日文版 GPT-Image-Harness 信息图

{"type":"日文信息图 Slides","style":"简洁扁平矢量演示图形，极简企业设计，白色背景，深海军蓝与青色点缀，精致浅灰色边框，清晰排版，平衡间距","canvas":{"aspect_ratio":"16:9"},"title_block":{"headline":"{argument name=\"headline text\" default=\"GPT-Image-Harness 的基本构成\"}","subheadline":"{argument name=\"subheadline text\" default=\"无需每次从零开始，专为图像生成打造的工作环境。\"}","headline_alignment":"center top","headline_style":"bold sans-serif, very large, dark navy","accent":"short horizontal teal line under headline"},"layout":{"sections":[{"title":"repository tree","position":"left large panel","count":1,"panel_style":"rounded rectangle with light gray outline and dark navy header bar","header_label":"GPT-Image-Harness/","content":{"type":"folder tree","root":"GPT-Image-Harness/","items_count":8,"items":["assets/","reference/","output/","prompts/","skills/","AGENTS.md","SPEC.md"],"hierarchy":"assets 包含 reference 和 output；根目录还包含 prompts、skills、AGENTS.md 和 SPEC.md","icon_style":"outline folder and document icons, teal for active folders, gray for nested guides"}},{"title":"feature cards","position":"right stacked column","count":5,"panel_style":"five separate rounded white cards with light gray outline, evenly spaced vertically","labels":["assets/reference","assets/output","prompts","skills","AGENTS.md / SPEC.md"],"descriptions":["管理角色设定图、过往页面及参考素材","整理生成的图像、修正版及候选作品","管理漫画用、修正版及不同用途的提示词","Playwright 等必要的 CLI 调用及辅助功能","明确运营规则与整体规范"],"icons":["image icon","folder icon","document icon","terminal icon","document icon"],"title_color":"teal","body_color":"dark gray"},{"title":"bottom message","position":"bottom full width banner","count":1,"panel_style":"rounded rectangle with light teal outline","content":{"icon":"teal circled checkmark","text":"{argument name=\"footer message\" default=\"通过整理素材、指令、输出内容及规则，实现高复现性的图像生成工作流。\"}","emphasis":"短语 高复现性的图像生成工作流 为青色加粗"}}],"composition":"title and subtitle centered at top, large repository diagram on left, five explanatory cards on right, summary banner across bottom"},"text_language":"Japanese","visual_details":{"left_panel_header_gradient":"dark navy to slightly lighter navy","folder_tree_line_color":"light gray","teal_hex_approx":"#22B8C7","navy_hex_approx":"#0F2236","background":"very light gray-white"}}

• 模型gpt-image-2
• 用途social-media-post
• 风格minimal, editorial
• 画幅Flexible

• 用途social-media-post
• 风格minimal, editorial
• 推荐规格1:1, 4:5, 16:9, or 9:16
• 模型GPT Image 2

把这个 日文版 GPT-Image-Harness 信息图 提示词当作起点。替换主体、产品细节、场景、光线、品牌颜色和输出画幅，同时保留 social-media-post 的用途和 minimal, editorial 的视觉方向。

• 主体替换成你的产品、人物、场景或界面对象。
• 用途保留或改写为 social media post、广告图、封面、海报或电商主图。
• 画幅按投放位置选择 1:1、4:5、16:9 或 9:16。
• 品牌线索补充颜色、材质、道具、文案安全区和最终使用平台。

先按目标平台选择画幅，再根据用途选择 1K、2K 或 4K。产品图和海报通常适合更高清晰度。

• 检查主体、光线、构图和画幅都足够明确。
• 检查如果图片要承载标题或广告文案，已经预留文字安全区。
• 检查生成前确认分辨率、额度成本和最终使用位置。

日文版 GPT-Image-Harness 信息图 的重点是先固定图片用途，再用主体、构图、光线、风格和画幅控制生成结果。它适合 social media post，并保留 minimal, editorial 的视觉方向。

• 能直接用吗？可以。先替换主体、品牌线索和画幅，再进入 Image2Studio 生成。
• 适合什么比例？按平台选择 1:1、4:5、16:9 或 9:16。