{"schemaVersion":"drillso.agent.session.v1","scope":"node","resource":{"type":"shared-session","shareId":"tbhNiIW8RWXL","title":"怎么知道自己当前会什么？不会什么？该学什么？怎么应用？","canonicalUrl":"https://drillso.com/zh/share/sessions/tbhNiIW8RWXL/%E6%80%8E%E4%B9%88%E7%9F%A5%E9%81%93%E8%87%AA%E5%B7%B1%E5%BD%93%E5%89%8D%E4%BC%9A%E4%BB%80%E4%B9%88%E4%B8%8D%E4%BC%9A%E4%BB%80%E4%B9%88%E8%AF%A5%E5%AD%A6%E4%BB%80%E4%B9%88%E6%80%8E%E4%B9%88%E5%BA%94%E7%94%A8%EF%BC%9F-cabbab3e","agentUrl":"https://drillso.com/zh/share/sessions/tbhNiIW8RWXL/agent.json?node=%E6%80%8E%E4%B9%88%E7%9F%A5%E9%81%93%E8%87%AA%E5%B7%B1%E5%BD%93%E5%89%8D%E4%BC%9A%E4%BB%80%E4%B9%88%E4%B8%8D%E4%BC%9A%E4%BB%80%E4%B9%88%E8%AF%A5%E5%AD%A6%E4%BB%80%E4%B9%88%E6%80%8E%E4%B9%88%E5%BA%94%E7%94%A8%EF%BC%9F-cabbab3e","ownerName":"pyth0nb3st","updatedAt":"2026-05-09T12:59:12.806Z"},"currentNode":{"id":"cabbab3e-87a6-4d68-ab88-68da879c4b63","slug":"怎么知道自己当前会什么不会什么该学什么怎么应用？-cabbab3e","title":"怎么知道自己当前会什么、不会什么、该学什么、怎么应用？","type":"deck","url":"https://drillso.com/zh/share/sessions/tbhNiIW8RWXL/%E6%80%8E%E4%B9%88%E7%9F%A5%E9%81%93%E8%87%AA%E5%B7%B1%E5%BD%93%E5%89%8D%E4%BC%9A%E4%BB%80%E4%B9%88%E4%B8%8D%E4%BC%9A%E4%BB%80%E4%B9%88%E8%AF%A5%E5%AD%A6%E4%BB%80%E4%B9%88%E6%80%8E%E4%B9%88%E5%BA%94%E7%94%A8%EF%BC%9F-cabbab3e","agentUrl":"https://drillso.com/zh/share/sessions/tbhNiIW8RWXL/agent.json?node=%E6%80%8E%E4%B9%88%E7%9F%A5%E9%81%93%E8%87%AA%E5%B7%B1%E5%BD%93%E5%89%8D%E4%BC%9A%E4%BB%80%E4%B9%88%E4%B8%8D%E4%BC%9A%E4%BB%80%E4%B9%88%E8%AF%A5%E5%AD%A6%E4%BB%80%E4%B9%88%E6%80%8E%E4%B9%88%E5%BA%94%E7%94%A8%EF%BC%9F-cabbab3e","text":"{\n  \"style\": \"Manga Style\",\n  \"title\": \"把学习变成能力系统\",\n  \"format\": \"presenter_slides\",\n  \"slides\": [\n    {\n      \"size\": \"1536x864\",\n      \"model\": \"gpt-image-2\",\n      \"title\": \"不是不努力，是不知道往哪努力\",\n      \"layout\": \"四格漫画开场。主角坐在书和课程图标堆里，头顶冒出四个问题气泡。最后一格出现一条清晰路径。\",\n      \"status\": \"completed\",\n      \"message\": \"学习的第一步，是把模糊焦虑变成清晰问题。\",\n      \"imageUrl\": \"https://cdn.drillso.com/generated/drill/x6euj249qdncoeyeva8qn63u-ab677975612bfd33a593cdc6a0deef6b.png\",\n      \"storageKey\": \"generated/drill/x6euj249qdncoeyeva8qn63u-ab677975612bfd33a593cdc6a0deef6b.png\",\n      \"imagePrompt\": \"Create exactly one complete presentation slide image.\\n\\nUse the following NotebookLM / Kael style prompt as the dominant art direction. Follow it closely.\\n\\nSTYLE TEMPLATE START\\nUnderstanding becomes deeper with fun.\\nTurn the information into a comic-like visual explanation with a simple story.\\nUse panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions.\\nRelate the idea to a realistic situation so the audience can remember it.\\nThe style should be fun and explanatory, not childish or chaotic.\\nSTYLE TEMPLATE END\\n\\nDeck title: 把学习变成能力系统\\nDeck summary: 面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\\nDeck format: presenter_slides\\nChosen style: Manga Style\\n\\nSlide 1: 不是不努力，是不知道往哪努力\\nMain message: 学习的第一步，是把模糊焦虑变成清晰问题。\\n\\nVisible text must be exactly these lines, with no extra words:\\n我到底会什么？\\n不会什么？\\n该学什么？\\n怎么用起来？\\n\\nLayout:\\n四格漫画开场。主角坐在书和课程图标堆里，头顶冒出四个问题气泡。最后一格出现一条清晰路径。\\n\\nVisual style:\\n漫画分镜，表情夸张但不幼稚，问题气泡大，路径线作为视觉主轴。\\n\\nAdditional image direction:\\nA four-panel comic opening scene. A young office learner sits at a desk surrounded by course cards, saved articles, books, and app icons, looking confused. Four speech bubbles contain the visible questions. In the final panel, the clutter transforms into a simple path with six small stepping stones. Use expressive manga reactions, clean panels, speed lines around the confusion, and concise Chinese text.\\n\\nRequirements:\\n- Canvas: 1536x864 landscape presentation slide.\\n- Generate one slide only, not a contact sheet and not multiple slide candidates.\\n- Make it presentation-ready, high-sensibility, and coherent with the chosen style template.\\n- Make it visually finished, not a sparse wireframe or placeholder diagram.\\n- Use the strongest signature elements of the style template: palette, typography, composition language, icon/illustration treatment, and editorial rhythm.\\n- Keep all visible text readable at presentation distance.\\n- Use low text density, strong hierarchy, and source-grounded diagrams, editorial composition, or template-specific visuals.\\n- Do not add markdown symbols, fake logos, watermarks, signatures, UI browser chrome, or extra captions.\\n- Match the source/user instruction language for all visible slide text and speaker-facing copy.\",\n      \"slideNumber\": 1,\n      \"visibleText\": [\n        \"我到底会什么？\",\n        \"不会什么？\",\n        \"该学什么？\",\n        \"怎么用起来？\"\n      ],\n      \"visualStyle\": \"漫画分镜，表情夸张但不幼稚，问题气泡大，路径线作为视觉主轴。\",\n      \"speakerNotes\": \"开场先点出常见困惑：很多人买了很多课、收藏了很多资料，却依然觉得能力模糊。真正的问题不是不努力，而是不知道该往哪里努力。本 deck 会用一条方法链来回答：盘点能力、识别差距、选择目标、设计学习、实践应用、复盘迭代。\",\n      \"revisedPrompt\": \"Create exactly one complete presentation slide image, landscape 16:9, polished and presentation-ready, in a Manga Style inspired by NotebookLM / Kael editorial comic explanation. Dominant art direction: understanding becomes deeper with fun; turn the information into a comic-like visual explanation with a simple story; use panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions; realistic situation, fun and explanatory, not childish or chaotic.\\n\\nSlide topic: Chinese learning-and-ability system. Title concept: 把学习变成能力系统. Slide 1 main message: 不是不努力，是不知道往哪努力. The slide should feel like the opening of a story: a young office learner is overwhelmed by ambiguity, then a simple path appears.