Ecommerce Selling Point Images

When to Use

Use this skill when the user wants marketplace-ready ecommerce selling point images from a real product photo. Typical requests include:

Create Amazon, Taobao, Tmall, Temu, Shopee, TikTok Shop, Xiaohongshu, Pinduoduo, JD, or cross-border product feature images.
Turn one original product photo into a visual selling point image with text overlays, benefit callouts, usage scenes, or platform styling.
Use optional reference images as style, layout, scene, or competitor inspiration while preserving the user’s actual product.
Generate localized product feature images in Chinese, English, Japanese, Korean, or another requested target language.

The workflow is optimized for one image per request or a small set of variations. For a full Tmall/Taobao listing set with main images and detail pages, prefer the broader tmall-product-images skill. Use $gpt-image-2 as the image generation/editing engine for actual files.

Inputs

Required: one original product image. Preserve the product’s shape, color, material, logo placement, packaging, proportions, and distinctive details.
Optional: one or more reference images. Use them as style, layout, scene, lighting, typography, or competitor-analysis references while preserving the original product. $gpt-image-2 image editing supports multiple images through images; keep the original product image first and then reference images in the order the user provided or in the order most relevant to the requested output.
Optional: selling point description. If absent, infer only visible product facts and ask one focused question when the missing selling point would materially change the image.
Optional: target marketplace or platform. Default to a general clean ecommerce style if unspecified.
Optional: target language for on-image text. Default to the user’s conversation language.
Optional: aspect ratio or output size. Default to square 1024x1024; use 1536x1024 for landscape and 1024x1536 for portrait unless the user asks for a supported smaller draft size.
Optional: output count. Default to 1. Pass n only when the user asks for multiple variations.
Optional: output format and quality. Default to PNG and quality: "high" for deliverable ecommerce images.

Execution

Use $gpt-image-2 for the actual image operation. This skill provides the ecommerce strategy, prompt structure, input ordering, and compliance constraints; $gpt-image-2 provides the execution path.

Follow $gpt-image-2 image editing mode.

Because this workflow always starts from an original product image, use GPT Image Editing through fusion-api.openai_image_edit_async_submit, as documented by $gpt-image-2. Do not run new oo capability discovery during normal use.

Upload local images that must be passed to $gpt-image-2:

oo file upload "<product-image-path>" --json
oo file upload "<reference-image-path>" --json

Pass each returned downloadUrl as an image reference in images, using the $gpt-image-2 shape:

{"image_url":"<downloadUrl>"}

Put the original product image first, followed by reference images.

Build a concise prompt with this structure:

Create a marketplace-ready ecommerce selling point image.
Primary product: preserve the product from image 1 exactly, including shape, proportions, color, material, packaging, labels, and logo placement.
Reference usage: use images 2+ only for style, layout, background, lighting, composition, typography, or scene inspiration; do not copy unrelated products, brands, watermarks, or text.
Marketplace: <platform or general ecommerce>.
Target language for visible text: <language>.
Selling point: <user-provided selling point, or conservative visible/inferred benefit>.
Composition: product-first, clean commercial layout, readable mobile-safe typography, clear visual hierarchy, enough whitespace, no fake platform badges.
Avoid: QR codes, phone numbers, social handles, external URLs, competitor marks, fake certificates, fake discounts, unsupported absolute claims, distorted product details, unreadable text, watermark.

When the user did not provide a selling point, avoid inventing proof-dependent claims. Use visible, low-risk benefits such as material appearance, compactness, storage, comfort, organization, portability, or usage context only when supported by the image.

Execute with $gpt-image-2.

The underlying command shape is:

oo connector run "fusion-api" \
  --action "openai_image_edit_async_submit" \
  --data @payload.json \
  --json

Example payload:

{
  "model": "gpt-image-2",
  "prompt": "Create a marketplace-ready ecommerce selling point image...",
  "images": [
    {"image_url": "<product-downloadUrl>"},
    {"image_url": "<optional-reference-downloadUrl>"}
  ],
  "output_format": "png",
  "quality": "high",
  "size": "1024x1024"
}

Use the exact field names from $gpt-image-2: output_format, quality, size, and optional n. If using a JSON file, write valid JSON with those field names. oo connector run --json usually waits internally and returns completed image results directly; if it returns a handle, follow $gpt-image-2 polling guidance.

Result Handling

Read GPT Image 2 results according to $gpt-image-2:

Image URLs are usually in .data.data[].url.
Revised prompts may be in .data.data[].revised_prompt.
Returned metadata may include .data.size, .data.quality, .data.output_format, .data.model, and .data.usage.

Download each returned HTTP image URL with oo file download "<url>" "<output-dir>" --name "<fileNameWithoutExtension>" --ext "<png|jpeg|webp|jpg>". oo file download prints Saved to: <path> and does not support --json.

Use a clear local output directory such as /Users/yunshi/Downloads/ecommerce-selling-point-images/<short-product-name>-<timestamp>/.
Name files predictably, for example selling-point-01.png, selling-point-02.png, etc.
Preview the generated image when practical, or return the local file path and a short note describing platform, language, aspect ratio, and selling point used.
If the task returns structured result data instead of direct URLs, report the actual returned fields and do not invent file URLs.

Failure Handling

Missing original product image: stop and ask for one product image.
Too many reference images: $gpt-image-2 supports multiple edit images, but if the set is noisy, redundant, or too large for a useful prompt, select the most relevant references and summarize the rest in the prompt.
Unsupported aspect ratio, format, size, quality, or count: use the nearest $gpt-image-2 supported value and mention the adjustment.
Missing selling point: either ask one focused question or use a conservative visible benefit; do not invent certifications, performance numbers, regulated claims, awards, warranties, or safety/health claims.
Upload failure: report the exact file that failed and retry only after the path or network issue is resolved.
Connector handle or timeout: follow $gpt-image-2 result/polling guidance before rerunning. Do not start a duplicate task only because a wait window ended.
GPT Image 2 connector failure or billing/auth blocker: report the exact blocker from oo output and the next useful action.