Languages: English | 简体中文
Remove image backgrounds to transparent PNG, resize and compress app assets, and convert sprite sheets to aligned GIFs for Claude Code and Codex — generate the raw image with your favorite model, then turn it into a production-ready asset here.
This is a skill (not a hosted API). It does not redistribute Codex’s bundled imagegen skill. Cutout, resize, sprite-sheet split, and GIF encoding are deterministic image processing.
Maintained successor to image-assets-skill. The skill is still named image-assets and lives in image-assets/. The old repository is an archived landing page.
- Remove background → transparent PNG (solid-color keying or model-based cutout)
- Resize, compress, convert PNG / WebP / AVIF / JPEG
- Sprite sheet → aligned animated GIF (N-grid split, feet-anchored, ping-pong)
- App icon sets from one source image
Ask the assistant to process an image after install, or see the full commands.
python3 image-assets/scripts/image_assets_cli.py app-transparent \
--input ./photos --output ./cutoutsRequires Node.js (for the skills CLI) and Python 3 (for the image processing helpers):
npx skills add x0c/image-skills -gThis installs the image-assets skill across all detected coding agents. Start a new session (or restart your agent) to load it.
If you use agentsync, maintain its canonical skill source instead of replacing shared tool-side links.
If you run multiple AI coding agents daily, check out the rest of the local-first toolkit:
| Repository | Description |
|---|---|
| corral | Run Claude Code, Codex, OpenCode, Cursor & Pi side-by-side in a terminal session manager, then continue from iPhone or Mac. |
| subswap | Check remaining quota and switch Claude Code, Codex, Cursor, Kimi Code & OpenCode accounts without logging out. |
| agentsync | Sync AGENTS.md, CLAUDE.md, skills, and MCP configs across coding agents from one local source of truth. |
| agent-doc-skills | Navigable project knowledge base skills (doc-init, doc-update, doc-compact) so coding agents don't rediscover your repo every session. |
Solid-color frame processing needs Python; compression/conversion also needs Node.js; AI cutout models download their local weights on first run. RMBG-2.0 is free for non-commercial use only (commercial use needs a BRIA license); BiRefNet is free including commercial use. Never upload access tokens.
Note: The internal skill prompts (SKILL.md) are currently written in Chinese.
Report problems in this repository’s Issues, with redacted errors and a minimal example.
MIT; model licenses apply separately.