gpt-images
A Claude skill that turns a plain brief into a production-ready prompt for OpenAI GPT Image models (ChatGPT Images 2.5): six-block prompts, the hard API limi…
Install with your AI
Paste into Claude Code, Cursor, or any agent — it reads the repo and wires the tool into your project.
Install and set up gpt-images (claude-skill project) into my current project. Found on https://claudeers.com/gpt-images Repo: https://github.com/lukasersil/gpt-images Homepage/docs: — Detected install method: claude-skill → # copy this skill into .claude/skills/gpt-images/ Category: skills. Platforms: cli, api. Read the repo's README for exact setup and env vars, then install it and wire it into my project. Claudeers Health Verdict: active; community-verified: false. Confirm the source before running anything.
# copy the skill dir into your project: # .claude/skills/gpt-images/ (or ~/.claude/skills/gpt-images/ for all projects)
git clone https://github.com/lukasersil/gpt-images
// compatibility
| Platforms | cli, api |
|---|---|
| Operating systems | — |
| AI compatibility | claude |
| License | MIT |
| Pricing | open-source |
| Language | — |
GPT Images: a prompt-writing skill for Claude
Give it a brief in plain language. It gives back an English prompt for OpenAI GPT Image models, a settings block for the surface you are on, and three levers for the next round.
Covers ChatGPT Images 2.5 (gpt-image-2.5-flare, gpt-image-2.5-sunburst) and gpt-image-2.
What is inside
| File | What it holds |
|---|---|
SKILL.md | the workflow: surface, intake, model choice, seven-block prompt, references, edits, handover |
references/official-spec.md | hard numbers: model IDs, size rules, quality levels, input limits, pricing, endpoints |
references/prompt-anatomy.md | text rendering, style transfer, transparency, five failure modes and their fixes |
references/recipes.md | seven skeletons: poster, article cover, product shot, portrait, infographic, logo, UI mockup |
Self-contained. No other skill required.
Install in Claude Code
git clone https://github.com/lukasersil/gpt-images.git ~/.claude/skills/gpt-images
That makes it available in every project. For one project only, clone into .claude/skills/gpt-images inside that project instead.
Copy the whole folder, not just SKILL.md. The references are half the skill, and a folder with only the top-level file loads and then quietly underperforms.
Install on claude.ai
Download gpt-images.zip from Releases, then go to Settings, Capabilities, and upload it.
Do not unzip and rezip it through Finder. macOS adds a wrapper folder and a __MACOSX sidecar, and the uploader rejects the result with "SKILL.md file must be in the top-level folder, not nested deeper." If you have to repack, use zip -rq gpt-images.zip gpt-images -x "*.DS_Store" from a terminal.
How to use it
Just describe what you need:
- "poster for an October workshop, A4, dark background"
- "product shot of this bottle on wet stone"
- "change only the jacket color, keep everything else"
- "logo on a transparent background, has to work at 32px"
The skill asks for whatever is missing, assumes the rest and marks its assumptions.
It does not generate images. It writes the prompt and then asks. Generation costs money and that call is yours.
One caveat
The numbers in references/official-spec.md are a snapshot from 8 September 2026, the day Images 2.5 shipped. Model catalogs move fast. Before betting a production pipeline on a limit, check the current OpenAI documentation. The prompting method in SKILL.md and references/prompt-anatomy.md ages far more slowly than the parameter table.
Pull requests that update the spec against newer documentation are welcome. Keep the "Unverified" section honest: if the docs are silent, say so instead of guessing.
License
MIT. See ATTRIBUTION.md for sources.
// faq
What is gpt-images?
A Claude skill that turns a plain brief into a production-ready prompt for OpenAI GPT Image models (ChatGPT Images 2.5): six-block prompts, the hard API limits, and edits that don't drift. MIT.. It is open-source on GitHub.
Is gpt-images free to use?
gpt-images is open-source under the MIT license, so it is free to use.
What category does gpt-images belong to?
gpt-images is listed under skills in the Claudeers registry of Claude-compatible tools.
// embed badge
[](https://claudeers.com/gpt-images)
// retro hit counter
[](https://claudeers.com/gpt-images)
// reviews
// guestbook
// related in Claude Skills
🔓An agentic skills framework & software development methodology that works.
🔓
🔓💫 Toolkit to help you get started with Spec-Driven Development
🔓AI coding assistant skill (Claude Code, Codex, OpenCode, Cursor, Gemini CLI, and more). Turn any folder of code, SQL schemas, R scripts, shell scripts, docs,…