Universal Field Study Skill

A developer has created a universal field study skill that converts any reference photo into a minimalist editorial illustration. The skill, available as a SKILL.md file, uses AI to identify the subject type and generate a simplified image with a warm cream background, limited colors, and sparse ink lines. It is designed for landscapes, animals, architecture, people, and other subjects, producing restrained, poetic visuals rather than realistic or cartoonish renderings.

将任意照片转译成极简编辑式 Field Study 插画。 适用内容： - 风景 - 动物 - 建筑 - 人物 - 鱼群、鸟群等重复主体 - 烟花、日落和强光源 - 室内、窗景和静物 - 上传一张参考照片 - 将 SKILL.md 提供给支持 Skill 或自定义指令的 AI - 告诉 AI：使用这个 Skill 处理当前图片 - 直接生成插画 暖奶油色背景、少量主色、克制墨线、大面积留白、非写实、非卡通。 完整指令见： SKILL.md | name | gc-universal-field-study | |---|---| | description | 将任意参考照片提炼成极简编辑式 Field Study 插画。自动识别主体类型，保留最有辨识度的关系，用少量颜色、细墨线、大面积留白和暖奶油色背景重构画面。适用于风景、建筑、动物、人物、车辆、鱼群、鸟群、烟花、室内物件及其他照片。 | 把用户上传的任意照片，转译成一张极简、克制、成熟的 editorial “field study” 插画。 这不是水彩滤镜，也不是对原照片进行完整重画。核心任务是： 先在内部判断以下内容： 不要在最终回复中输出分析过程。 F0EBDD 。适用于：热气球、月亮、路灯、单个人、单只动物、小船、小物件。 适用于：山野、森林、农田、海岸、沙漠、雪地、城市远景。 根据当前照片动态生成英文图像编辑提示词。不得复制固定的“帐篷、森林、山、雾”等具体名词，除非当前图片中确实存在这些元素。 提示词应包含以下结构： 使用下面的结构，并将方括号内容根据当前图片动态替换： Transform Image A into an extremely minimal editorial “field study” illustration. Image A is the sole content and composition reference. Do not repaint the full photograph. Extract only the essential visual relationships from the image: describe the main subject, its memorable action or silhouette, its approximate position, and the minimum supporting elements required for recognition . Use a flat warm cream background, exact color F0EBDD. Keep the background clean and mostly empty, with no visible paper grain, scenery fill, border, decorative frame, or typography. Preserve the subject’s identity, direction, key gesture, and approximate spatial relationship, but simplify everything aggressively. insert subject-specific transformation instructions Use only 2–5 dominant colors sampled from Image A. Render them as flat, muted watercolor shapes with slightly irregular edges. Use sparse, thin, imperfect ink lines only for essential contours, structure, expression, or one simple ground/horizon line. Remove all realistic detail that is not required for immediate recognition: fine texture, material rendering, wrinkles, hair or fur strands, individual leaves, small architectural parts, complex reflections, gradients, shading, realistic lighting, depth rendering, and detailed brushwork. Treat the scene as one compact visual specimen rather than a complete photograph. The full motif should occupy approximately chosen percentage of the canvas and remain chosen position , leaving generous quiet negative space. Let all marks stop or fade naturally before reaching the canvas edges. The result should feel restrained, poetic, observational, mature, and editorial, like a contemporary naturalist’s sketchbook study. Not a conventional watercolor painting. Not realistic. Not cartoonish. Not cute. Not decorative wall art. No text, numbers, labels, color swatches, frame, UI, photograph, or typography. Output only the finished minimalist illustration. 每次生成时都要确保： 生成前，在内部确认： 若任一项不符合，先修改提示词再生成。