gpt-image-1-5

GPT Image 1.5 provides a CLI-based interface for generating and modifying visual assets using OpenAI models. This tool integrates directly into the local development environment, outputting files based on textual prompts. It supports standard image creation, full-canvas transformations, and pixel-precise inpainting via mask files. Users maintain control over image quality, dimensions, and background transparency by passing specific arguments. When performing edits, the script handles either entire-image style changes or specific area modifications using provided masks. By automating file naming with timestamps and naming conventions, the tool maintains an organized asset library in the current directory. It is intended for rapid prototyping, asset iteration, and creative tasks within a terminal-centric workflow, ensuring results are saved locally without needing a web browser.

When to Use This Skill

• •Creating custom UI icons or background graphics for software projects
• •Rapidly iterating on design mockups by applying style changes to existing prototypes
• •Inpainting specific regions of an image to remove or replace unwanted elements
• •Generating assets with alpha channels for web or app design

How to Invoke This Skill

Example prompts that trigger this skill in Claude Code, Cursor, or Antigravity:

• “Create a high-res image of a futuristic city skyline
• “Generate a square icon for a coffee shop app
• “Change the background of this image to a beach
• “Add a mountain range to the background of my landscape photo
• “Make this image look like an oil painting
• “Create a tall vertical draft of a forest

Pro Tips

• 💡Utilize the `--mask` parameter for highly targeted edits, ensuring only the specified area is modified for precise control over your image manipulations.
• 💡Experiment with the `--quality` and `--size` parameters to balance output fidelity with generation time, especially for prototyping or final asset creation.
• 💡When editing, ensure your `--prompt` clearly describes the desired changes or what should appear in the masked area for optimal results.

What this skill does

• •Generate original imagery from descriptive text prompts
• •Perform full-image edits to alter styles or compositions
• •Conduct localized inpainting using transparent PNG masks
• •Customize output size between square, portrait, and landscape formats
• •Toggle between low, medium, and high quality rendering settings
• •Control background transparency for generated assets

When not to use it

• ✕High-fidelity professional print design requiring specific color profiles or vector exports
• ✕Mass-batch image processing where significant latency or API cost management is needed
• ✕Tasks requiring manual brush-based fine-tuning or vector path manipulation

Example workflow

1. User executes generate_image.py to create a base image of a workspace
2. Script saves the file to the current working directory with a timestamp
3. User creates a mask image in a photo editor highlighting the computer screen area
4. User executes the script again with --input-image and --mask paths
5. Script replaces the screen content while preserving the surrounding workspace
6. User confirms the file path returned by the script output

Prerequisites

• –uv environment installed
• –OpenAI API key available via CLI flag or environment variable
• –Base image file for editing workflows
• –PNG mask file for inpainting

Pitfalls & limitations

• !Running the script inside the skill directory instead of the project directory may clutter internal folders
• !Large high-quality renders can increase API latency and execution time
• !Masks must be correctly aligned with the input image dimensions for accurate inpainting

FAQ

How it compares

Unlike manual web-based image tools, this skill allows for programmatic integration and file management within a terminal workflow without ever leaving your IDE.

Source & trust

Recommended Rules

Recommended Workflows

Recommended MCP Servers