@komereji12 shares how to batch-process local images in Claude Code with jezweb design-assets' image-processing skill: resize, convert to WebP, trim white edges
Xiaohongshu posts and web work often need one width, WebP, and logo whitespace trimmed. Overseas repo jezweb/claude-skills (MIT, ~1.1k★) ships design-assets → image-processing: it doesn't generate images, only processes local files; `bin/img-process` (Pillow) covers resize/convert/trim/thumbnail/optimise/og-card/batch. Chinese poster @komereji12 shared the install lines and natural-language triggers. We cloned the repo and ran img-process with Pillow 13: JPG→WebP ~47KB→5KB; transparent logo trim 400×300→261×201; opaque white-canvas "whitespace" did not shrink—trim needs an alpha channel, not visual white. We did not run `/plugin install` inside a live Claude Code session.
Visit site image-processing (jezweb/claude-skills design-assets) ↗- Pricing
- Free
- Region
- Global
- Where it's from
- Overseas
- Added
- Nov 4, 2026
- Tested on
- Nov 4, 2026
- Version we used
- jezweb/claude-skills (cloned 2026-11-04); img-process CLI in repo bin/; Pillow 13.0.0
- Original author
- 折腾笔记 @komereji12 ↗
- Our test
- Partly tested
- Cost / time
- Skill and CLI are free/open source; you need Claude Code (or another SKILL-capable agent) plus local Pillow.
The task
Resize, convert, trim transparent padding or compress a local batch of photos/logos in Claude Code via a skill—without opening Photoshop.
Steps
- In Claude Code: `/plugin marketplace add jezweb/claude-skills`, then `/plugin install design-assets@jezweb-skills`.
- Install Pillow: `pip install Pillow` (required by the skill and `img-process`; macOS `sips` / Node `sharp` / ffmpeg are partial alternatives).
- Optional: put repo `bin/img-process` on PATH so the agent calls the CLI instead of regenerating scripts.
- Trigger in natural language, e.g. "convert everything in ./photos to webp at width 1080 into ./out" or "trim this logo and leave 10px padding".
- Trial one or two files before a big batch; paths must be local—the skill does not fetch remote CDNs.
Prompt
Convert every photo in ./photos to webp at width 1080 into ./out; process one sample first and wait for my OK before the full batch.
What the author says
@komereji12: design-assets' image-processing only edits existing files—it doesn't generate images. Resize, trim logo padding, PNG/JPG/WebP convert, compress, batch thumbnails. Install the marketplace + design-assets plugin, then a Python image lib; natural language triggers it. MIT, ~1.1k★; get local paths right and trial a couple files first.
Our test
Cloned jezweb/claude-skills; read image-processing SKILL.md and bin/img-process. With Pillow 13.0.0 ran trim/convert/batch on synthetic samples. Did not run Claude Code `/plugin` flow.
What we found when we tried it:photo.jpg 2000×1500 ~47.6KB → photo.webp same size ~5.4KB. Opaque white logo-raw stayed 400×300 after trim; transparent logo-alpha trim+padding10 → 261×201. batch wrote both webps successfully.
How to get in
GitHub jezweb/claude-skills; needs the Claude Code plugin marketplace. Share post in Chinese; skill docs in English.
Who it suits
People who batch social/web assets; developers already in Claude Code who don't want another image app.
Pitfalls
- SKILL marks compatibility: claude-code-only; other agents must copy SKILL.md or call img-process themselves.
- Our run: trim did not shrink an opaque white canvas (400×300 stayed); a transparent-bg logo did (400×300→261×201 with padding). Make padding transparent—or cut out—before trim.
- batch convert without --width only changes format; say the width or pass the flag.
- We validated the shipped CLI only—not a live Claude Code `/plugin` install and auto-trigger.
More in this category
- Turn a selfie into a transparent animated sticker setImageSeedream 5.5JimengHuashu skillSkill is free; image gen uses Jimeng credits◐ Partly testedChina
- Two minutes to give agents GPT stills and GIFsImageopenai-imageMCPGrok BotMCP free; images use OpenAI credits◐ Partly testedOverseas
- One photo plus a minimal paper-print twinImageCodexSkillImage genSkill free; needs Codex image generation◐ Partly testedChina
- Paint one dieline, fold a low-poly GLB with a Fold clipImageOrigamiClaude CodeBlenderSkill free; paint ~$0.10–0.25 each◐ Partly testedOverseas
- Pixel VFX as code—plus a Claude refine loopImagepixelfxClaude Codepixel VFXLib+skill free; new FX need Claude◐ Partly testedOverseas