# see-through-anime-layer-decomposition

Splits a single anime illustration into semantic layers and exports a layered PSD with depth maps for 2.5D animation.

- skill · `type=skill` · Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, OpenCode · 759 installs · ★ 88
- Author: reason-machines
- License: not stated · policy=link-only · description only; the file stays with the author
- Source: https://github.com/reason-machines/trending-skills/tree/HEAD/skills/see-through-anime-layer-decomposition/SKILL.md
- Updated: 2026-10-09
- HTML: https://affpapa.org/en/nevibecode/skills/see-through-anime-layer-decomposition

## What it does

This skill hands the agent working knowledge of See-through: it takes one anime character image and separates it into layers — hair, face, eyes, clothing, accessories. Occluded regions get inpainted so each layer is complete, and drawing order is inferred from pseudo-depth.

The pipeline runs on diffusion models. The main script `inference/scripts/inference_psd.py` takes a file path or a folder plus the `--save_to_psd` flag, and writes results to `workspace/layerdiff_output/`. After that, `heuristic_partseg.py` can split layers further: `seg_wdepth` by depth, `seg_wlr` left to right. Models are pulled from HuggingFace on first use, so there is no separate download step.

What you need up front: a conda environment with Python 3.12, PyTorch built for CUDA 12.8, and the packages in `requirements.txt`. Annotators such as detectron2, SAM2 or mmdet come from their own requirements files — install only the ones you actually use. Run every script from the repository root, or asset paths will not resolve.

## Install

### Claude Code (native platform)

1. **Skills CLI (Vercel)**
```bash
npx skills add reason-machines/trending-skills --skill see-through-anime-layer-decomposition -a claude-code
```

2. **Manually: put the files in place**
- File: `~/.claude/skills/see-through-anime-layer-decomposition/SKILL.md` — Copy the whole skill folder; for one project use .claude/skills/see-through-anime-layer-decomposition/

### Codex

1. **Skills CLI (Vercel)**
```bash
npx skills add reason-machines/trending-skills --skill see-through-anime-layer-decomposition -a codex
```

2. **Manually: put the files in place**
- File: `.agents/skills/see-through-anime-layer-decomposition/SKILL.md` — For all projects use ~/.agents/skills/

### Cursor

1. **Skills CLI (Vercel)**
```bash
npx skills add reason-machines/trending-skills --skill see-through-anime-layer-decomposition -a cursor
```

2. **Manually: put the files in place**
- File: `.cursor/skills/see-through-anime-layer-decomposition/SKILL.md` — Cursor also reads .agents/skills/ and .claude/skills/

### GitHub Copilot

1. **GitHub CLI**
```bash
gh skill install reason-machines/trending-skills see-through-anime-layer-decomposition
```

2. **Manually: put the files in place**
- File: `.github/skills/see-through-anime-layer-decomposition/SKILL.md`

### Gemini CLI

1. **Skills CLI (Vercel)**
```bash
npx skills add reason-machines/trending-skills --skill see-through-anime-layer-decomposition -a gemini-cli
```

2. **Platform CLI**
```bash
gemini skills install https://github.com/reason-machines/trending-skills --path skills/see-through-anime-layer-decomposition
```

### Windsurf (Devin Desktop)

**Manually: put the files in place**
- File: `.devin/skills/see-through-anime-layer-decomposition/SKILL.md` — The legacy .windsurf/skills/ path still works

### OpenCode

**Skills CLI (Vercel)**
```bash
npx skills add reason-machines/trending-skills --skill see-through-anime-layer-decomposition -a opencode
```

### Claude.ai

**Claude.ai**
- Zip the folder and upload it in Settings → Capabilities → Skills

Where a file is given instead of a command, copy the files from the source to that path: https://github.com/reason-machines/trending-skills/tree/HEAD/skills/see-through-anime-layer-decomposition/SKILL.md

## When to use it
- split an anime character into separate layers for animation
- turn one image into a layered PSD file
- inpaint occluded body parts on each layer
- infer drawing order from pseudo-depth
- stratify layers by depth or left-right

## Example prompts
- decompose this anime illustration into layers and save a PSD
- run inference_psd.py over my image folder and export PSDs
- split the handwear layer by depth with heuristic_partseg.py

## Similar
- [anime-soccer](https://affpapa.org/en/nevibecode/skills/anime-soccer.md) — skill · 665 installs
- [Anime Garden](https://affpapa.org/en/nevibecode/mcp/anime-garden.md) — MCP server · ★ 1374
- [layer-3d](https://affpapa.org/en/nevibecode/skills/layer-3d.md) — skill · 3300 installs
- [fal-vision](https://affpapa.org/en/nevibecode/skills/fal-vision.md) — skill · 2700 installs
- [nv-generate-mr](https://affpapa.org/en/nevibecode/skills/nv-generate-mr.md) — skill · 1800 installs
- [nv-generate-mr-brain-finetune](https://affpapa.org/en/nevibecode/skills/nv-generate-mr-brain-finetune.md) — skill · 1800 installs

From the same set: [reason-machines/trending-skills](https://affpapa.org/en/nevibecode/by/reason-machines/trending-skills.md) — 49 cards

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md