GPT Prompt Studio

Turn a visual idea into a complete image prompt.

Describe the intended use, supply a reference or bring an existing prompt. Your Agent specifies composition, subject, text and light in a complete prompt. Series share a visual system; targeted edits preserve everything else.

Complete image prompts and reusable masters
Commercial composition, reference roles and eight portrait controls
Copyable HTML workbench and TXT / Markdown

Available now · v1.0.0

USD 9.99 · Single purchase, or get access through Pro.

What would you like to make?

A poster series

Four products need variation within one brand system.

One visual master and four complete image prompts.

Understand a reference

I like a reference but cannot describe how it works.

An evidence-based visual breakdown and a complete prompt for a new subject.

Revise a portrait prompt

I want one edit without changing the whole person or scene.

A complete revision preserving the locked scene.

Four drinks, one visual system

Actual output · original fictional example

The input

The original fictional example supplies four drinks, headlines and accent colors.

The work

Aspect ratio, cup placement, type hierarchy and light stay fixed while declared variables change.

Browse all output examples

Actual export · four campaign prompts — The included exporter expands an original fictional coffee campaign into complete prompts. This screenshot shows text output; no image generation is claimed.
Chinese workbench · same workflow — Actual Chinese output from the package, including complete prompts and variant navigation.

From your material to the finished files

Set the visual requirements
Provide your brief or images and let the Agent separate fixed requirements from variables. For a small edit, include the full original prompt.
Get complete prompts
Posters start with reading order; portraits use camera, pose, styling and related controls. Each variant is complete and ready to use.
Review, then generate if needed
Save the prompts in an offline workbench if useful. To make images, ask your Agent to use your available image tool and inspect the result.

Bring this task into your Agent

Prepare these materials

Specify use, aspect ratio, quantity, exact copy and elements to preserve. Give each reference a role: identity, composition, outfit or color.

Sign in to acquire the package.
Extract the entire folder into your project: .agents/skills/gpt-prompt-studio/SKILL.md (Codex) or .claude/skills/gpt-prompt-studio/SKILL.md (Claude Code).
Open a new chat in that project, attach your materials and replace the bracketed fields in the brief below. Ask the Agent to check package instructions and dependencies first.

Use the gpt-prompt-studio Skill. Create four 4:5 poster prompts for [brand/product], with themes [four themes]. Lock layout, type hierarchy and light; vary only theme and accent. Export complete prompts and a workbench, without generating images.

Installation & troubleshooting guide

The first result can be revised

Change only the lipstick to wine red. Preserve identity, outfit, pose, camera and lighting, and return the entire revised prompt.

What the package includes

Reusable visual masters, complete series prompts, reference breakdowns and full revisions with locked requirements preserved.
Offline HTML prompt workbench, editable JSON, Markdown and individual TXT files with revision hashes.

What you provide separately

Your own Codex or Claude Code model access. Prompt instructions have no software dependencies; the optional exporter needs Node.js 22+ and no npm packages.
Image understanding and generation use your available tools. Model subscriptions and image credits are separate.

Before you start

Does this include unlimited image generation?

You receive the prompt workflow and exporter. Model access, image-generation tools and credits come from your own account and are billed separately.

Can it reproduce a reference exactly?

It analyzes observable visual rules and assigns reference roles. It does not guarantee pixel-exact reproduction or grant rights to reference images.

Delivers prompts and a prompt workbench, not generated images or print-ready artwork. No third-party prompt collection, stock photos, private personas, provider accounts or generation credits. Independent SiSkill product; not an official OpenAI product.

More to make

Creator Content Workflow — Turn research notes into evidence-linked articles, reviewed social drafts and an organized publishing handoff.
Xiaohongshu Note Cover — Create a bold headline, useful checklist or owned-visual cover with semantic title lines. Export a 1080×1440 PNG, editable HTML and JSON.
Open Graph Image — 1200x630 share images for blog posts, product launches and docs pages in English or Chinese, readable in small link previews, delivered with og:image tags.
NotebookLM Video — Turn your sources into a reviewed video brief, then generate an explainer in your own Google notebook.