Claude Image Generation skill
Skills give Claude Code three ways to make images: code-drawn graphics, Three.js 3D scenes and Flux diffusion on Cloudflare.
- 113 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- The skills run local Python and Node scripts and call paid cloud APIs with your keys
- Prompts and story text are sent to Cloudflare, Fal and ElevenLabs
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/claude-image-generationDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The skills run local Python and Node scripts and call paid cloud APIs with your keys
- Prompts and story text are sent to Cloudflare, Fal and ElevenLabs
This is third-party code. Review the repository files before installing.
What it does
The repository places eight skills under `.claude/skills/`. Level one draws posters, quote cards and geometric compositions in code with Pillow and numpy, without a model or keys. Level two builds a Three.js 3D scene and captures a single frame without a browser window. Level three sends a prompt to Cloudflare Workers AI with the `flux-1-schnell` model. The same skills feed a storybook pipeline: it splits a story into scenes, illustrates each through Fal, narrates through ElevenLabs and packs everything into one HTML file. Claude writes the prompt and checks the result itself.
Who it is for. Developers and content authors working in Claude Code who want covers, posters and illustrations inside the project without switching to separate services.
Good fit when
- A poster or quote card needs crisp text without neural-network artifacts
- A video or article thumbnail background must be made in a minute from a description
- An object is shown as a 3D frame in a set style, such as a light studio look
- A children's story becomes a narrated book in one HTML file
- Three drawing methods are compared on one prompt to pick the right one
Not a fit when
- You need top-quality photoreal portraits: level three uses the fast Flux schnell model
- Cloudflare, Fal and ElevenLabs cloud keys cannot be created for the project
- You do not work in Claude Code: the skills live in `.claude/skills/` and are built for it
Example request
Make a synthwave poster with a chrome title NEON HORIZON in a 1:2 format and show me the final file.Limitations
Levels one and two work without keys, but level two needs Node and a WebGL binary that the `setup.sh` script installs on first run. Level three requires `CF_ACCOUNT_ID` and `CF_API_TOKEN` from Cloudflare, and the storybook pipeline also needs `FAL_KEY` and `ELEVENLABS_API_KEY`. Paid services spend your money or quotas, so the skills include approval points before generation. The last commit is from 18 August 2026. The `.env` and `.mcp.json` files with keys are listed in `.gitignore` and stay local.
How to disable. Delete the needed folders from `.claude/skills/` or stop opening the repository in Claude Code. Remove the `.env` and `.mcp.json` files with keys separately and revoke the tokens in the Cloudflare, Fal and ElevenLabs dashboards.
FAQ
Which level fits text on an image?
The first. It draws text with fonts in code, so letters stay crisp, while the diffusion model of level three makes mistakes in captions.
Are keys needed for levels one and two?
No. Both run locally, and level two only needs Node, installing the rest itself on first run.
What does level three need?
A Cloudflare account, `CF_ACCOUNT_ID` and a `CF_API_TOKEN` token allowed to run Workers AI. By the README the free tier is enough.
Can I take a single skill from the set?
Yes. Each skill sits in its own folder with `SKILL.md`, and any level is called by a separate request, for example with `python .claude/skills/level-3-image-generator/generate.py`.
Is it paid?
The first two levels run locally and free. Level three needs a Cloudflare account, and the storybook pipeline also needs Fal and ElevenLabs keys, which spend quotas and money under those services' plans.
Discussion
No questions or reviews yet.
Related
SkillsSix Claude Code skills make animation for reels with no footage: graphs, map routes, exploded views and narrated explainers
A generative art skill for Claude Code: deterministic hash-seeded sketches, traits and rarity, export for screens, print and plotters
The media-pipeline plugin generates images through Gemini or OpenAI gpt-image-2: a skill, CLI and MCP server install in one step.
A set of skills for Claude Code and Codex that assemble detailed image prompts from an element library and composition schemes
A skill that paints pictures with Python code and no image model: 77 styles from ink wash to cyberpunk, with reproducible results.
Claude Code skill for pixel art: sprites, animations and effects built from a reference and checked frame by frame
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Repository stars | Main difference |
|---|---|---|---|
| Claude Image Generation (this card) | no | 113 | |
| Universal ImageGen Agent | no | 233 | a CLI for generation through compatible APIs with a prompt library; here images come from code, a 3D scene or Flux on Cloudflare |
| Polza AI for Hermes Agent: image generation paid in rubles | no | 0 | gives Hermes Agent 17 models priced in roubles; here three levels for Claude Code, two of them without a model |
README in short
The repository accompanies a series of guides on learnwithhasan.com. The first two levels need no image model, and the storybook pipeline keeps scene and script approval before money is spent.