SkillAgentSearch skills...

UnrealCopilot

An Editor Plugin for Unreal Engine 5.6 (or later) that acts as a virtual technical artist, responding to user prompts by executing python code to perform editor actions. Unreal Copilot integrates Github models to provide access to a variety of LLMs and centralized quotas and billing via the user's Github account.

Install / Use

/learn @atgoldberg/UnrealCopilot
About this skill

Quality Score

0/100

Supported Platforms

GitHub Copilot
Zed

README

Unreal Copilot Plugin

Unreal Copilot is an integrated LLM agent plugin for Unreal Engine.
It acts as a technical artist development partner, enabling users to prompt the selected model to perform Unreal technical art and design tasks. The plugin generates and executes Python scripts directly in the Unreal Editor to streamline workflows for asset creation, level editing, and pipeline automation.


Features

  • 🔹 AI-Driven Technical Art Assistant
    Generate and execute Python code in-editor via natural language prompts.

  • 🔹 Task Automation
    Automate repetitive Unreal Editor tasks (level creation, asset setup, animation pipelines).

  • 🔹 Extensible LLM Integration
    Supports external model providers (GitHub Copilot, OpenAI, etc.) for flexible model access and billing.

  • 🔹 Iterative Development Framework
    Development follows a phased roadmap with clear milestones and deliverables.


Roadmap

Development is structured into iterative phases:

  1. Phase 1 – Core Plugin Skeleton

    • Create Unreal Engine plugin framework.
    • Enable Python execution inside the editor.
    • Establish communication layer with LLM API.
  2. Phase 2 – Prompt-to-Python Workflow

    • Send natural language prompts to the model.
    • Parse responses into executable Python scripts.
    • Execute safely in Unreal Editor.
  3. Phase 3 – Tooling & Workflow Expansion

    • Add support for asset creation, level editing, and animation scripting.
    • Introduce error-handling and rollback systems.
  4. Phase 4 – Advanced Features

    • Multi-agent workflows (planner, implementer, reviewer).
    • GitHub/Git integration for issue-based task execution.
    • Marketplace-ready distribution.

Requirements

  • Unreal Engine 5.6+
  • Python Editor Script Plugin (enabled in UE)
  • Access to a supported LLM (OpenAI API, GitHub Copilot, etc.)
  • Visual Studio 2022 (for plugin development)

Installation

  1. Clone this repository into your Unreal project’s Plugins/ folder:

    git clone https://github.com/YourOrg/UnrealCopilot.git
    
  2. Enable the plugin in Unreal Editor (Edit → Plugins → Installed).

  3. Restart the editor.


Usage

  1. Open the Unreal Copilot Panel inside Unreal Editor.
  2. Enter a prompt (e.g., “Create a new material with a scrolling texture.”).
  3. The model generates Python code → executes it in the editor.
  4. Review the results and iterate with follow-up prompts.

Development

  • Code style follows Unreal Engine C++ and Python guidelines.
  • Each feature is developed in its own branch and merged after review.
  • Agents produce a post-mortem.md after each task, documenting implementation details and test results.

References


License

MIT License – You are free to use, modify, and distribute this plugin under the terms of the MIT license.


Authors

Developed by Athomas Goldberg and collaborators.

View on GitHub
GitHub Stars8
CategoryDevelopment
Updated29d ago
Forks1

Languages

C++

Security Score

70/100

Audited on Mar 3, 2026

No findings