Skill

image-generation

This skill should be used when the user asks to "generate an image", "create an image", "edit an image", "modify an image", "make a picture", "create an icon", "design a logo", "make a mockup", "create a banner", "generate a thumbnail", "make an illustration", "create a transparent PNG", or discusses image generation, DALL-E, GPT-Image, visual assets, or image editing using the imagegen MCP tools.

From imagegen

Install

Run in your terminal

npx claudepluginhub thorrak/imagegen-plugin --plugin imagegen

Tool Access

This skill uses the workspace's default tool permissions.

Supporting Assets

View in Repository

examples/image-editing-example.md

examples/text-to-image-example.md

references/parameters.md

Skill Content

Similar Skills

cache-components

Guides Next.js Cache Components and Partial Prerendering (PPR) with cacheComponents enabled. Implements 'use cache', cacheLife(), cacheTag(), revalidateTag(), static/dynamic optimization, and cache debugging.

cache-components

138.6k

claude-opus-4-5-migration

2 files

Migrates code, prompts, and API calls from Claude Sonnet 4.0/4.5 or Opus 4.1 to Opus 4.5, updating model strings on Anthropic, AWS, GCP, Azure platforms.

claude-opus-4-5-migration

83.2k

evaluation-methodology

1 file

Details PluginEval's skill quality evaluation: 3 layers (static, LLM judge), 10 dimensions, rubrics, formulas, anti-patterns, badges. Use to interpret scores, improve triggering, calibrate thresholds.

plugin-eval

32.9k

Stats

Stars0

Forks0

Last CommitFeb 13, 2026

Actions

View Source View Plugin View on GitHub View README

Image Generation with imagegen MCP

Available Tools

Use the two MCP tools provided by the imagegen server:

text-to-image

Generate new images from text prompts.

Required parameters:

text: The generation prompt
outputPath: Absolute path where the image file should be saved

Optional parameters:

model: gpt-image-1 (default), dall-e-2, or dall-e-3
size: Dimensions (default 1024x1024)
quality: auto (default), standard, or hd
output_format: png (default), jpeg, or webp
style: vivid (default) or natural (DALL-E 3 only)
background: auto (default), transparent, or opaque
n: Number of images (default 1)
moderation: low (default) or auto
output_compression: 0-100 (default 100)

image-to-image

Edit existing images using text instructions.

Required parameters:

images: Array of absolute file paths to source images
prompt: Description of the desired edits
outputPath: Absolute path where the edited image should be saved

Optional parameters:

mask: Path to a PNG mask (transparent areas indicate where to edit)
model: gpt-image-1 (default) or dall-e-2
size: Dimensions (default 1024x1024)
quality: auto (default) or standard
output_format: png (default), jpeg, or webp
n: Number of variations (default 1)
output_compression: 0-100 (default 100)

Output Path Handling

Always save generated images to a generated-images/ directory in the current project:

Construct the output path as an absolute path: <project-root>/generated-images/<descriptive-filename>.<format>
Create the generated-images/ directory with mkdir -p before calling the tool
Use descriptive filenames based on what was generated (e.g., hero-banner.png, app-icon-512.png)
Never use UUIDs or random strings as filenames
If generating multiple images, use numbered suffixes: icon-1.png, icon-2.png

After generation, confirm the file path to the user so they can find the image.

Model Selection Guide

Choose the model based on the use case:

gpt-image-1 (recommended default):

Best overall quality and prompt adherence
Supports transparent backgrounds (with PNG/WebP)
Supports image editing with masks
Use for: most use cases, professional assets, icons, UI elements

dall-e-3:

Strong artistic interpretation
Supports style parameter (vivid/natural) and hd quality
Limited to n=1 per request, text-to-image only (no editing)
Use for: artistic/creative images, illustrations

dall-e-2:

Fastest generation, lower cost, lower quality
Supports image editing
Use for: quick drafts, small thumbnails, rapid iteration

Size Constraints by Model

Model	Allowed Sizes
gpt-image-1	1024x1024, 1536x1024, 1024x1536, auto
dall-e-3	1024x1024, 1792x1024, 1024x1792
dall-e-2	256x256, 512x512, 1024x1024

Always verify the requested size is valid for the chosen model. If the user requests a size not supported by their chosen model, suggest the closest available size.

Prompt Best Practices

When crafting prompts for image generation:

Be specific and descriptive about the subject, style, composition, and mood
Include details about lighting, color palette, and perspective when relevant
Specify the art style (photorealistic, flat design, watercolor, pixel art, etc.)
For UI assets, mention the intended use (icon, button, banner, background)
For transparent backgrounds, set background: "transparent" and use PNG or WebP format

Transparent Background Images

To generate images with transparent backgrounds:

Set background to "transparent"
Use output_format of "png" or "webp" (JPEG does not support transparency)
Only supported by gpt-image-1
Ideal for icons, logos, UI elements, and overlays

Image Editing Workflow

When editing images with image-to-image:

Verify the source image exists and note its path
Describe the desired changes clearly in the prompt
Use a mask if only specific regions should be edited (transparent areas = edit zones)
Only gpt-image-1 and dall-e-2 support editing
The source image should be a supported format (PNG recommended)

Error Handling

If the MCP server is not running, instruct the user to verify their OPENAI_API_KEY environment variable is set
If a model or size is not supported, suggest valid alternatives
If generation fails due to content policy, suggest rephrasing the prompt
If the output directory cannot be created, ask the user about an alternative location

Quick Access Commands

For quick access, users can use the /generate-image and /edit-image slash commands instead of describing the task conversationally.

Parameter Reference

For edge cases involving parameter validation or uncommon parameter combinations, consult the full parameter reference at ${CLAUDE_PLUGIN_ROOT}/skills/image-generation/references/parameters.md.

Examples

For complete tool call examples, see ${CLAUDE_PLUGIN_ROOT}/skills/image-generation/examples/.