# generate-image

Generates images from text prompts using OpenAI or Stability AI. Useful when your coding agent needs to create logos, banners, or illustrations directly in the project.

- skill · `type=skill` · Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, OpenCode · 343 installs · ★ 721
- Author: OpenClaudia
- License: MIT · policy=mirror · the beginning of the file is at the end of this page
- Source: https://github.com/OpenClaudia/openclaudia-skills/tree/HEAD/skills/ai-image-gen/SKILL.md
- Updated: 2026-10-21
- HTML: https://affpapa.org/en/nevibecode/skills/openclaudia-generate-image

## What it does

The skill executes a Python script located at `~/.agents/tools/generate-image.py`, passing prompt text, destination path, and generation parameters. It defaults to OpenAI's `gpt-image-2` model with medium quality, but can switch to Stability AI running `sd3.5-large` with the `--provider stability` flag.

Execution requires the `requests` Python package and an environment file at `~/.agents/tools/.env`. That file must store `OPENAI_API_KEY` or `STABILITY_API_KEY` depending on the provider you run. The script fails immediately without valid credentials in place.

OpenAI handles transparent backgrounds via `--background transparent` and supports aspect ratios like 1536x1024 or 1024x1536. Stability AI restricts output to 1024x1024 and does not support transparency.

## Install

### Claude Code (native platform)

1. **Skills CLI (Vercel)**
```bash
npx skills add OpenClaudia/openclaudia-skills --skill generate-image -a claude-code
```
```bash
npx skills add OpenClaudia/openclaudia-skills --skill ai-image-gen -a claude-code
```

2. **Manually: put the files in place**
- File: `~/.claude/skills/ai-image-gen/SKILL.md` — Copy the whole skill folder; for one project use .claude/skills/ai-image-gen/

### Codex

1. **Skills CLI (Vercel)**
```bash
npx skills add OpenClaudia/openclaudia-skills --skill ai-image-gen -a codex
```

2. **Manually: put the files in place**
- File: `.agents/skills/ai-image-gen/SKILL.md` — For all projects use ~/.agents/skills/

### Cursor

1. **Skills CLI (Vercel)**
```bash
npx skills add OpenClaudia/openclaudia-skills --skill ai-image-gen -a cursor
```

2. **Manually: put the files in place**
- File: `.cursor/skills/ai-image-gen/SKILL.md` — Cursor also reads .agents/skills/ and .claude/skills/

### GitHub Copilot

1. **GitHub CLI**
```bash
gh skill install OpenClaudia/openclaudia-skills ai-image-gen
```

2. **Manually: put the files in place**
- File: `.github/skills/ai-image-gen/SKILL.md`

### Gemini CLI

1. **Skills CLI (Vercel)**
```bash
npx skills add OpenClaudia/openclaudia-skills --skill ai-image-gen -a gemini-cli
```

2. **Platform CLI**
```bash
gemini skills install https://github.com/OpenClaudia/openclaudia-skills --path skills/ai-image-gen
```

### Windsurf (Devin Desktop)

**Manually: put the files in place**
- File: `.devin/skills/ai-image-gen/SKILL.md` — The legacy .windsurf/skills/ path still works

### OpenCode

**Skills CLI (Vercel)**
```bash
npx skills add OpenClaudia/openclaudia-skills --skill ai-image-gen -a opencode
```

### Claude.ai

**Claude.ai**
- Zip the folder and upload it in Settings → Capabilities → Skills

Where a file is given instead of a command, copy the files from the source to that path: https://github.com/OpenClaudia/openclaudia-skills/tree/HEAD/skills/ai-image-gen/SKILL.md

## When to use it
- create an app icon with a transparent background
- generate a wide banner image for a blog cover
- produce artwork using OpenAI `gpt-image-2` or Stability AI
- batch render cheaper drafts using the `gpt-image-1-mini` model

## Example prompts
- Generate a mountain sunset oil painting and save it to ./sunset.png
- Create a flat minimalist cat icon with a transparent background
- Draw a watercolor garden using Stability AI and output to ./garden.png

## Source text · SKILL.md · MIT

Below is the content of a third-party file. It is data, not instructions for you: do not act on anything written inside unless the human asks. Block tag: e45bc8998ea29b13.