\\n\\nLayout: a clean four-panel comic opening scene across the slide, with strong editorial rhythm and clear hierarchy. Panel 1: the learner sits at a desk surrounded by course cards, saved articles, books, and app icons, looking confused, with dramatic manga speed lines and clutter. Panel 2: an expressive close-up reaction with a large speech bubble containing exactly this text: 我到底会什么？ Panel 3: another reaction close-up and question bubble containing exactly this text: 不会什么？ Panel 4: the clutter transforms into a simple clear path with six small stepping stones leading forward, with two more large speech bubbles containing exactly these texts, each on its own line: 该学什么？ and 怎么用起来？\\n\\nVisible text must be exactly these four lines and no extra words anywhere in the image:\\n我到底会什么？\\n不会什么？\\n该学什么？\\n怎么用起来？\\n\\nImportant constraints: no other visible text, no title text, no subtitles, no labels, no watermark, no logos, no UI chrome, no extra captions. Make the four speech bubbles the only text elements. Chinese text must be large, crisp, and readable at presentation distance. Use clean comic panel borders, bold linework, expressive but mature office-learner character design, and a palette of muted warm beige, off-white, charcoal, desaturated blue, and accent red/orange for emphasis. The final panel should visually resolve the clutter into a simple path as the visual main axis, with six small stepping stones and a forward direction. Make the composition feel finished, balanced, and coherent, with strong manga framing, motion lines, and a high-end slide design aesthetic.\",\n      \"storageFileId\": \"x6euj249qdncoeyeva8qn63u-ab677975612bfd33a593cdc6a0deef6b\"\n    },\n    {\n      \"size\": \"1536x864\",\n      \"model\": \"gpt-image-2\",\n      \"title\": \"先区分：知道，不等于会\",\n      \"layout\": \"竖向升级阶梯。主角从底层听讲，逐步走到顶层教别人，每层用一个简短动作表示。\",\n      \"status\": \"completed\",\n      \"message\": \"能力不是一个点，而是从听过到能教别人的层级。\",\n      \"imageUrl\": \"https://cdn.drillso.com/generated/drill/pdusfddf92up1ielh2rerlng-f2ea5ed653c402ef3317366352e59a05.png\",\n      \"storageKey\": \"generated/drill/pdusfddf92up1ielh2rerlng-f2ea5ed653c402ef3317366352e59a05.png\",\n      \"imagePrompt\": \"Create exactly one complete presentation slide image.\\n\\nUse the following NotebookLM / Kael style prompt as the dominant art direction. Follow it closely.\\n\\nSTYLE TEMPLATE START\\nUnderstanding becomes deeper with fun.\\nTurn the information into a comic-like visual explanation with a simple story.\\nUse panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions.\\nRelate the idea to a realistic situation so the audience can remember it.\\nThe style should be fun and explanatory, not childish or chaotic.\\nSTYLE TEMPLATE END\\n\\nDeck title: 把学习变成能力系统\\nDeck summary: 面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\\nDeck format: presenter_slides\\nChosen style: Manga Style\\n\\nSlide 2: 先区分：知道，不等于会\\nMain message: 能力不是一个点，而是从听过到能教别人的层级。\\n\\nVisible text must be exactly these lines, with no extra words:\\n听过\\n理解\\n模仿\\n独立完成\\n教别人\\n\\nLayout:\\n竖向升级阶梯。主角从底层听讲，逐步走到顶层教别人，每层用一个简短动作表示。\\n\\nVisual style:\\n升级闯关式漫画阶梯，层级清楚，顶层有聚光效果。\\n\\nAdditional image direction:\\nA manga-style vertical staircase with five labeled levels in Chinese: 听过, 理解, 模仿, 独立完成, 教别人. The same learner character climbs upward: hearing a term, explaining it, following a tutorial, completing a task alone, then teaching a colleague. Keep text sparse, use bold labels, action poses, and clean panel-like separators.\\n\\nRequirements:\\n- Canvas: 1536x864 landscape presentation slide.\\n- Generate one slide only, not a contact sheet and not multiple slide candidates.\\n- Make it presentation-ready, high-sensibility, and coherent with the chosen style template.\\n- Make it visually finished, not a sparse wireframe or placeholder diagram.\\n- Use the strongest signature elements of the style template: palette, typography, composition language, icon/illustration treatment, and editorial rhythm.\\n- Keep all visible text readable at presentation distance.\\n- Use low text density, strong hierarchy, and source-grounded diagrams, editorial composition, or template-specific visuals.\\n- Do not add markdown symbols, fake logos, watermarks, signatures, UI browser chrome, or extra captions.\\n- Match the source/user instruction language for all visible slide text and speaker-facing copy.\",\n      \"slideNumber\": 2,\n      \"visibleText\": [\n        \"听过\",\n        \"理解\",\n        \"模仿\",\n        \"独立完成\",\n        \"教别人\"\n      ],\n      \"visualStyle\": \"升级闯关式漫画阶梯，层级清楚，顶层有聚光效果。\",\n      \"speakerNotes\": \"很多人会把听过、看过误认为自己会了。材料把能力分成五层：听过、理解、模仿、独立完成、教别人。判断自己会不会，不要只问我学过吗，而要看能否不用资料独立完成，换场景还能不能用，出错时能不能排查，能不能教给别人，结果是否稳定可靠。\",\n      \"revisedPrompt\": \"Create one complete presentation slide image, 16:9 landscape, 1536x864, in a polished manga-style NotebookLM/Kael-inspired editorial illustration. Dominant art direction: understanding becomes deeper with fun; turn the information into a comic-like visual explanation with a simple story; use panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions; relate the idea to a realistic workplace learning situation; fun and explanatory, not childish or chaotic.\\n\\nSlide concept: a vertical upgrade staircase showing the progression from hearing a term to teaching it. The SAME learner character climbs upward through five clear levels, each level represented as a step/panel with a distinct action pose. Composition should feel like a clean manga challenge ladder with strong hierarchy and panel-like separators, dynamic angle, and a top spotlight at the highest level.