This project provides a structured, multi-agent framework for AI-assisted independent game development, addressing the common lack of organization in single-session AI development. It targets developers using Claude Code, offering a simulated game studio environment with specialized AI agents to guide projects from concept to release, ensuring better coherence and quality.

How It Works

The core architecture comprises 48 specialized AI agents organized hierarchically into Director, Department Head, and Expert tiers, mimicking a real game studio structure. Agents operate via a structured delegation model, allowing vertical and horizontal communication, with defined conflict resolution and change propagation mechanisms. Development is guided by 37 slash commands (skills) for specific workflows, 8 automated hooks for validation (e.g., Git commits, asset changes), and 11 path-scoped rules enforcing coding standards. This approach provides structure, specialized expertise, and automated checks, preventing common AI development pitfalls while ensuring user control over all decisions.

Quick Start & Requirements

• Primary install/run command: Clone the repository (git clone ...), navigate into the directory, and initiate a Claude Code session (claude). Use the /start command to begin the guided workflow or jump directly to specific skills like /brainstorm or /setup-engine.
• Prerequisites: Git, Claude Code CLI (npm install -g @anthropic-ai/claude-code). jq and Python 3 are recommended for full hook validation but are optional.
• Links: UPGRADING.md for migration guides, CLAUDE.md for main configuration, and various files within .claude/ and docs/ for detailed agent, skill, hook, rule, and setup information.

Highlighted Details

• Features 48 specialized agents covering design, programming, art, audio, narrative, QA, and production.
• Includes 37 skills (slash commands) for workflows such as /design-review, /code-review, /sprint-plan, and /brainstorm.
• Automated hooks provide validation on Git commits, pushes, asset changes, and session lifecycle events.
• Path-scoped rules enforce coding standards based on file location (e.g., src/gameplay/, src/ai/).
• Offers engine-specific agent sets for Godot 4, Unity, and Unreal Engine 5.
• Employs a strict collaboration protocol where agents present options and await user decisions.

Maintenance & Community

This project is under active development. Community interaction, questions, and project showcases are encouraged via GitHub Discussions. Bug reports and feature requests should be submitted through GitHub Issues.

Licensing & Compatibility

The project is released under the MIT License, permitting broad use, modification, and distribution, including for commercial purposes, subject to the standard MIT license terms.

Limitations & Caveats

This project functions as a template and requires user customization, including adding/removing agents, editing prompts, and modifying skills or rules. Optional tools (jq, Python 3) are recommended for full hook functionality; hooks degrade gracefully if these are missing. As an actively developed project, expect ongoing changes and potential evolution of its architecture and features.