FlowCay logo
FlowCay
GPT-Image-2 Ecommerce Skill

GPT-Image-2 Ecommerce Skill

Claude Code Skill for one-click e-commerce creative generation powered by GPT-Image-2 with 25 built-in scene templates.

Install command

npx skills add https://github.com/buluslan/gpt-image2-ecommerce
SkillOpen SourceAll Platforms
Added on Popularity 18

Details

Overview

GPT-Image-2 Ecommerce Skill (by buluslan, MIT, v0.1.0 released Apr 2026) is an e-commerce creative generation skill that drives OpenAI's GPT-Image-2 model through the Codex CLI from any AI agent. It ships 25 professional, structured prompt templates covering the most common e-commerce visual scenarios (grouped into 5 categories: listing essentials, atmosphere, ad creative, social content, and technical display). You describe what you need in plain language, the agent auto-matches the best template, fills in the variables, and executes one codex exec command to render the image.

Key Features

  • 25 built-in scene templates across 5 categories:
  • Listing essentials: white/plain background hero image, multi-angle grid, size/spec diagram, product bundles.
  • Atmosphere: lifestyle scene, flat lay, macro detail, model display, virtual try-on.
  • Ad creative: poster/banner, creative concept ad, magazine cover, luxury atmosphere, sports campaign.
  • Social content: Xiaohongshu/Instagram/TikTok, UGC buyer-show, livestream screenshot.
  • Technical display: exploded view, ghost mannequin, device mockup (SaaS/app), packaging design, infographic/A+.
  • Plus seasonal campaigns and storefront scenes — each template offers 4 style variants.
  • Structured JSON prompts: each template decomposes the prompt into fields — subject, background, lighting, composition, quality standards — for consistent professional output (studio lighting and standard composition on listing images).
  • Smart template matching: natural-language request → intent recognition (scene type + product info + style) → best template → prompt assembly → image generation.
  • Reference-image consistency: pass in product photos to keep brand consistency across generations.
  • Anti-AI-sense processing: UGC/livestream/social templates apply CCD retro effects so images look like real user shots, not AI renders.
  • Hybrid invocation mode: CLI (codex exec) with automatic HTTP-service fallback; works on Claude Code, OpenClaw, Cursor, Windsurf or directly from a terminal.
  • 6-step workflow in SKILL.md: intent recognition → template matching → prompt assembly → generation call → temp-file cleanup → optimization suggestions.
  • Zero config beyond prerequisites: only needs Codex CLI installed and logged in (optional Claude Code for Skill-mode auto-matching).

How to Use

  1. Use a Skill-compatible agent such as Claude Code, Cursor, or Codex.
  2. Install from the terminal (Node.js required):
npx skills add https://github.com/buluslan/gpt-image2-ecommerce

Or paste the install prompt from this tool page into your agent and ask it to install for you. 3. Restart / reload the agent so the new Skill is loaded. 4. Invoke the Skill with its slash commands or by asking the agent to run the workflow described in Overview / Key Features. 5. Review outputs before publishing — treat model-generated copy, images, and data as drafts.

Pricing

  • The Skill itself: free and open source (MIT) on GitHub — clone it and use it.
  • Image generation cost: all generation goes through your own GPT-Image-2 usage via Codex CLI, so you pay OpenAI's image-generation API rates based on your volume.

Best For

E-commerce sellers, operations staff and social-media creators who need fast, consistent product visuals — main images, lifestyle scenes, UGC-style shots, ad creatives, packaging and technical detail images — driven by natural language from any AI agent, without a designer or dedicated SaaS subscription.

FAQ

What do I need to run it?

Codex CLI installed and logged in (it drives GPT-Image-2). Any agent that can run shell commands — Claude Code, OpenClaw, Cursor, Windsurf or a plain terminal.

How do I use it?

Describe the image in natural language, e.g. "generate a sports campaign image for an electric bike" or "create a set of premium skincare white-background main images". The agent matches a template and runs Codex CLI to generate.

What scenarios are covered?

25 templates: hero/white-background, lifestyle, flat lay, macro, posters, social media (Xiaohongshu/Instagram/TikTok), UGC buyer-shots, models, before/after, packaging, infographics/A+, concept ads, spec diagrams, bundles, livestream, try-on, exploded views, ghost mannequin, grids, magazine covers, seasonal, luxury, mockups, storefronts and sports ads.

Can it keep product identity consistent?

Yes — you can pass reference product images so generated results stay consistent with your brand/product across the set.

Is it free?

The Skill is free (MIT). You only pay OpenAI's GPT-Image-2 image-generation API costs for the images you generate.