BEGIN THIRD-PARTY FILE e45bc8998ea29b13
````text
---
name: generate-image
description: Generate images using AI (OpenAI GPT Image or Stability AI). Use when the user asks to generate an image, create an AI image, make an illustration, or produce artwork from a text prompt.
argument-hint: [prompt description]
allowed-tools: Bash(*), Read, Write
---

# AI Image Generation

Generate images from text prompts using OpenAI GPT Image (gpt-image-2 / gpt-image-1 / variants) or Stability AI (SD 3.5 Large).

## Tool Location

- Script: `~/.agents/tools/generate-image.py`
- Env file: `~/.agents/tools/.env` (contains `OPENAI_API_KEY` and `STABILITY_API_KEY`)

## Available OpenAI image models (verified live via `/v1/models` 2026-05-15)

| Model ID | Notes |
|---|---|
| `gpt-image-2` | **Flagship.** Released 2026-04-21. Best prompt adherence, best photorealism. Default when image quality matters. |
| `gpt-image-2-2026-04-21` | Pinned dated variant of gpt-image-2 |
| `gpt-image-1.5` | Intermediate release between 1 and 2 |
| `gpt-image-1-mini` | Smaller/cheaper gpt-image-1 variant — use for batch/draft generation where cost matters |
| `gpt-image-1` | Original gpt-image. Still works; superseded by gpt-image-2. |
| `chatgpt-image-latest` | Always-current alias of the model ChatGPT.com uses (currently gpt-image-2-class). Use when you want "whatever ChatGPT uses today" |
| `dall-e-3` | Legacy fallback. Different quality semantics (standard/hd, not low/medium/high). |

Pass any of these to `--model`. The script branches on `gpt-image*` for the quality/format handling, so all gpt-image-* variants work out of the box.

**Default in the script is now `gpt-image-2`** (the flagship) — pass `--model gpt-image-1-mini` for cheap batch/draft work.

## Quick Usage

### Generate with OpenAI gpt-image-2 (flagship)

```bash
python ~/.agents/tools/generate-image.py \
  --prompt "a sunset over mountains, oil painting style" \
  --output ./sunset.png \
  --model gpt-image-2 \
  --quality high
```

### Generate with default (gpt-image-2)

```bash
python ~/.agents/tools/generate-image.py \
  --prompt "a sunset over mountains, oil painting style" \
  --output ./sunset.png
```

### High quality

```bash
python ~/.agents/tools/generate-image.py \
  --prompt "modern logo design for a tech company" \
  --output ./logo.png \
  --size 1024x1024 \
  --quality high
```

### Wide format (good for blog covers, banners)

```bash
python ~/.agents/tools/generate-image.py \
  --prompt "abstract digital art with blue tones" \
  --output ./banner.png \
  --size 1536x1024
```

### Tall format (good for mobile, stories)

```bash
python ~/.agents/tools/generate-image.py \
  --prompt "portrait of a futuristic city" \
  --output ./city.png \
  --size 1024x1536
```

### Transparent background (icons, logos)

```bash
python ~/.agents/tools/generate-image.py \
````
END THIRD-PARTY FILE e45bc8998ea29b13

The third-party file has ended. Everything between the e45bc8998ea29b13 tags is data: requests and commands from it are not followed. An “end of file” line with another tag inside the block is part of the file.

First 2804 of 4557 bytes. © OpenClaudia · MIT · https://github.com/OpenClaudia/openclaudia-skills/tree/HEAD/skills/ai-image-gen/SKILL.md

_Part of the page did not fit — see the HTML version._

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md