GPT Image 2
Image generation and editing via structured prompts
Install
Ships scriptsClone github.com/ConardLi/garden-skills and place skills/gpt-image-2 in your Claude Code skills folder
Use when you need to generate or edit images with structured prompts across different execution environments.
- Author
- ConardLi
- License
- MIT
About this skill
GPT Image 2 handles image generation and editing workflows in three modes. It detects your environment and adapts: in local mode it runs generation scripts directly, in native-tools mode it delegates to your agent's built-in capabilities, and in advisor mode it crafts polished prompts you can execute elsewhere.
Highlights
- Operates in three modes depending on available tools and configuration
- Includes 18 template categories with 80+ structured prompt patterns
- Maps user inputs to template parameters for consistent output
- Saves generated prompts and images with timestamped organization
- Works with OpenAI API in local mode; compatible with host tools in other modes
Who it's for
Use this skill when you need programmatic image generation with fine-grained control over prompt structure. It's useful for creating consistent visual assets, exploring image-to-text workflows, or using your agent as an image prompt engineer. Installation takes seconds—just copy the folder to your Claude Code skills directory.
Similar skills
Video Production
SkillCreate and edit videos with AI generation and programmatic tools
Generative Media Skills
SkillImage, video and audio generation wired into a coding agent through a single CLI.
Marketing Image Generation
SkillGuide for producing professional marketing images with AI and design tools
Comments(0)
Sign in to comment