godogen

Generate complete Godot games from natural-language design briefs.

Updated Aug 11, 2026
One-click install
npx skills add https://github.com/zerox-core/AI_Game --skill godogen-zerox-core
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: godogen
Source: https://github.com/zerox-core/AI_Game/tree/main/.claude/skills/godogen
Command: npx skills add https://github.com/zerox-core/AI_Game --skill godogen-zerox-core

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires rembg, onnxruntime-gpu, and includes scripts (resource) components.

What problem does it solve? Building a complete Godot game from scratch requires coordinating architecture design, scene construction, asset generation, and visual verification — a long, error-prone process. This Skill orchestrates that entire pipeline from a single natural-language description. ## Core Features & Use Cases - End-to-End Game Pipeline: Decomposes a design brief into a task plan, scaffolds a Godot C# project, generates scenes programmatically, and verifies results with screenshot and video capture. - Budgeted Asset Generation: Plans and generates images, textures, animated sprites, and 3D GLB models via Gemini, Grok, and Tripo3D within a user-provided cost budget. - Visual Verification: Captures deterministic frame sequences and encodes proof videos so broken placement, scale, or motion is caught instead of trusted from code alone. - Use Case: Describe a 3D driving game with a $5 asset budget; the Skill produces a reference image, task plan, project scaffold, generated car model and textures, verified gameplay, and a final video proof bundle. ## Quick Start Use godogen to build a Godot game from this design brief: a top-down 2D space shooter with three enemy types and a score HUD.

Frequently Asked Questions about godogen

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate a Godot game from a text description?

Provide a natural-language design brief and the Skill runs a staged pipeline: visual target image, task decomposition, project scaffold, asset generation, task execution, and capture-based verification. Progress is tracked in PLAN.md, STRUCTURE.md, MEMORY.md, and ASSETS.md so work can resume after interruptions.

How does AI asset generation work for Godot games?

The asset planner budgets and generates images via Gemini or Grok, converts approved images to GLB models via Tripo3D, and can rig biped characters with retargeted animations. Costs are tracked per operation, from 2 cents for a texture to about 92 cents for a rigged character with three animations.

Can I export a Godot project as an Android APK?

Yes, the Skill includes an Android build stage that enables ETC2/ASTC texture compression, creates export_presets.cfg, and exports a debug APK using the Godot .NET editor with Mono. It requires the .NET 9 SDK, OpenJDK 17, Android SDK, and export templates.

Does the pipeline verify the game actually works visually?

Yes, it captures deterministic frame sequences with --fixed-fps and encodes a 15-30 second video.mp4 proof bundle under screenshots/result/{N}/. Code alone is not trusted; broken placement, scale, or motion must be caught by inspecting captured frames.

What are the limitations of AI-generated game assets?

Generators handle left/right orientation poorly, so one direction is generated and flipped at runtime. Minimum 1K generation resolution makes tiny sprites muddy, and video-extracted animations need correct frame-rate playback around 24 fps to avoid choppiness.