메뉴
Analyze image content with multi-language output and comprehensive type detection. Use when user mentions image analysis, analyze picture, photo analysis, visual inspection, 图片分析, 图像识别, 看图, 分析图片, 截图, 图表.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Browser automation platform controlling Chrome for navigation, clicking, input, screenshots, data extraction and AI exploration. Use when user mentions browser automation, web scraping, screenshot capture, page navigation, Chrome control, 浏览器自动化, 网页操作, 截图, 数据提取, AI探索.
Control 2D CAD software (AutoCAD, GstarCAD, ZWCAD) via Python COM interface with natural language commands. Covers full drafting workflow: shapes, text, dimensions, layers, file save. Use when user mentions AutoCAD, 2D CAD, CAD automation, engineering drawing, CAD绘图, 工程制图, 图纸绘制, DWG文件, 浩辰CAD, 中望CAD.
Academic research and learning framework for CNKI (China National Knowledge Infrastructure). Provides paper reading guidance, literature management, and research methodology support. Use when user mentions CNKI, academic papers, literature review, research methods, 论文阅读, 文献管理, 知网, 研究方法, 学术研究.
Automatically organize desktop files by type into categorized folders such as Documents, Images, Scripts, Data, Archives, and Others. Use when user mentions file organization, clean desktop, organize files, 整理桌面, 文件分类, 整理文件.
High-fidelity HTML prototyping, interactive demos, slide decks, animations, design exploration with expert review. Embodies different design personas (UX designer, animator, slide designer, prototyper). Use when user mentions prototype, design demo, interactive prototype, HTML demo, animation demo, design variant, UI mockup, app prototype, design exploration, 做原型, 设计Demo, 交互原型, HTML演示, 动画Demo, 设计探索.
Convert documents (PDF, Word, Excel, PPT, images, audio) to Markdown using Microsoft MarkItDown. Full workflow: format detection, conversion, OCR, LLM image description, batch processing. Use when user mentions document conversion, PDF to markdown, Word to markdown, OCR, 文档转换, PDF转Markdown, Word转换, 文件格式转换.
| name | vision-analyzer |
| description | Analyze image content with multi-language output and comprehensive type detection. Use when user mentions image analysis, analyze picture, photo analysis, visual inspection, 图片分析, 图像识别, 看图, 分析图片, 截图, 图表. |
| user-invocable | true |
| model | sonnet |
| trigger-keywords | ["图片","图像","截图","照片","图标","图表","视觉","分析图片","描述图片","识别图片","图片内容","图像分析","看图说话","图片描述","image","picture","photo","screenshot","chart","graph","diagram","analyze image","describe image","what's in this image",".png",".jpg",".jpeg",".gif",".webp",".bmp"] |
基于 Qwen-VL 的自动图片分析系统，为 OpenCode 提供强大的视觉能力。
当以下情况出现时自动触发图片分析：
.png、.jpg、.jpeg、.gif、.webp、.bmp等扩展名http://或https://的图片链接只需在对话中提到图片或相关关键词，系统会自动分析：
用户：帮我看看这个截图 C:\Users\曾金昌\Pictures\Screenshots\test.png
AI： [自动分析图片并提供详细描述]
可以直接请求图片分析：
用户：请分析这张图片
用户：描述一下这个截图的内容
用户：这是什么图片？
用户：分析图片：https://example.com/image.jpg
系统会自动检测或根据用户请求使用相应语言：
用户：用英文分析这张图片
用户：Describe this image in English
用户：この画像を分析してください（日文）
技能使用以下默认配置：
qwen-vl-plushttps://dashscope.aliyuncs.com/compatible-mode/v1~/.ds_cache/view_image/确保以下Python包已安装：
pip install openai httpx pillow
[视觉分析结果] (耗时2.3秒 | 详细程度: 详细)
------------------------------------------------------------
这张图片显示了一个软件的代码编辑界面。界面分为左右两个主要部分：
左侧是一个文件浏览器，显示目录结构，包含多个文件夹如"src"、"tests"、"docs"等。
右侧是代码编辑器，显示Python代码，内容是关于数据处理的函数定义。
界面顶部有菜单栏，包含"File"、"Edit"、"View"等选项。代码编辑区域使用深色主题，
语法高亮显示关键字、字符串和注释。底部状态栏显示行号、列号和文件编码信息。
整体布局整洁专业，是典型的IDE（集成开发环境）界面。
------------------------------------------------------------
可以通过上下文指定分析详细程度：
支持以下语言代码：zh(中文)、en(英文)、ja(日文)、ko(韩文)、fr(法文)、es(西班牙文)、de(德文)
可以指定特定的分析问题：
系统会优雅处理各种错误情况：
所有错误都会提供友好的中文提示和解决方案建议。
本技能设计为与OpenCode无缝集成：