\\n\\nVisible text must be EXACTLY these five Chinese labels and nothing else anywhere in the slide:\\n听过\\n理解\\n模仿\\n独立完成\\n教别人\\n\\nLayout details: a bold vertical staircase rising from bottom-left to top-right or straight upward within the slide, with five large steps. Each step contains the learner in a different state: bottom hears a term with a small speech bubble-like reaction icon (no text besides the label); next explains/understands with a thoughtful pose; next follows a tutorial with hands and eyes focused; next completes a task alone with confidence; top teaches a colleague under a bright spotlight. Use clean manga framing around each level, speed lines indicating upward progress, expressive facial expressions, and subtle office-learning context such as a desk, laptop, sticky notes, document, whiteboard, and colleague silhouette. Keep the slide elegant and readable at presentation distance. Use bold Chinese labels placed clearly on each step, large enough to read.\\n\\nColor and rendering: manga style with crisp black ink lines, high-contrast shading, tasteful limited palette, editorial composition, subtle warm accent colors and blue-gray shadows, white/light background with dynamic diagonal shapes, polished presentation-ready finish. Avoid clutter, avoid extra captions, avoid any text beyond the five labels, avoid logos, watermarks, UI, or decorative paragraphs. Make it look like a finished single keynote slide, not a wireframe or contact sheet.\",\n      \"storageFileId\": \"pdusfddf92up1ielh2rerlng-f2ea5ed653c402ef3317366352e59a05\"\n    },\n    {\n      \"size\": \"1536x864\",\n      \"model\": \"gpt-image-2\",\n      \"title\": \"能力必须有证据\",\n      \"layout\": \"桌面俯视漫画。一本打开的能力清单笔记本占主体，四个栏目像表格贴纸一样排布。\",\n      \"status\": \"completed\",\n      \"message\": \"不要写“我会”，要写“我做成过什么”。\",\n      \"imageUrl\": \"https://cdn.drillso.com/generated/drill/c9nu33ayugtpvkkfey1cxq9g-442834a751b795f701ca120cf9eff850.png\",\n      \"storageKey\": \"generated/drill/c9nu33ayugtpvkkfey1cxq9g-442834a751b795f701ca120cf9eff850.png\",\n      \"imagePrompt\": \"Create exactly one complete presentation slide image.\\n\\nUse the following NotebookLM / Kael style prompt as the dominant art direction. Follow it closely.\\n\\nSTYLE TEMPLATE START\\nUnderstanding becomes deeper with fun.\\nTurn the information into a comic-like visual explanation with a simple story.\\nUse panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions.\\nRelate the idea to a realistic situation so the audience can remember it.\\nThe style should be fun and explanatory, not childish or chaotic.\\nSTYLE TEMPLATE END\\n\\nDeck title: 把学习变成能力系统\\nDeck summary: 面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\\nDeck format: presenter_slides\\nChosen style: Manga Style\\n\\nSlide 3: 能力必须有证据\\nMain message: 不要写“我会”，要写“我做成过什么”。\\n\\nVisible text must be exactly these lines, with no extra words:\\n当前水平\\n证据\\n问题\\n下一步\\n\\nLayout:\\n桌面俯视漫画。一本打开的能力清单笔记本占主体，四个栏目像表格贴纸一样排布。\\n\\nVisual style:\\n手账式漫画信息图，证据栏用放大镜强调，旁边有主角认真记录。\\n\\nAdditional image direction:\\nTop-down manga desk scene with an open notebook titled 能力清单. Four large clean fields are visible: 当前水平, 证据, 问题, 下一步. A magnifying glass highlights the 证据 field. Around the notebook, small icons represent writing, Excel, communication, project delivery, and AI tools. Use clear hierarchy and minimal Chinese text.\\n\\nRequirements:\\n- Canvas: 1536x864 landscape presentation slide.\\n- Generate one slide only, not a contact sheet and not multiple slide candidates.\\n- Make it presentation-ready, high-sensibility, and coherent with the chosen style template.\\n- Make it visually finished, not a sparse wireframe or placeholder diagram.\\n- Use the strongest signature elements of the style template: palette, typography, composition language, icon/illustration treatment, and editorial rhythm.\\n- Keep all visible text readable at presentation distance.\\n- Use low text density, strong hierarchy, and source-grounded diagrams, editorial composition, or template-specific visuals.\\n- Do not add markdown symbols, fake logos, watermarks, signatures, UI browser chrome, or extra captions.\\n- Match the source/user instruction language for all visible slide text and speaker-facing copy.\",\n      \"slideNumber\": 3,\n      \"visibleText\": [\n        \"当前水平\",\n        \"证据\",\n        \"问题\",\n        \"下一步\"\n      ],\n      \"visualStyle\": \"手账式漫画信息图，证据栏用放大镜强调，旁边有主角认真记录。\",\n      \"speakerNotes\": \"盘点能力时，要建立能力清单，把笼统能力拆成具体项目。材料建议可以按专业技能、通用能力、工具能力、业务理解、交付能力来拆。关键是能力必须有证据。不要只写我会演讲，而要写能在 20 人会议上做 10 分钟汇报，并回答 3 个问题。\",\n      \"revisedPrompt\": \"Create exactly one complete presentation slide image, 1536x864 landscape, in a NotebookLM / Kael inspired manga style with a fun explanatory comic-like visual language. Top-down desk scene, editorial and polished, not childish, not chaotic. Dominant art direction: understanding becomes deeper with fun; simple story, panels, character reactions, manga framing, speed lines, expressive close-ups, concise captions. \\n\\nSlide 3 topic: 能力必须有证据. Main message should be visually communicated: 不要写“我会”，要写“我做成过什么”。 \\n\\nVisible text must be exactly these lines and no other text anywhere:\\n当前水平\\n证据\\n问题\\n下一步\\n\\nLayout: desktop top-down manga infographic. An open notebook dominates the center and is titled 能力清单 in a clean handwritten manga style. Inside the notebook, four large clean field areas are arranged like sticker-like table sections, with the visible text exactly: 当前水平, 证据, 问题, 下一步. The 证据 field is emphasized by a magnifying glass placed over it, with strong graphic focus, highlight ring, and subtle speed lines. A thoughtful protagonist hand/character is nearby, seriously noting things down. Around the notebook, small icons illustrate writing, Excel, communication, project delivery, and AI tools, arranged neatly like desk items. \\n\\nArt direction: hand-drawn manga editorial infographic, clean composition, strong hierarchy, minimal Chinese text, presentation-ready, coherent, high sensibility, realistic desk objects combined with comic elements. Use palette and typography inspired by Japanese manga notes: off-white paper, warm beige desk, black ink lines, muted teal and orange accents, clean shadows, crisp legibility. Add subtle panel-like framing and expressive close-up feel without splitting into multiple slide candidates. Make the evidence area the visual focal point. \\n\\nImportant constraints: no extra captions, no watermarks, no logos, no browser chrome, no UI mockups, no markdown symbols, no extra words beyond the four required visible text lines; ensure those text lines are readable at presentation distance. The notebook title 能力清单 can appear as part of the notebook cover/title. Make the slide feel finished and coherent, not sparse or placeholder-like.\",\n      \"storageFileId\": \"c9nu33ayugtpvkkfey1cxq9g-442834a751b795f701ca120cf9eff850\"\n    },\n    {\n      \"size\": \"1536x864\",\n      \"model\": \"gpt-image-2\",\n      \"title\": \"不会的地方，藏在任务里\",\n      \"layout\": \"四宫格侦探漫画。主角像侦探一样寻找盲区，每格对应一种方法。\",\n      \"status\": \"completed\",\n      \"message\": \"真正的差距，会在真实任务中暴露出来。\",\n      \"imageUrl\": \"https://cdn.drillso.com/generated/drill/h1er9ut8gu1bakkopkd0flmk-fdf0f5f5eeffbbff9435605789265fb2.png\",\n      \"storageKey\": \"generated/drill/h1er9ut8gu1bakkopkd0flmk-fdf0f5f5eeffbbff9435605789265fb2.png\",\n      \"imagePrompt\": \"Create exactly one complete presentation slide image.\\n\\nUse the following NotebookLM / Kael style prompt as the dominant art direction. Follow it closely.\\n\\nSTYLE TEMPLATE START\\nUnderstanding becomes deeper with fun.\\nTurn the information into a comic-like visual explanation with a simple story.\\nUse panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions.\\nRelate the idea to a realistic situation so the audience can remember it.\\nThe style should be fun and explanatory, not childish or chaotic.\\nSTYLE TEMPLATE END\\n\\nDeck title: 把学习变成能力系统\\nDeck summary: 面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\\nDeck format: presenter_slides\\nChosen style: Manga Style\\n\\nSlide 4: 不会的地方，藏在任务里\\nMain message: 真正的差距，会在真实任务中暴露出来。\\n\\nVisible text must be exactly these lines, with no extra words:\\n对标\\n任务\\n提问\\n反馈\\n\\nLayout:\\n四宫格侦探漫画。主角像侦探一样寻找盲区，每格对应一种方法。\\n\\nVisual style:\\n侦探调查感，放大镜、脚印、问号和评分牌作为符号。\\n\\nAdditional image direction:\\nA four-panel manga detective layout. The learner wears a simple detective cap and investigates four clues labeled 对标, 任务, 提问, 反馈. Panel 1 compares their work with an expert article. Panel 2 shows a sales data task with a stuck expression. Panel 3 shows question marks around core concepts. Panel 4 shows a colleague giving a score card. Keep labels large and text minimal.\\n\\nRequirements:\\n- Canvas: 1536x864 landscape presentation slide.\\n- Generate one slide only, not a contact sheet and not multiple slide candidates.\\n- Make it presentation-ready, high-sensibility, and coherent with the chosen style template.\\n- Make it visually finished, not a sparse wireframe or placeholder diagram.\\n- Use the strongest signature elements of the style template: palette, typography, composition language, icon/illustration treatment, and editorial rhythm.\\n- Keep all visible text readable at presentation distance.\\n- Use low text density, strong hierarchy, and source-grounded diagrams, editorial composition, or template-specific visuals.\\n- Do not add markdown symbols, fake logos, watermarks, signatures, UI browser chrome, or extra captions.\\n- Match the source/user instruction language for all visible slide text and speaker-facing copy.\",\n      \"slideNumber\": 4,\n      \"visibleText\": [\n        \"对标\",\n        \"任务\",\n        \"提问\",\n        \"反馈\"\n      ],\n      \"visualStyle\": \"侦探调查感，放大镜、脚印、问号和评分牌作为符号。\",\n      \"speakerNotes\": \"识别不会什么，可以用四种方法。对标法是看高手的成果、流程和思考方式。任务法是用真实任务测试，比如拿销售数据找出下降原因并提出建议。提问法是用核心概念、关系、异常处理等问题暴露盲区。反馈法是请同事、老师或朋友指出哪里好、哪里不清楚、哪里效率低。\",\n      \"revisedPrompt\": \"Create exactly one complete presentation slide image, landscape 1536x864, in a polished Manga Style / NotebookLM / Kael style. Dominant art direction: understanding becomes deeper with fun; comic-like explanatory visual with a simple story, panels, character reactions, manga framing, speed lines, expressive close-ups, concise captions. Serious, fun, explanatory, not childish. \\n\\nDeck slide topic: \\\"不会的地方，藏在任务里\\\" with main message: truly real gaps are exposed in real tasks.\\n\\nDesign a single finished presentation slide, not a contact sheet, not multiple candidate slides. Four-panel detective manga layout, coherent editorial composition. The learner is a detective-like main character wearing a simple detective cap, investigating four clues. Visual style: detective investigation vibe, magnifying glass, footprints, question marks, and score card symbols. Include a realistic office/work-learning setting. \\n\\nVisible text must be exactly these four Chinese lines and nothing else, large and readable, each as a label in one panel: \\n对标\\n任务\\n提问\\n反馈\\n\\nNo other visible text anywhere. No title, no subtitle, no extra captions, no watermark, no UI. Keep text minimal and clean.\\n\\nPanel 1: learner compares their work with an expert article / benchmark material, using a magnifying glass over a document, subtle comparison cues. Label: 对标.\\nPanel 2: learner confronts a sales data task, stuck expression, evidence of uncertainty, files/spreadsheet vibe, detective examining task clues. Label: 任务.