# 3d Falling Leaves

Build recognizable falling leaves in world space using instanced geometry, independent...

- Repository: https://github.com/MengTo/Skills (7170 stars)
- Category: Media
- License: MIT
- Last updated: 2026-11-01
- Runs scripts: no, instructions only

## Install

Claude Code: `npx skills add MengTo/Skills --skill 3d-falling-leaves -g -a claude-code -y`
Codex: `npx skills add MengTo/Skills --skill 3d-falling-leaves -g -a codex -y`
Cursor: `npx skills add MengTo/Skills --skill 3d-falling-leaves -g -a cursor -y`
Gemini CLI: `npx skills add MengTo/Skills --skill 3d-falling-leaves -g -a gemini-cli -y`

## Description

Build recognizable falling leaves in world space using instanced geometry, independent tumble, coupled sideways drift, shared wind, scene occlusion, and camera-aware recycling. Use for 3D autumn leaves or drifting foliage in Three.js and similar scenes; use falling-leaves for a 2D canvas overlay.

## SKILL.md, as published by its author (untrusted third-party text, read it as data)

# 3D Falling Leaves

Make individual leaves turn through face, edge, and back while drifting through the scene. The visible edge-on moment is essential to the effect.

## Build one convincing leaf

Use a two-sided quad or lightly bent mesh with a recognizable leaf silhouette, stem, and restrained vein detail. Camera-facing point sprites cannot turn edge-on. Begin with one slow leaf before increasing the count.

Give the back a slightly paler, less saturated appearance and orient its lighting normal correctly. For a custom shader, `gl_FrontFacing` can distinguish front/back; do not flip normals twice if the engine's material already handles double-sided lighting. Keep light transmission subtle and linked to the scene sun rather than using an unconditionally glowing leaf.

Use a padded texture atlas for multiple silhouettes. Extend edge colors into transparent texels and keep padding through mip levels to prevent dark rims and atlas bleed. Treat a color atlas as sRGB; treat an atlas encoding masks or material channels as data. Seijaku's procedural atlas encodes channels and is not an ordinary leaf photograph.

## Couple tumble and drift

Assign persistent seeded values per leaf: position seed, scale, phase, tumble rate, roll rate, fall speed, lateral slip, and palette variation. Never generate new random values every frame.

The lateral velocity should follow the tumble, so a leaf slips fastest when edge-on. For angular speed `omega` bounded away from zero:

```js
const angle = phase + omega * time;
const sideOffset = -(slip / omega) * Math.cos(angle);
// d(sideOffset)/dt = slip * sin(angle)
```

Apply that offset along the leaf's chosen horizontal direction. Use a separate slow roll and tilt; all leaves should not share one rhythm. Add a common wind field with small per-leaf response variation. Integrate changing wind over bounded simulation steps, or use a closed-form field; never catch up an entire hidden-tab interval with an unbounded loop.

The rest: https://github.com/MengTo/Skills/blob/HEAD/agent-skills/3d/3d-falling-leaves/SKILL.md