Team T-Shirt Design

Problem to solve

A class or club T-shirt usually needs to do three things: match the team's theme, keep the front and back styles consistent, and give a quick preview of how the print will look on a real garment. Prompting an image model directly can lead to conflicting front/back elements, unpredictable print placement, and no try-on reference. This skill turns a class name, keywords, and style into a reviewable T-shirt design workflow with front/back artwork and a three-person preview.

How it works

The required inputs are name, key_word, and style. The skill first validates the request and generates prompts for the front artwork, back artwork, and try-on image. The back design stays stylistically aligned with the front but is simplified. It then submits asynchronous front and back image-generation tasks and returns a task_id. The caller must poll the task status until the front artwork, back artwork, and three-person try-on image URLs are available. The output is typically three 1:1 images: a front design, a simplified back design, and a model try-on showing two female and one male figures, two facing forward and one from the back.

Scope and cautions

• Real generation requires showapi_appkey; without it, the skill may run in simulation mode, which should be treated as a preview rather than a production result.
• Tasks are asynchronous only; the task_id is the handle used for polling.
• Each polling response should be surfaced to the user promptly instead of waiting silently for the final result.
• This skill is best for class uniforms, club merchandise, and graduation-season garment previews, not for vector cutting files, print proofing, or garment manufacturing.