What problem does it solve? Game projects built by AI agents often fail because architectural decisions are made inconsistently or left implicit. This Skill runs a structured, step-by-step workflow that produces a complete architecture document — engine selection, networking model, physics, save systems, and implementation patterns — so every AI agent implements against the same decisions. ## Core Features & Use Cases - Guided Decision Workflow: Walks through sequential step files covering engine choice, rendering, physics, networking, input, AI, persistence, and deployment, with a decision catalog mapping requirements to compatible technology stacks. - Engine-Specific Knowledge: Loads deep knowledge fragments for Godot, Unity, Unreal, Phaser, or Roblox only for the selected engine, plus a catalog of MCP servers that give AI agents direct editor access. - Validation Checklist: Ships a checklist that verifies decision completeness, version specificity, starter template integration, and AI-agent implementability before coding begins. - Use Case: A solo developer planning a multiplayer roguelike uses this Skill to decide between Godot and Unity, select a netcode model, define save schema conventions, and produce an architecture document that coding agents follow without ambiguity. ## Quick Start Ask the AI to design the game architecture for your project by describing your game concept and target platforms.