Claude Skill MartBrowse skillsWhat is a Skill?
 
← Back to list

Baoyu Image Gen

One CLI-backed skill that generates images through 10+ providers — OpenAI, Google, DashScope, MiniMax, Replicate and more.

Image & VideoIntermediate★ 24,749⑂ 2,764AI score 8/10Last updated: Jul 4, 2026

What it does

Ask Claude to "generate/create/draw an image" and this skill takes over. It wraps a dozen official image APIs — OpenAI GPT Image 2, Azure OpenAI, Google, OpenRouter, DashScope (Qwen-Image), Z.AI GLM-Image, MiniMax, Jimeng, Seedream, Replicate, Agnes — behind a single TypeScript CLI (scripts/main.ts).

Highlights:

  • Text-to-image plus reference-image edits (--ref)
  • Aspect ratios (--ar 16:9), explicit sizes, normal/2k quality presets
  • Parallel batch generation from saved prompt files (--batchfile, --jobs)
  • Persistent preferences in EXTEND.md (default provider, model, quality, save path)
  • Up to 3 automatic retries per image with per-image failure reporting
  • A dedicated codex-cli provider that uses your Codex login instead of an OpenAI API key

The identity-preservation guidance for reference images is unusually detailed and helps avoid the classic failure where the model invents a look-alike stranger.

Who it's for

  • Power users who switch between multiple image APIs depending on the task
  • Content creators who need batches of article/newsletter illustrations
  • Anyone wanting to compare Chinese models (Qwen-Image, Jimeng, Seedream, GLM-Image) alongside Western ones
  • Teams with an existing prompt list who want throughput instead of one-off calls

Examples

  1. Single image: "Draw a snowy mountain landscape, 16:9, 2K quality" → main.ts --prompt ... --ar 16:9 --quality 2k
  2. Reference edit: pass 2–3 portrait photos and say "same person, just change the scene to a Tokyo street at night" — the skill applies its identity-preservation prompt rules
  3. Article illustration batch: turn outline.md + a prompts/ folder into batch.json via build-batch.ts, then run --jobs 4 to render 10 illustrations in parallel

· · · Install guide · · ·

Install in the Claude app (no terminal)
  1. Download the ZIP with the button below.
  2. In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
  3. Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
↓ Download ZIP
Install in Claude Code

Let Claude do it — paste this into Claude Code

Install the skill I found on Claude Skill Mart.
Copy the skills/baoyu-image-gen folder from the GitHub repo JimLiu/baoyu-skills into my ~/.claude/skills/baoyu-image-gen/.
When it's done, tell me in one line what this skill can do.

Install with a command instead

git clone https://github.com/JimLiu/baoyu-skills.git /tmp/baoyu-skills && mkdir -p ~/.claude/skills && cp -r /tmp/baoyu-skills/skills/baoyu-image-gen ~/.claude/skills/

⚠ This is a third-party skill. Check the source repository before installing.

  1. Get a runtime: run bun --version. If missing, install with brew install oven-sh/bun/bun (macOS) or follow https://bun.sh. npx -y bun also works.
  2. Clone the repo: git clone https://github.com/JimLiu/baoyu-skills.git /tmp/baoyu-skills
  3. Install the skill: mkdir -p ~/.claude/skills && cp -r /tmp/baoyu-skills/skills/baoyu-image-gen ~/.claude/skills/
  4. Set API keys for whichever provider you plan to use, e.g. export OPENAI_API_KEY=sk-... or export GOOGLE_API_KEY=.... You can also store them in ~/.baoyu-skills/.env.
  5. Restart Claude Code, then try: "generate an image of a cat".
  6. Complete first-time setup: the skill asks for default provider, model, quality and output folder, then writes ~/.baoyu-skills/baoyu-image-gen/EXTEND.md. Generation is blocked until this file exists.
  7. (Optional) For batches, save your prompts as files, build batch.json with scripts/build-batch.ts, and run the CLI with --batchfile batch.json --jobs 4.
View source on GitHub ↗License: MIT