AgentSkillsCN

pontasan-genimg

AI图像生成技能。每当您需要在实现前端代码（HTML/CSS/TSX）时，创建新的图像资源（JPEG、PNG、SVG等），如图标、插图、背景、Logo或横幅，都可主动调用此技能。当所需的图像资产缺失或尚不存在时，无需等待用户明确指示，即可主动调用此技能。

SKILL.md
--- frontmatter
name: pontasan-genimg
description: AI image generation skill. Proactively use this skill whenever you need to create new image resources (JPEG, PNG, SVG, etc.) such as icons, illustrations, backgrounds, logos, or banners while implementing frontend code (HTML/CSS/TSX). When a required image asset is missing or does not yet exist, invoke this skill on your own initiative without waiting for explicit user instructions.

Usage

Based on the prompt content, determine the appropriate image to generate and run the following script. The script generates an image according to your instruction (<IMAGE_SPEC_JSON>) and writes the generated image file path to standard output. Retrieve the image file path from standard output and use it as part of your task. .ts files can be executed directly by Node.js, so TypeScript compilation is not required.

bash
# Set timeout to 600000ms (10 minutes) as image generation can take a long time
npm --prefix .claude/skills/pontasan-genimg/script run generate -- <IMAGE_SPEC_JSON>

<IMAGE_SPEC_JSON> Description

IMAGE_SPEC_JSON must be a JSON object with the following fields:

[{ "filePath": "The expected output image file path (must be an absolute path)", "mime": "The MIME type of the image", "prompt": "The prompt used to generate the image", "mode": "fast or quality. fast: high speed generation. quality: slower but higher quality. quality has strict usage limits, so use it only when high quality is truly needed. Default to fast in most cases." }]