What problem does it solve? Building 3D RTS or large-scale web games in Three.js often leads to performance bottlenecks: duplicate asset loads, excessive draw calls, unbounded animation updates, and runtime memory allocation that stalls the main thread. This Skill enforces a disciplined asset pipeline so thousands of units render smoothly within strict CPU and GPU budgets. ## Core Features & Use Cases - Unified Asset Caching: Implements an AssetManager that loads each GLB exactly once, caches it globally, and clones instances via SkeletonUtils without duplicating materials. - Compression & Texture Pipeline: Binds DRACOLoader, MeshoptDecoder, and KTX2Loader to GLTFLoader for minimal network payloads and GPU texture memory. - Performance Guardrails: Enforces draw call limits, LOD meshes, animation culling by camera distance, shadow restrictions, and object pooling with typed arrays. - Use Case: When building an RTS game with hundreds of on-screen units, use this Skill to refactor naive loader code into a pooled, compressed, single-draw-call asset architecture that passes the pre-shipping audit checklist. ## Quick Start Use the threejs-glb-pipeline skill to refactor my GLTFLoader code into a cached asset manager with Draco compression and SkeletonUtils cloning.