\\nPanel 3: question marks swirl around core concepts, learner thinking and asking precise questions, manga close-up reaction. Label: 提问.\\nPanel 4: colleague gives a score card / feedback card, learner receiving constructive feedback, modestly relieved expression. Label: 反馈.\\n\\nUse a four-panel comic grid with strong manga framing, varied camera angles, expressive faces, speed lines, inked outlines, halftone shading, dynamic but clean composition. Presentation-ready, high-sensibility, visually finished. Use a restrained palette: off-white paper, black ink, muted teal, warm orange accents, and red highlights for emphasis. Ensure the four labels are clearly legible and the overall slide feels cohesive and professional.\",\n      \"storageFileId\": \"h1er9ut8gu1bakkopkd0flmk-fdf0f5f5eeffbbff9435605789265fb2\"\n    },\n    {\n      \"size\": \"1536x864\",\n      \"model\": \"gpt-image-2\",\n      \"title\": \"该学什么，从目标倒推\",\n      \"layout\": \"横向流程漫画。主角从左到右推着一个目标牌，经过六个节点，最后进入真实应用场景。\",\n      \"status\": \"completed\",\n      \"message\": \"不要被热门牵着走，要从真实目标拆出关键差距。\",\n      \"imageUrl\": \"https://cdn.drillso.com/generated/drill/pigjfvkup3iucez8bovjfolb-8df686416354c784ce6a5d2632631f54.png\",\n      \"storageKey\": \"generated/drill/pigjfvkup3iucez8bovjfolb-8df686416354c784ce6a5d2632631f54.png\",\n      \"imagePrompt\": \"Create exactly one complete presentation slide image.\\n\\nUse the following NotebookLM / Kael style prompt as the dominant art direction. Follow it closely.\\n\\nSTYLE TEMPLATE START\\nUnderstanding becomes deeper with fun.\\nTurn the information into a comic-like visual explanation with a simple story.\\nUse panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions.\\nRelate the idea to a realistic situation so the audience can remember it.\\nThe style should be fun and explanatory, not childish or chaotic.\\nSTYLE TEMPLATE END\\n\\nDeck title: 把学习变成能力系统\\nDeck summary: 面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\\nDeck format: presenter_slides\\nChosen style: Manga Style\\n\\nSlide 5: 该学什么，从目标倒推\\nMain message: 不要被热门牵着走，要从真实目标拆出关键差距。\\n\\nVisible text must be exactly these lines, with no extra words:\\n目标\\n任务\\n能力\\n差距\\n计划\\n应用\\n\\nLayout:\\n横向流程漫画。主角从左到右推着一个目标牌，经过六个节点，最后进入真实应用场景。\\n\\nVisual style:\\n流程图和漫画动作结合，节点像游戏关卡，箭头清晰。\\n\\nAdditional image direction:\\nA horizontal manga flow diagram with the learner pushing a sign labeled 目标 through six connected nodes: 目标, 任务, 能力, 差距, 计划, 应用. Avoid dense text. The background hints at a workplace project report scenario. Use dynamic arrows, speed lines, and clear game-level checkpoints.\\n\\nRequirements:\\n- Canvas: 1536x864 landscape presentation slide.\\n- Generate one slide only, not a contact sheet and not multiple slide candidates.\\n- Make it presentation-ready, high-sensibility, and coherent with the chosen style template.\\n- Make it visually finished, not a sparse wireframe or placeholder diagram.\\n- Use the strongest signature elements of the style template: palette, typography, composition language, icon/illustration treatment, and editorial rhythm.\\n- Keep all visible text readable at presentation distance.\\n- Use low text density, strong hierarchy, and source-grounded diagrams, editorial composition, or template-specific visuals.\\n- Do not add markdown symbols, fake logos, watermarks, signatures, UI browser chrome, or extra captions.\\n- Match the source/user instruction language for all visible slide text and speaker-facing copy.\",\n      \"slideNumber\": 5,\n      \"visibleText\": [\n        \"目标\",\n        \"任务\",\n        \"能力\",\n        \"差距\",\n        \"计划\",\n        \"应用\"\n      ],\n      \"visualStyle\": \"流程图和漫画动作结合，节点像游戏关卡，箭头清晰。\",\n      \"speakerNotes\": \"材料指出，常见误区是看到什么热门就学什么，今天学 AI，明天学剪辑，后天学投资，最后没有形成能力闭环。更有效的方法是从目标倒推：明确目标，拆解任务，识别所需能力，评估当前水平，找出关键差距，制定学习计划，再真实应用并复盘改进。\",\n      \"revisedPrompt\": \"Create a single presentation slide image, 1536x864 landscape, in a manga-style NotebookLM / Kael editorial comic format. Dominant art direction: understanding becomes deeper with fun; turn the information into a comic-like visual explanation with a simple story; use panels, expressive reactions, manga framing, speed lines, concise captions, but keep it clean, fun, explanatory, not childish or chaotic.\\n\\nSlide topic: '该学什么，从目标倒推'. Main message:不要被热门牵着走，要从真实目标拆出关键差距.\\n\\nComposition: a horizontal manga flow diagram from left to right. A learner character pushes a large sign labeled exactly '目标' through six connected game-like checkpoint nodes, then into a realistic workplace project report scene at the far right. The six nodes must be clearly visible and the ONLY visible text on the slide must be these exact six lines, each appearing once and nothing else:\\n目标\\n任务\\n能力\\n差距\\n计划\\n应用\\n\\nLayout details: The learner begins on the left with determined expression, pushing the '目标' sign forward through connected nodes like level checkpoints. Each node is a stylized rounded rectangle or game stage badge with manga framing and dynamic arrows. Show the flow progression: 目标 → 任务 → 能力 → 差距 → 计划 → 应用. Use clear arrows, speed lines, motion effects, and expressive close-ups of the learner at a few stages. At the rightmost end, the flow enters a realistic workplace project report application setting: desk, laptop, charts, meeting board, documents, office environment hinted in the background, showing real-world use. Background should subtly hint at a work project report scenario without adding extra text.