Skip to main content
Official AI workflow

PixiJS Skills

Teach your AI coding agent how to correctly use PixiJS v8 with the official pixijs-skills collection. One install, 40+ agents supported.

View pixijs-skillsAbout Agent Skills
Works with+40 more
pixijs-skills quick start
install → prompt → build
Install

Pick the path that fits your setup

The universal npx skills command works for most agents. The IDE-specific routes below are useful if you prefer native plugin or rule workflows.

Universal (recommended)

Works with 40+ agents

Drop the skill collection into your agent with a single command. No config required.

npx skills add https://github.com/pixijs/pixijs-skills

Claude Code

Use Claude's built-in plugin marketplace. Run this from Claude Code's slash-command prompt.

/plugin marketplace add pixijs/pixijs-skills
Agent Skills docs

Cursor

Settings → Rules → Add Rule → Remote Rule (GitHub) and paste pixijs/pixijs-skills. Or use the universal command above.

Cursor Rules docs

VS Code + Copilot

Copilot doesn't load skill files, so copy the official copilot-instructions.md into your repo instead.

curl -o .github/copilot-instructions.md \
  https://raw.githubusercontent.com/pixijs/pixijs-skills/main/.github/copilot-instructions.md
Copilot customization
Use

Prompt like you normally would

Once installed, describe your PixiJS task and the agent pulls in the relevant skills. You can also call the skill out by name when you want to be direct.

Example prompts

# Ask naturally, the agent picks up the skill when it fits
Build a PixiJS v8 scene with Assets, Container, Sprite, and ticker-based animation.

# Or be explicit
Use the pixijs skill and refactor this scene to follow current v8 patterns.
Why skills

Better PixiJS context for coding agents

Instead of hand-rolling prompts or pasting raw docs, point your agent at the official skill collection maintained by the PixiJS team.

Official & maintained

Authored and updated by the PixiJS team alongside the library itself.

40+ agents supported

One install command covers Cursor, Claude Code, Codex, Windsurf, Copilot, and more.

Targeted expertise

25 focused skills across Application, Assets, Graphics, Filters, Mesh, Performance, and more.

Raw docs

Need plain-text documentation too?

Prefer the skills workflow whenever possible, but these files remain useful for tools that accept a llms.txt URL or for custom GPTs.

llms.txt

Lightweight navigation index optimized for quick context loading.

  • Links to all documentation pages
  • Short descriptions for each section
  • Organized by category
DownloadView

llms-medium.txt

Complete guides and tutorials without the full API reference.

  • Complete guides and tutorials
  • Optimized for medium context windows
  • No API reference (smaller size)
DownloadView

llms-full.txt

Complete API reference with all details in a single file.

  • Full documentation for all classes & functions
  • Parameters, methods, and properties
  • Working code examples
DownloadView

llms.txt is a standardized format for AI-readable project documentation. It complements the official skills when you want raw context alongside a task-aware workflow.