hand-drawn-styles skill
A skill that assembles ready image prompts in hand-drawn styles from a set of tested recipes: kids drawing, minimal line, crayon, Ghibli look and more
- 2,665 ★ on the repository
- Last commit
- MIT
Minimal access
Why this level
- The skill only builds prompt text and takes no action in external systems
Install
- Works in
Manual install
git clone https://github.com/threerocks/hand-drawn-styles.git ~/.claude/skills/hand-drawnSKILL.md is picked up automatically by triggers.
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- The skill only builds prompt text and takes no action in external systems
This is third-party code. Review the repository files before installing.
What it does
The skill turns a scene description into a ready image-generation prompt using pre-built hand-drawn recipes. It picks a style by name or trigger, fills the content into a template, handles aspect ratio and returns the final prompt text. For a complex stable style it offers a multi-step process with intermediate corrections. The skill does not draw images itself; it prepares the text you hand to a generation model.
Who it is for. For designers, content teams and marketers who need consistent prompts for hand-drawn images.
Good fit when
- You need a prompt for an image in a specific hand-drawn style
- You need a series of illustrations in one recognizable style
- You need to quickly pick a style for a kids or family format
Not a fit when
- You need photorealism or strict technical graphics
- You have no access to an image-generation model
Example request
Assemble a prompt for a family illustration in the crayon style, 3 to 4 ratioLimitations
The skill outputs prompt text only; generation is done by an external model. The instructions and some styles are in Chinese, and the recipes cover a specific set of tested styles. The stable style needs the full multi-step process, otherwise the result is treated as intermediate.
How to disable. Delete the skill folder from ~/.claude/skills or your agent's skills directory.
FAQ
Does the skill generate images itself?
No, it only prepares the prompt. Generation is done by an external model you pass the text to.
Which agents does it work with?
Any agent that reads custom instructions: Claude Code, Cursor, Codex CLI, Gemini CLI, Cline, Windsurf and others.
Discussion
No questions or reviews yet.
Related
SkillsGPT-Image2 Style Library
gpt-image-2-style-library
A skill picks a matching template and style from a library of GPT-Image2 cases and assembles it into a finished prompt
A set of skills for Claude Code and Codex that assemble detailed image prompts from an element library and composition schemes
A skill that turns a picture into a generator prompt: frame breakdown, an enhanced version, three variants and platform editions.
A Claude Code skill that compiles a vague request into a finished gpt-image-2 prompt and checks it with a validator script
Fifty-one Codex skills that turn a photo into a poster, avatar, comic or illustration in a chosen visual style.
Five image-generation skills: Qwen-Image prompts, logos with shared geometry, diary comics, transparent PNGs and process dioramas
README in short
The README describes a set of tool-agnostic hand-drawn prompt recipes. The user picks a style, fills in the content, and the skill returns a ready prompt for a generation model. It ships with tested built-in styles and a separate multi-step process for the stable style. The skill installs by cloning the repository into the skills directory and fits agents that read custom instructions.
README badge
Original SKILL.md
--- name: hand-drawn description: Use when users ask for a hand-drawn or illustrated image prompt, name one of the repository's numbered styles or the 3.1 stable variant, or mention triggers. Produces prompt text or a formal reference-bearing call bundle and does not generate images. --- # hand-drawn style prompt generator 1. Follow the 5-step flow in PROTOCOL.md: pick style, take recipe, fill placeholders, handle ratio, output prompt. 2. Take the full template by number from STYLES.md. 3. When scripts are available prefer scripts/render_prompt.py, without manual abbreviation or synonym rewriting. 4. The stable style 3.1 for production requires the full three-stage workflow. 5. Other styles output the final prompt only; the skill does not generate images.