🤖 Macbourg AI Agents

— A mini software factory powered by AI agents.
One prompt → Plan + Code + Tests + Review.

Inspired by the Antfarm architecture: each agent has a single responsibility and no agent reviews its own work.

---

Architecture

User Prompt
    ↓
📋 Planner   — Breaks request into dev tasks
    ↓
💻 Developer — Writes production Python code
    ↓
🧪 Tester    — Creates pytest unit tests
    ↓
✅ Verifier  — Reviews code & tests, scores quality
    ↓
Final Output

---

Quick Start

1. Clone & setup

git clone https://github.com/swordenkisk/macbourg-agents.git
cd macbourg-agents
python setup.py

2. Add your API key

# Edit .env — add your Anthropic or OpenAI key
nano .env

ANTHROPIC_API_KEY=sk-ant-...

3. Activate venv & run

# Mac / Linux
source .venv/bin/activate

# Windows
.\.venv\Scripts\activate

# Run
python main.py

4. Or pass a prompt directly

python main.py "Build a Python REST API for user authentication"

---

Project Structure

macbourg-agents/
├── main.py              ← Entry point
├── setup.py             ← One-command setup
├── requirements.txt     ← Dependencies
├── .env.example         ← API key template
├── .gitignore
├── src/
│   ├── agents.py        ← 4 agent definitions
│   ├── pipeline.py      ← Pipeline orchestrator
│   └── llm_client.py    ← Anthropic / OpenAI client
└── tests/
    └── test_agents.py   ← Pytest unit tests

---

Run Tests

pytest tests/ -v

---

Supported LLMs

Provider	Model (default)	Key in .env
Anthropic	claude-sonnet-4-20250514	ANTHROPIC_API_KEY
OpenAI	gpt-4o-mini	OPENAI_API_KEY

The client auto-detects which key is present. Anthropic takes priority.

---

Web UI Version

A React-based visual pipeline dashboard is also included in macbourg-agents.jsx — runs inside Claude.ai artifacts.

---

License

MIT — built by @swordenkisk