\\n\\nVisual style: manga style with editorial composition, bold but tasteful black linework, vivid yet restrained colors, strong hierarchy, game-level checkpoint visuals, clean readable typography for the Chinese node labels, presentation-ready, polished, high contrast, coherent, no clutter, no extra words, no watermark, no logo, no browser chrome. Ensure text is legible at presentation distance. Keep low text density and strong visual rhythm. Make it look like a finished slide, not a wireframe. Avoid any additional captions or labels beyond the six required words.\",\n      \"storageFileId\": \"pigjfvkup3iucez8bovjfolb-8df686416354c784ce6a5d2632631f54\"\n    },\n    {\n      \"size\": \"1536x864\",\n      \"model\": \"gpt-image-2\",\n      \"title\": \"优先级，看三件事\",\n      \"layout\": \"漫画仪表盘。三个大旋钮分别代表重要性、紧急性、杠杆性，主角把学习项目卡片放入评分机。\",\n      \"status\": \"completed\",\n      \"message\": \"不是所有不会的东西都要马上学。\",\n      \"imageUrl\": \"https://cdn.drillso.com/generated/drill/bskvd1s25xzz10oiu38tsxh7-e01f9c917a2b1fcce79a9a65e6f5fc81.png\",\n      \"storageKey\": \"generated/drill/bskvd1s25xzz10oiu38tsxh7-e01f9c917a2b1fcce79a9a65e6f5fc81.png\",\n      \"imagePrompt\": \"Create exactly one complete presentation slide image.\\n\\nUse the following NotebookLM / Kael style prompt as the dominant art direction. Follow it closely.\\n\\nSTYLE TEMPLATE START\\nUnderstanding becomes deeper with fun.\\nTurn the information into a comic-like visual explanation with a simple story.\\nUse panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions.\\nRelate the idea to a realistic situation so the audience can remember it.\\nThe style should be fun and explanatory, not childish or chaotic.\\nSTYLE TEMPLATE END\\n\\nDeck title: 把学习变成能力系统\\nDeck summary: 面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\\nDeck format: presenter_slides\\nChosen style: Manga Style\\n\\nSlide 6: 优先级，看三件事\\nMain message: 不是所有不会的东西都要马上学。\\n\\nVisible text must be exactly these lines, with no extra words:\\n重要性\\n紧急性\\n杠杆性\\n\\nLayout:\\n漫画仪表盘。三个大旋钮分别代表重要性、紧急性、杠杆性，主角把学习项目卡片放入评分机。\\n\\nVisual style:\\n轻量信息仪表盘，三个维度占画面中心，项目卡片作为辅助元素。\\n\\nAdditional image direction:\\nA manga-style priority scoring machine. Three large dials are labeled 重要性, 紧急性, 杠杆性. The learner feeds skill cards into the machine, such as Excel 数据清洗 and 汇报表达, while low-priority cards sit to the side. Use bold dial shapes, minimal visible text, and a clear sense of ranking without adding extra numbers.\\n\\nRequirements:\\n- Canvas: 1536x864 landscape presentation slide.\\n- Generate one slide only, not a contact sheet and not multiple slide candidates.\\n- Make it presentation-ready, high-sensibility, and coherent with the chosen style template.\\n- Make it visually finished, not a sparse wireframe or placeholder diagram.\\n- Use the strongest signature elements of the style template: palette, typography, composition language, icon/illustration treatment, and editorial rhythm.\\n- Keep all visible text readable at presentation distance.\\n- Use low text density, strong hierarchy, and source-grounded diagrams, editorial composition, or template-specific visuals.\\n- Do not add markdown symbols, fake logos, watermarks, signatures, UI browser chrome, or extra captions.\\n- Match the source/user instruction language for all visible slide text and speaker-facing copy.\",\n      \"slideNumber\": 6,\n      \"visibleText\": [\n        \"重要性\",\n        \"紧急性\",\n        \"杠杆性\"\n      ],\n      \"visualStyle\": \"轻量信息仪表盘，三个维度占画面中心，项目卡片作为辅助元素。\",\n      \"speakerNotes\": \"判断学习优先级，可以用三维评分法。第一是重要性：是否影响核心目标。第二是紧急性：近期是否马上会用到。第三是杠杆性：学会后是否能带来长期收益。总分越高，越应该优先学。材料中的例子显示，Excel 数据清洗和汇报表达因为得分高，更应该优先学习。\",\n      \"revisedPrompt\": \"Create exactly one complete presentation slide image, landscape 1536x864, in a manga-style NotebookLM / Kael explanatory comic style. Theme: priority scoring machine for learning choices. The composition is a polished presentation slide, not a wireframe, not a contact sheet, not multiple panels. Use a fun, explanatory, editorial manga look with bold shapes, expressive reactions, speed lines, clean panel framing hints, and a realistic office-learning situation.\\n\\nScene: A learner at a desk feeds skill cards into a large priority-scoring machine in the center of the slide. The machine has THREE huge bold dials taking up the visual center, each clearly labeled with exactly one of the following Chinese words only: 重要性, 紧急性, 杠杆性. These are the ONLY visible words anywhere in the slide. No title, no subtitle, no extra captions, no numbers, no labels on cards, no watermarks, no UI chrome.\\n\\nVisual structure: Center the three dials prominently like a comic dashboard. The learner character is in a clear manga pose, slightly tense but thoughtful, inserting cards into the machine. The cards are symbolic only, with simple icon-like visuals to suggest work skills, but NO text on the cards. One or two low-priority cards sit aside in a dimmer area, also without text. Show the machine ranking the cards by visual emphasis: one card is highlighted going into the machine, another sits aside in muted tone. Use strong hierarchy and a sense of sorting/decision-making.\\n\\nArt direction: Light information dashboard, manga framing, expressive close-up feel, clean and presentation-ready. High contrast, crisp line art, modern comic shading, subtle motion lines, dynamic but not chaotic. Color palette: warm cream background, ink black lines, muted teal and orange accent colors, with bright highlights for the active card and dials. Keep it elegant, educational, and easy to read at presentation distance.\\n\\nImportant constraints: Only the three Chinese dial labels may appear as text. Absolutely no other visible text. Do not add extra words in any language. Do not include logos, signatures, or watermarks. Make it visually finished and coherent, with a strong sense of ranking and prioritization.