hatayama/unity-cli-loop

What it solves

Unity CLI Loop allows AI agents to autonomously develop, test, and operate Unity projects. It removes the need for a human to manually compile code, run tests, check logs, and manipulate the Unity Editor, enabling a continuous, AI-driven development loop.

How it works

The project provides a CLI and a set of "Skills" (tool definitions) that LLMs can use to interact with the Unity Editor. It exposes essential Unity operations—such as compilation, log retrieval, and scene hierarchy analysis—as tools. A key feature is the execute-dynamic-code tool, which allows the AI to run arbitrary C# code within the editor to perform complex tasks without needing a massive library of predefined tools.

Who it’s for

Unity developers who use AI-powered IDEs or agents (like Claude Code, Cursor, or Codex) and want to automate the iterative process of coding, testing, and refining their game or application.

Highlights

• Autonomous Dev Loop: AI can compile, run tests, and inspect logs to fix bugs independently.
• Dynamic C# Execution: Enables the AI to perform almost any Editor operation via dynamic code snippets.
• PlayMode Automation: Tools for simulating mouse and keyboard input, recording and replaying gameplay, and verifying behavior.
• Visual Feedback: Ability to take screenshots of specific Editor windows to provide the AI with visual context.
• Integration Ready: Supports installation via Unity Package Manager or OpenUPM and integrates with popular LLM tools via CLI or MCP.