txt

---

6️⃣ Authenticate to Azure

az login

This allows DefaultAzureCredential to work automatically.

---

7️⃣ Configure Environment Variables

Copy the example file:

copy .env.example .env

(or on Mac/Linux)

cp .env.example .env

Then update .env with your values:

AZURE_AI_PROJECT_ENDPOINT=https://<your-resource>.services.ai.azure.com/api/projects/<your-project>
AZURE_AGENT_ID=asst_xxxxxxxxxxxxxxxxx

⚠️ Important:

• AZURE_AGENT_ID must be the actual agent ID (not name)
• .env is not committed to git (by design)

---

▶️ Run the Basic Test

python src/basic_test.py

✅ Expected Output

• Successful authentication message
• Agent retrieval confirmation
• A response from your agent

---

🧪 Troubleshooting

❌ ValueError: No value for given attribute

Cause: Missing or incorrect environment variable Fix: Verify .env contains:

AZURE_AGENT_ID=...

---

❌ Authentication Errors

If DefaultAzureCredential fails:

az login

Or ensure your account has access to the Foundry project.

---

❌ .env Not Loading

Ensure:

• .env is in the project root
• Script uses:

load_dotenv(Path(__file__).resolve().parents[1] / ".env")

---

🧱 Project Structure

.
├── src/
│   ├── basic_test.py
│   ├── image_test.py
│
├── .env.example
├── requirements.txt
├── README.md

---

🔐 Security Notes

• .env is excluded via .gitignore
• Never commit credentials or secrets
• Use Azure RBAC for access control

---

🚧 Future Enhancements

• Reusable agent client wrapper
• CLI interface for agent interaction
• Structured logging and error handling
• Integration with APIs / front-end apps
• Multi-agent orchestration patterns

---

💡 Why This Project Matters

This project demonstrates:

• Real-world Azure AI Foundry integration
• Secure credential handling
• Clean Python environment setup
• Practical agent interaction patterns

It is intentionally designed to be:

✅ Simple to run ✅ Easy to extend ✅ Useful as a learning and portfolio artifact

---

📬 Contributions / Ideas

Feel free to fork and expand:

• Add new agent workflows
• Integrate with cloud services (AWS/Azure hybrid)
• Build UI layers or APIs