🧠 Jules Code Team Template

A Drop-in Multi-Agent AI Coding Squad

Stop coding alone. This template scaffolds a complete "Coding Squad" architecture into your new project. It turns a standard LLM session into a simulated team of specialized agents who debate, critique, and verify your code before it ships.

---

⚡ Quick Start

Do not clone this repo manually. Use the initialization script to wake up the agents and configure your project.

1. Initialize the Squad

Run the onboarding script to configure your team, set governance, and unpack the project structure. (Requires Python 3+. No external dependencies or installation needed).

python template_source/scripts/init_project.py

2. Answer the "Interview"

Brain will wake up and ask you:

• Mission: What are we building? (SaaS, Game, Library?)
• Governance: Democracy (Debate) or Dictatorship (Speed)?
• Risk Tolerance: High (Move fast) or Low (Security first)?

The script will then unpack the agent system, configure the personas, and prepare your repository for work.

---

💻 Usage (Direct Execution)

Once initialized, you can run the agent system directly from the command line:

python3 -m src.main --task "Optimize the database query"

Or process a specific file:

python3 -m src.main --file path/to/file.py

Key Arguments:

• --task <description>: Natural language description of the task.
• --file <path>: Path to a file to be processed.

---

👥 Meet The Squad

This template is not just prompts; it's a System of Adversarial Interoperability. Agents have conflicting goals to ensure balance.

Agent	Role	Focus
Brain 🧠	The Architect	Dialectic simulation. Resolves disputes and enforces the Roadmap.
Boom 💥	Product	Features & "Vibes". Wants to ship MVP now.
Bolt ⚡	Performance	Speed & Efficiency. Hates O(n^2) complexity and bloat.
Sentinel 🛡️	Security	Paranoid defense. Checks for injection, CVEs, and bad patterns.
Scope 🔬	QA & Testing	The Cynic. "Everything breaks." Writes edge cases and stress tests.
Scribe 📜	Docs	The Historian. Ensures TEAM_MEMORY.md is updated and code is readable.
Orbit 🛰️	DevOps	Infrastructure & CI/CD. Handles release management and containers.
Palette 🎨	UX/UI	Accessibility & Design. Fights for the user's experience.

---

🏗️ Architecture

The system operates via a structured execution graph managed by the Nexus Bus and Graph Executor.

graph TD
    User[User Input] --> Brain[🧠 Brain: Architect]
    Brain -->|Generate Graph| NexusBus[📡 Nexus Bus]
    NexusBus -->|Validate| Schema[📜 JSON Schema]
    NexusBus -->|Dispatch| Executor[🤖 Graph Executor]

    subgraph "Execution Loop"
        Executor -->|Traverse| Node[Graph Node]
        Node -->|Action| ToolRegistry[🛠️ Tool Registry]
        ToolRegistry -->|Invoke| Function[Python Function]
        Function -->|Result| Node
    end

    Node -->|Success| NextNode[Next Node]
    Node -->|Failure| Repair[Repair Loop]
    Repair -->|Retry| Node

Loading

🚀 Why use this?

1. Vibe Coding with Seatbelts: Enjoy the speed of AI coding, but with Sentinel and Scope ensuring you don't commit security vulnerabilities or logic bugs.
2. Context Awareness: The system uses smart_ingest.py (available in template_source/scripts/) to maintain a token-optimized "Memory" of your project, so you don't have to copy-paste files constantly.
3. Clean Architecture: All agent logic is hidden in .agents/. Your src/ folder stays clean.

---

🐛 Known Issues

Please see BUGS.md for a list of known issues and audit findings.

🤝 Contributing

We welcome "Vibe Coders" and prompt engineers! Please read CONTRIBUTING.md before submitting a Pull Request.

📄 License

This project is licensed under the GNU Affero General Public License v3.0.