gpt-image-2-skill
A Rust skill and CLI for agents to generate and edit images with GPT Image 2, make transparent PNGs and answer in JSON.
- 146 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Prompts and input images go to the selected provider and the Codex route uses its credentials.
- API keys are read from the environment, Keychain or a config file.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/gpt-image-2-skillDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Prompts and input images go to the selected provider and the Codex route uses its credentials.
- API keys are read from the environment, Keychain or a config file.
This is third-party code. Review the repository files before installing.
What it does
The repository ships one Rust engine in four forms: the `gpt-image-2-skill` CLI, a desktop app, a Docker web service and a Claude Code skill. The skill calls the CLI through a Node wrapper, and the CLI talks to OpenAI `gpt-image-2`, a compatible API at another base URL, or the Codex login. Every command returns JSON with an `ok` field and a machine-readable error code, while progress goes to stderr as JSONL. It covers generation, mask edits, sizes up to 4K and a transparent PNG pipeline: the image is drawn on a control matte and the background is then cut out locally.
Who it is for. Developers and designers who ask an agent to make illustrations, icons and transparent-background assets straight from the terminal.
Good fit when
- An agent draws an illustration or cover while working on a project.
- You need a logo or icon on a transparent background with a cutout quality check.
- An image is edited by mask without touching the rest.
- A script or CI calls generation and parses the JSON answer.
Not a fit when
- You have no OpenAI key, compatible API or Codex login.
- You need image models other than OpenAI ones.
- Keys cannot be stored on the machine running the agent.
Example request
Draw a red apple icon on a transparent background at 2K, save it to `./apple.png` and verify the cutout quality.Limitations
Generation is billed at the selected provider's rate. The `~/.codex/auth.json` route reuses Codex credentials, so check the terms of your subscription. The skill downloads the binary from GitHub releases when needed. Keys come from the environment, Keychain or a file; keep them out of the repository. History and config live in `$CODEX_HOME/gpt-image-2-skill/`. All four surfaces share one engine, so the binary version should match the skill version. Homebrew and Cargo installs need those package managers.
How to disable. Remove `~/.claude/skills/gpt-image-2-skill` and, if installed, the global package with `npm uninstall -g gpt-image-2-skill`. Clear config and history in `$CODEX_HOME/gpt-image-2-skill/`.
FAQ
Which providers are supported?
Three: OpenAI with `gpt-image-2`, a compatible API via `--openai-api-base`, and Codex with an `auth.json` file. Commands are identical; only the `--provider` flag changes.
How do I get a transparent PNG?
With `transparent generate`, `transparent extract` and `transparent verify`. The background is cut locally and checked against quality profiles, so provider-native transparency is not required.
How do I install the skill?
With the README `npx skills add` command or by copying the `skills/gpt-image-2-skill` folder into `~/.claude/skills/`.
Is a separate binary needed?
No. The skill wrapper searches several locations and downloads the matching release if none is found.
Can I edit an image with a mask?
Yes. The `images edit` command takes the source image and a mask, and the rest of the picture is unchanged. See `--help` for the exact flags.
Discussion
No questions or reviews yet.
Related
SkillsA skill and CLI for generating and editing images with GPT Image 2 and 2.5 plus a library of ready prompts
Claude Image (GPT Image 2)
claude-image
A skill teaching agents to work with GPT Image 2. OpenAI API access is needed and it is unavailable from Russia without workarounds
GPT-Image2 Style Library
gpt-image-2-style-library
A skill picks a matching template and style from a library of GPT-Image2 cases and assembles it into a finished prompt
Sprite Gen
sprite-gen
A skill for Codex and Claude and a Python CLI: builds 2D game sprites from one base image, removes the background and assembles animation atlases
A Claude Code skill that compiles a vague request into a finished gpt-image-2 prompt and checks it with a validator script
Image generation and editing from any agent: the imagegen-agent CLI and a skill with the Codex prompt library for compatible APIs
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Repository stars | Main difference |
|---|---|---|---|
| gpt-image-2-skill (this card) | no | 146 | |
| GPT-Image2-Skill | no | 6,011 | another gpt-image-2 skill; this one is a standalone Rust CLI with a JSON protocol and three providers |
| GPT Image 2 Prompts by YouMind | no | 10,249 | a collection of ready prompts; this tool runs generation and editing itself |
README in short
The skill only forwards arguments to the Rust engine, so the CLI, app and web version give the same result. Result storage and webhook details apply to the app and Docker version.