What does image-gen do?

image-gen generates and edits images using multiple AI providers via a CLI, handling provider selection, fallbacks, and saving results.

Which providers can I use?

You can use auto (default) or explicitly specify one of: openai, stability, bfl, ideogram, fal, gemini, replicate, clipdrop.

Where are outputs saved?

Outputs and a JSON summary are saved in the .image-gen directory in your project.

image-gen

Scanned

npx machina-cli add skill shipdeckai/claude-skills/image-gen --openclaw

Files (1)

SKILL.md

2.9 KB

Image Generation

Generate images via CLI tool image-gen which handles provider selection, fallbacks, and image saving.

Important: Run commands using node with the plugin's CLI path:

node ${CLAUDE_PLUGIN_ROOT}/dist/cli.bundle.cjs <command>

Generate an Image

node ${CLAUDE_PLUGIN_ROOT}/dist/cli.bundle.cjs generate --prompt "description of image" [--provider auto] [--width 1024] [--height 1024]

Output: JSON with file paths to saved images in .image-gen/ directory.

Edit an Image

node ${CLAUDE_PLUGIN_ROOT}/dist/cli.bundle.cjs edit --image /path/to/image.png --prompt "edit instructions" [--provider auto]

List Configured Providers

node ${CLAUDE_PLUGIN_ROOT}/dist/cli.bundle.cjs providers

Provider Selection

auto (default): Intelligent selection based on prompt content
Explicit: openai, stability, bfl, ideogram, fal, gemini, replicate, clipdrop

Quick guide:

Text/logos/typography → ideogram or recraft
Photorealism → bfl or stability
Fast iterations → fal
General purpose → openai
Image editing → openai, stability, bfl, gemini, clipdrop

For detailed provider capabilities, see references/providers.md.

Environment Setup

Add at least one API key to your shell profile (~/.zshrc or ~/.bashrc):

# Add to ~/.zshrc (or ~/.bashrc)
export OPENAI_API_KEY="sk-..."       # For DALL-E
export BFL_API_KEY="..."             # For FLUX (recommended for quality)
export IDEOGRAM_API_KEY="..."        # For text/logos (best typography)
export GEMINI_API_KEY="..."          # For Google Imagen
export STABILITY_API_KEY="..."       # For Stable Diffusion
export FAL_API_KEY="..."             # For fast iterations
export REPLICATE_API_TOKEN="..."     # For various models

After adding, run source ~/.zshrc or restart your terminal.

Example Workflows

Generate a logo:

node ${CLAUDE_PLUGIN_ROOT}/dist/cli.bundle.cjs generate --prompt "Modern minimalist logo for TechStartup with the text 'NOVA'" --provider ideogram --width 1024 --height 1024

Generate product shot:

node ${CLAUDE_PLUGIN_ROOT}/dist/cli.bundle.cjs generate --prompt "Professional product photography of a sleek smartphone on marble surface, soft lighting" --provider bfl

Edit an image:

node ${CLAUDE_PLUGIN_ROOT}/dist/cli.bundle.cjs edit --image ./photo.png --prompt "Remove the background and make it transparent" --provider clipdrop

Source

git clone https://github.com/shipdeckai/claude-skills/blob/main/plugins/image-gen/skills/image-gen/SKILL.mdView on GitHub

Overview

image-gen is a CLI tool that generates and edits images using multiple AI providers (OpenAI DALL-E, Stability AI, BFL FLUX, Ideogram, FAL, Gemini, Replicate, Clipdrop). It automatically selects providers, handles fallbacks, and saves outputs to the .image-gen directory. Use it for logos, product visuals, marketing graphics, or photo edits like background removal.

How This Skill Works

Run commands via node ${CLAUDE_PLUGIN_ROOT}/dist/cli.bundle.cjs with generate or edit. It automatically selects a provider by default (auto) or uses an explicit provider, and it saves the generated images or edits as files. The CLI outputs a JSON summary with file paths in the .image-gen directory.

When to Use It

You need a new logo or illustration
You want professional product shots or marketing visuals
You need to edit an image (retouch, crop, or background removal)
You require rapid visual concept iterations
You want to compare outputs across providers (auto vs explicit)

Quick Start

Step 1: Prepare your prompt and optional dimensions, using auto provider by default
Step 2: Run the command: node ${CLAUDE_PLUGIN_ROOT}/dist/cli.bundle.cjs generate --prompt "your description" [--provider auto] [--width 1024] [--height 1024]
Step 3: Open the .image-gen directory to view the saved image paths and the JSON output

Best Practices

Craft clear, specific prompts to guide style, composition, and elements
Leverage provider strengths: typography-focused tasks with Ideogram, photorealism with BFL or Stability, quick iterations with FAL
Start with default auto provider to explore options, then lock in a provider for consistency
Specify dimensions (width/height) when you need precise aspect ratios or canvases
Review the .image-gen outputs and iterate by refining prompts or trying different providers

Example Use Cases

Generate a Modern minimalist logo for TechStartup with the text 'NOVA' using ideogram
Create a professional product shot of a sleek smartphone on a marble surface with soft lighting (bfl)
Edit an image to remove the background and make it transparent (clipdrop)
Explore typography-focused logos or wordmarks using Ideogram for a brand
Run rapid iterations of a banner image using fal for fast turnaround

Frequently Asked Questions

Add this skill to your agents