jimeng-infographic-skill

This repository contains a skill file for Jimeng, a Chinese AI image generation platform made by ByteDance, the company behind TikTok. Jimeng has an Agent mode where users can load custom skill files to give the AI additional instructions for specific tasks. This particular skill is designed to generate infographic posters.

The skill guides the user through a four-step process. First, you choose a language (Chinese or English), an image model version, a size, and a visual style. Second, you provide a title, color preferences, and the content you want displayed, and you can optionally upload a reference image. Third, the AI infers a structure for the information, fills in any gaps, and asks you to confirm before generating four poster variations. Fourth, you review the results and either accept them or ask for adjustments.

All four generated images share the same theme, style, and color palette but use different layouts. One centers the content with the title at the top, one splits information into left and right columns, one arranges content in a vertical flow, and one emphasizes key numbers or data points in large type. The skill supports five visual styles, including flat illustration, isometric 3D, and data visualization, and five aspect ratios ranging from square to portrait and landscape.

To use it, you open Jimeng's Agent mode, drag the single skill file into the chat window or upload it through the add-skill interface, and then follow the prompts. The README is written in Chinese and describes the feature set clearly. The repository contains only the one skill definition file.