Agentic AI - Multi-Agent Software Development System

Autonomous AI agents that automate the complete software development lifecycle

Built with LangChain + Anthropic Claude | Python 3.9+

---

🎯 What is This?

This project implements an Agentic AI architecture - a system where multiple specialized AI agents autonomously collaborate to handle software development tasks from requirements analysis to code review.

Unlike traditional LLM assistants that respond to single queries, these agents:

• 🧠 Plan multi-step workflows autonomously
• 🤝 Collaborate through structured context passing
• 🛠️ Use tools (Git, file system, tests, etc.)
• 💾 Remember decisions and execution history
• 🔄 Iterate until objectives are met

🏗️ Architecture

The system features 5 core specialized agents coordinated by an orchestrator:

Agent	Role	What It Does
🎯 Orchestrator	Workflow Manager	Plans and coordinates multi-agent workflows
📋 Requirements	Analysis	Transforms user stories into structured specifications
🏛️ Design	Architecture	Proposes software architecture and design patterns
💻 Development	Coding	Generates production-ready code with best practices
🧪 Tests	Quality Assurance	Creates comprehensive test suites (unit/integration/functional)
✅ Quality	Code Review	Reviews code and suggests improvements

Optional Agents (extensible):

• CI/CD Agent: Build pipelines and deployment
• Ops Agent: Observability and monitoring

⚡ Quick Start (5 minutes)

Automated Setup

# 1. Run setup script
chmod +x setup.sh
./setup.sh

# 2. Add your Anthropic API key
echo "ANTHROPIC_API_KEY=sk-ant-your-key-here" > .env

# 3. Activate environment
source venv/bin/activate

# 4. Run example workflow
python examples/simple_workflow.py

# 5. Check outputs
ls -la output/

Manual Setup

# Create and activate virtual environment
python3 -m venv venv
source venv/bin/activate  # Windows: venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

# Configure API key
cp .env.example .env
# Edit .env and add: ANTHROPIC_API_KEY=sk-ant-...

# Run example
python examples/simple_workflow.py

Get API Key: https://console.anthropic.com/

Project Structure

.
├── src/
│   ├── agents/          # Agent implementations
│   ├── tools/           # Tool integrations (git, tests, build)
│   ├── memory/          # Memory and observability layer
│   ├── config/          # Configuration management
│   └── orchestrator.py  # Main orchestrator
├── examples/            # Example workflows
├── tests/               # Test suite
└── data/                # Memory storage (gitignored)

Usage

Example: Implement a Feature

from src.orchestrator import Orchestrator

# Initialize orchestrator
orchestrator = Orchestrator()

# Define objective
objective = """
Implement a user authentication feature with:
- Login endpoint
- JWT token generation
- Password hashing
- Unit tests
"""

# Execute workflow
result = await orchestrator.execute(objective)
print(result)

Development

# Run tests
pytest tests/

# Format code
black src/

# Lint code
ruff src/

📚 Documentation

QUICKSTART.md | Detailed installation and usage guide

ARCHITECTURE.md | Technical architecture and design patterns

🚀 Example Workflows

Try these objectives:

# Simple API endpoint
"Create a REST API endpoint for user registration with validation"

# CLI application
"Build a command-line todo list with add, list, and complete commands"

# Data processing
"Implement a CSV to JSON converter with data validation"

# Full feature
"Design and implement user authentication with JWT tokens and tests"

Results are saved to output/ with:

• 📋 specifications.md - Requirements and acceptance criteria
• 🏛️ architecture.md - Design and component breakdown
• 💻 implementation.md - Production-ready code
• 🧪 tests.md - Comprehensive test suite
• ✅ code_review.md - Quality analysis and suggestions

🛠️ Tech Stack

Core:

• Python 3.9+
• LangChain (agent framework)
• LangChain-Anthropic (Claude integration)
• Anthropic Claude Sonnet 4.5

Development:

• pytest (testing)
• Pydantic (data validation)
• asyncio (concurrency)
• Black (formatting)
• Ruff (linting)

📝 License & Attribution

Educational project for IA Pour le Génie Logiciel course.

🆘 Troubleshooting

Import errors?

source venv/bin/activate
pip install -r requirements.txt

API key issues?

cat .env | grep ANTHROPIC_API_KEY