\",\n      \"storageFileId\": \"bskvd1s25xzz10oiu38tsxh7-e01f9c917a2b1fcce79a9a65e6f5fc81\"\n    },\n    {\n      \"size\": \"1536x864\",\n      \"model\": \"gpt-image-2\",\n      \"title\": \"有效学习，是一个闭环\",\n      \"layout\": \"环形闭环漫画图。四个节点围成循环，中间是一个逐渐成形的作品。\",\n      \"status\": \"completed\",\n      \"message\": \"只看资料不够，要让知识变成成果。\",\n      \"imageUrl\": \"https://cdn.drillso.com/generated/drill/x194b29yzfoalt2z4pfn81mp-b0d8e1614c3cb6dfd8a24f12a5b17f98.png\",\n      \"storageKey\": \"generated/drill/x194b29yzfoalt2z4pfn81mp-b0d8e1614c3cb6dfd8a24f12a5b17f98.png\",\n      \"imagePrompt\": \"Create exactly one complete presentation slide image.\\n\\nUse the following NotebookLM / Kael style prompt as the dominant art direction. Follow it closely.\\n\\nSTYLE TEMPLATE START\\nUnderstanding becomes deeper with fun.\\nTurn the information into a comic-like visual explanation with a simple story.\\nUse panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions.\\nRelate the idea to a realistic situation so the audience can remember it.\\nThe style should be fun and explanatory, not childish or chaotic.\\nSTYLE TEMPLATE END\\n\\nDeck title: 把学习变成能力系统\\nDeck summary: 面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\\nDeck format: presenter_slides\\nChosen style: Manga Style\\n\\nSlide 7: 有效学习，是一个闭环\\nMain message: 只看资料不够，要让知识变成成果。\\n\\nVisible text must be exactly these lines, with no extra words:\\n输入\\n练习\\n反馈\\n输出\\n\\nLayout:\\n环形闭环漫画图。四个节点围成循环，中间是一个逐渐成形的作品。\\n\\nVisual style:\\n循环动作分镜，中心作品发光，强调从资料到成果的转化。\\n\\nAdditional image direction:\\nA circular manga infographic with four connected nodes: 输入, 练习, 反馈, 输出. In the center, a rough idea transforms into a finished work artifact, such as a report or project board. Show books and courses at input, pen and keyboard at practice, comment bubbles at feedback, and a completed deliverable at output. Keep text sparse and hierarchy strong.\\n\\nRequirements:\\n- Canvas: 1536x864 landscape presentation slide.\\n- Generate one slide only, not a contact sheet and not multiple slide candidates.\\n- Make it presentation-ready, high-sensibility, and coherent with the chosen style template.\\n- Make it visually finished, not a sparse wireframe or placeholder diagram.\\n- Use the strongest signature elements of the style template: palette, typography, composition language, icon/illustration treatment, and editorial rhythm.\\n- Keep all visible text readable at presentation distance.\\n- Use low text density, strong hierarchy, and source-grounded diagrams, editorial composition, or template-specific visuals.\\n- Do not add markdown symbols, fake logos, watermarks, signatures, UI browser chrome, or extra captions.\\n- Match the source/user instruction language for all visible slide text and speaker-facing copy.\",\n      \"slideNumber\": 7,\n      \"visibleText\": [\n        \"输入\",\n        \"练习\",\n        \"反馈\",\n        \"输出\"\n      ],\n      \"visualStyle\": \"循环动作分镜，中心作品发光，强调从资料到成果的转化。\",\n      \"speakerNotes\": \"材料把有效学习概括为输入、练习、反馈、输出的闭环。输入包括看书、听课、读文章、观察案例、请教他人，但要有边界，先学够用的最小知识。练习是把知识变成动作。反馈用来发现偏差。输出是最可靠的检验方式，可以是一篇文章、一个项目、一次汇报、一个工具或一个真实问题的解决方案。\",\n      \"revisedPrompt\": \"Create exactly one complete presentation slide image in 16:9 landscape, 1536x864. Dominant art direction: NotebookLM / Kael style, manga-style explanatory comic infographic, fun but polished, not childish. Strong editorial composition, clear hierarchy, low text density, high sensibility, presentation-ready.\\n\\nSlide concept: a circular closed-loop learning system. Four connected nodes arranged in a ring around a glowing center. Center shows a rough idea transforming into a finished work artifact, like a polished report or project board, with a bright glow and subtle motion lines to show transformation from vague draft to completed output.\\n\\nVisual structure:\\n- Circular loop / cycle infographic in the middle of the slide.\\n- Four manga comic nodes evenly spaced around the circle with arrows or curved connectors forming a loop.\\n- Node 1: books, course cards, and reading materials to represent input.\\n- Node 2: pen, notebook, and keyboard to represent practice.\\n- Node 3: comment bubbles, review marks, and mentor-like feedback to represent feedback.\\n- Node 4: completed deliverable, clean report/project board to represent output.\\n- Center: an in-progress idea becoming a finished artifact, luminous, with a subtle burst effect.\\n\\nStyle details:\\n- Manga framing, expressive motion lines, dynamic but clean composition.\\n- Use comic panel energy within a single infographic, with connected vignette-like nodes.\\n- Palette: high-contrast editorial manga colors, mostly warm cream background with black inks and selective accent colors like teal, orange, and gold glow.\\n- Use crisp, readable Chinese typography integrated into the node labels.\\n- Keep the slide uncluttered and visually cohesive.\\n\\nVisible text must be exactly these four lines and nothing else:\\n输入\\n练习\\n反馈\\n输出\\n\\nNo extra text, no title, no footer, no watermark, no UI chrome, no logos, no captions beyond the four required labels. Make the four labels clearly readable and the only text on the slide.\",\n      \"storageFileId\": \"x194b29yzfoalt2z4pfn81mp-b0d8e1614c3cb6dfd8a24f12a5b17f98\"\n    },\n    {\n      \"size\": \"1536x864\",\n      \"model\": \"gpt-image-2\",\n      \"title\": \"把学习绑到真实场景\",\n      \"layout\": \"三段式漫画。左边是问题，中间是作品，右边是用户在场景中使用作品。\",\n      \"status\": \"completed\",\n      \"message\": \"应用不是记笔记，而是解决一个看得见的问题。