How do I install the Create World skill?

Run npx skillmds@latest add frank890417/create-world in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Create World safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Create World?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Create World free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Create World?

frank890417 (@frank890417) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Create World by frank890417 — Agent Skill

Name: Create World
Author: frank890417

Create World

Scaffold a new open-audiovisual work (a World + example page) from a natural-language description of the performance. Use when the user asks to "create a world", "make a new work/piece/visual", or describes an audiovisual idea they want performable (e.g. "a jellyfish world that blooms on consonant chords").

frank890417 72e7140 2.3 KB Updated Sep 17, 2026

File contents

create-world

Turn a described performance idea into a runnable example. The chassis (MIDI, mapping, timeline, console, monitor) already exists — you only write the World.

Steps

Extract the performance design from the description (ask only if truly ambiguous):
- What inputs matter? (chords → @openav/chord, body → @openav/pose, loudness/onsets → @openav/audio, knobs/keys → @openav/midi)
- What are the performable params? (3–6 is right; each must audibly/visibly matter)
- What events trigger discrete reactions? (note-on, chord/event, audio/onset)
- Does it have acts? → scenes + automation keyframes.
Scaffold: copy examples/01-hello-particles/ → examples/<nn>-<slug>/ (next free number). For chord-driven works, 02-chord-garden is the better donor; for camera-driven, 03-pose-field.
Write the World per docs/writing-a-world.md:
- continuous control ONLY via params (the mapper/timeline drive them)
- discrete reactions via signals.on(...) in init
- dispose() must release everything (worlds get hot-swapped)
- renderer: default to the built-in 2D createCanvas; use three.js via CDN import map only when 3D is essential.
Wire the assembly in main.js: params → timeline (scenes with performer notes — they show in the teleprompter) → mapper routes as data (declared defaults; the performer can re-learn live) → console + monitor lines verbatim from the donor example.
Verify before handing back:
- node serve.js → the page loads with zero console errors
- QWERTY fallback produces visible reaction (never require hardware for a demo)
- Space plays the timeline and something evolves over time
- npm test still green if you touched any package logic (avoid — see AGENTS.md rule 2)
Hand back: URL path, the param list (what to map to knobs), the scene list (what the timeline does), and which signals it reacts to.

---name: create-worlddescription: Scaffold a new open-audiovisual work (a World + example page) from a natural-language description of the performance. Use when the user asks to "create a world", "make a new work/piece/visual", or describes an audiovisual idea they want performable (e.g. "a jellyfish world that blooms on consonant chords").---# create-worldTurn a described performance idea into a runnable example. The chassis (MIDI,mapping, timeline, console, monitor) already exists — you only write the World.## Steps1. **Extract the performance design from the description** (ask only if truly ambiguous): - What *inputs* matter? (chords → `@openav/chord`, body → `@openav/pose`, loudness/onsets → `@openav/audio`, knobs/keys → `@openav/midi`) - What are the *performable params*? (3–6 is right; each must audibly/visibly matter) - What *events* trigger discrete reactions? (note-on, chord/event, audio/onset) - Does it have *acts*? → scenes + automation keyframes.2. **Scaffold**: copy `examples/01-hello-particles/` → `examples/<nn>-<slug>/` (next free number). For chord-driven works, `02-chord-garden` is the better donor; for camera-driven, `03-pose-field`.3. **Write the World** per [docs/writing-a-world.md](../../../docs/writing-a-world.md): - continuous control ONLY via params (the mapper/timeline drive them) - discrete reactions via `signals.on(...)` in `init` - `dispose()` must release everything (worlds get hot-swapped) - renderer: default to the built-in 2D `createCanvas`; use three.js via CDN import map only when 3D is essential.4. **Wire the assembly** in `main.js`: params → timeline (scenes with performer notes — they show in the teleprompter) → mapper routes as data (declared defaults; the performer can re-learn live) → console + monitor lines verbatim from the donor example.5. **Verify before handing back**: - `node serve.js` → the page loads with zero console errors - QWERTY fallback produces visible reaction (never require hardware for a demo) - Space plays the timeline and something evolves over time - `npm test` still green if you touched any package logic (avoid — see AGENTS.md rule 2)6. **Hand back**: URL path, the param list (what to map to knobs), the scene list (what the timeline does), and which signals it reacts to.

Frequently asked questions