WebGPU Skill

This skill helps any agent design, implement, and debug WebGPU applications and GPU compute pipelines. It is framework-agnostic and focuses on reusable WebGPU/WGSL patterns.

What this skill covers

• WebGPU initialization, device setup, and surface configuration
• Compute pipelines, workgroup sizing, and storage buffer layout
• Render pipelines, render passes, and post-processing patterns
• GPU/CPU synchronization and safe readback strategies
• Performance and debugging practices
• Architecture patterns: modular passes, phase-based simulation, and capability handling
• Use cases: rendering, compute, ML training/inference, grid simulations, and systems modeling

Core principles

• Choose a capability strategy: fallback runtime, reduced mode, or fail fast.
• Avoid full GPU readbacks in hot paths; use localized queries or small readback buffers.
• Structure simulation with phases (state, apply, integrate, constrain, correct) to keep WGSL cohesive.
• Use spatial grids or other spatial indexing for neighbor queries and high particle counts.
• Build modular passes so render and compute stages stay composable and testable.

How to use this skill

When asked to build a WebGPU feature:

1. Confirm the target platform and WebGPU support expectations.
2. Propose a resource layout (buffers, textures, bind groups) with a simple data model.
3. Sketch the pipeline graph (compute vs render passes) and dependencies.
4. Provide minimal working code and scale up with performance constraints.
5. Choose a capability strategy when WebGPU is unavailable.

Deliverable checklist

• Clean WebGPU init and error handling
• A buffer layout with alignment notes (16-byte struct alignment for WGSL)
• A pass graph with clear read/write ownership (ping-pong textures if needed)
• Explicit notes on readback and when it is safe
• Optional fallback or reduced mode for critical functionality

Quick reference

See REFERENCE.md for a compact WebGPU cheat sheet and docs/ for deeper patterns, including docs/use-cases.md and docs/simulation-patterns.md.