EE_SKILLS

Read and analyze images using a vision-capable model. Supports custom focus: user can specify what to focus on (layout, code, errors, data, etc.) via prompt parameter. Use when the user asks to read, describe, analyze, or transcribe content from an image file (PNG, JPG, GIF, WebP, BMP, TIFF). Trigger keywords: read image, describe image, analyze image, what's in this image, look at this, 读图, 看图, 分析图片, 描述图片, 截图, 图片内容, 关注排版, 重点看.

Generate a polished PNG for a PowerPoint slide with image generation, save it into the project, then redraw it as a PowerPoint-friendly editable SVG. Use when the user asks to make a PPT page/infographic/slide visual, first use image gen, convert a generated PNG into SVG, redraw a PNG as SVG, or keep text easy to edit in PowerPoint.