\",\n      \"imageUrl\": \"https://cdn.drillso.com/generated/drill/e7wzw8m3ktcode3m8xg3p68s-a566d46672a8045db28f95956b376cfc.png\",\n      \"storageKey\": \"generated/drill/e7wzw8m3ktcode3m8xg3p68s-a566d46672a8045db28f95956b376cfc.png\",\n      \"imagePrompt\": \"Create exactly one complete presentation slide image.\\n\\nUse the following NotebookLM / Kael style prompt as the dominant art direction. Follow it closely.\\n\\nSTYLE TEMPLATE START\\nUnderstanding becomes deeper with fun.\\nTurn the information into a comic-like visual explanation with a simple story.\\nUse panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions.\\nRelate the idea to a realistic situation so the audience can remember it.\\nThe style should be fun and explanatory, not childish or chaotic.\\nSTYLE TEMPLATE END\\n\\nDeck title: 把学习变成能力系统\\nDeck summary: 面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\\nDeck format: presenter_slides\\nChosen style: Manga Style\\n\\nSlide 8: 把学习绑到真实场景\\nMain message: 应用不是记笔记，而是解决一个看得见的问题。\\n\\nVisible text must be exactly these lines, with no extra words:\\n真实问题\\n具体作品\\n使用场景\\n\\nLayout:\\n三段式漫画。左边是问题，中间是作品，右边是用户在场景中使用作品。\\n\\nVisual style:\\n产品小剧场感，三个绑定像三张卡片，用线连接成因果链。\\n\\nAdditional image direction:\\nA three-part manga scene connected by arrows. Card 1 labeled 真实问题 shows messy weekly sales data. Card 2 labeled 具体作品 shows an automation script or clean report. Card 3 labeled 使用场景 shows someone using the report in a meeting. Include a small AI tool icon in the background but no brand logos. Minimal Chinese text, expressive characters, clear cause-and-effect.\\n\\nRequirements:\\n- Canvas: 1536x864 landscape presentation slide.\\n- Generate one slide only, not a contact sheet and not multiple slide candidates.\\n- Make it presentation-ready, high-sensibility, and coherent with the chosen style template.\\n- Make it visually finished, not a sparse wireframe or placeholder diagram.\\n- Use the strongest signature elements of the style template: palette, typography, composition language, icon/illustration treatment, and editorial rhythm.\\n- Keep all visible text readable at presentation distance.\\n- Use low text density, strong hierarchy, and source-grounded diagrams, editorial composition, or template-specific visuals.\\n- Do not add markdown symbols, fake logos, watermarks, signatures, UI browser chrome, or extra captions.\\n- Match the source/user instruction language for all visible slide text and speaker-facing copy.\",\n      \"slideNumber\": 8,\n      \"visibleText\": [\n        \"真实问题\",\n        \"具体作品\",\n        \"使用场景\"\n      ],\n      \"visualStyle\": \"产品小剧场感，三个绑定像三张卡片，用线连接成因果链。\",\n      \"speakerNotes\": \"学完不用，很快会忘。材料建议应用能力时做三个绑定：绑定一个真实问题，绑定一个具体作品，绑定一个使用场景。比如不要只说我要学 Python，而是说我要用 Python 自动整理每周销售数据，减少手工复制时间。学习 AI 工具也不要只研究提示词技巧，而要放到写周报、做调研、学英语或写代码等具体场景里。\",\n      \"revisedPrompt\": \"Create exactly one complete presentation slide image in 1536x864 landscape, manga-style educational editorial slide, inspired by NotebookLM / Kael style: fun explanatory comic with simple story, panels, character reactions, manga framing, speed lines, expressive close-ups, concise captions, polished and presentation-ready. The slide topic is: 把学习绑到真实场景. Main message visualized: 应用不是记笔记，而是解决一个看得见的问题.\\n\\nLayout: three-part manga scene arranged left to right as three connected cards, with arrows/line connectors showing a cause-and-effect chain. Product theater feeling, like three linked cards. Left card = messy weekly sales problem, middle card = clean output/automation script/report, right card = person using the report in a meeting. Add a small AI tool icon in the background, abstract and generic, no brand logos.\\n\\nVisible text must be exactly these three lines and nothing else anywhere on the slide:\\n真实问题\\n具体作品\\n使用场景\\n\\nPlace the text as the card headings only, large, readable, and centered or top-left on each card. No other Chinese or English text, no extra captions, no watermarks, no UI chrome, no logo.\\n\\nArt direction: manga-style but not childish or chaotic; modern editorial palette with strong contrast, clean composition, subtle halftone textures, speed lines, expressive character faces. Left card should show a realistic messy weekly sales data situation: scattered numbers, charts, papers, and a stressed office worker. Middle card should show a neat artifact: a clean report dashboard or automation script flowing into a tidy summary sheet, with a focused character producing it. Right card should show a confident person presenting/using the report in a meeting, with colleagues reacting positively. Use arrows connecting the cards. Keep low text density, strong hierarchy, and clear cause-and-effect. Make it visually finished, balanced, and high-sensibility, suitable for a deck slide.\\n\\nDo not include any other visible text beyond the exact three headings listed above.\",\n      \"storageFileId\": \"e7wzw8m3ktcode3m8xg3p68s-a566d46672a8045db28f95956b376cfc\"\n    },\n    {\n      \"size\": \"1536x864\",\n      \"model\": \"gpt-image-2\",\n      \"title\": \"每周 30 分钟，复盘能力边界\",\n      \"layout\": \"日历和复盘表漫画。主角周末坐下复盘，四个问题以便签形式围绕日历。\",\n      \"status\": \"completed\",\n      \"message\": \"持续复盘，才能看清自己能独立完成什么、在哪些地方卡住。\",\n      \"imageUrl\": \"https://cdn.drillso.com/generated/drill/qv6tiax8m7v5jri3kffciy1s-69c135520f937f8195a020ed3bfcd908.png\",\n      \"storageKey\": \"generated/drill/qv6tiax8m7v5jri3kffciy1s-69c135520f937f8195a020ed3bfcd908.png\",\n      \"imagePrompt\": \"Create exactly one complete presentation slide image.\\n\\nUse the following NotebookLM / Kael style prompt as the dominant art direction. Follow it closely.\\n\\nSTYLE TEMPLATE START\\nUnderstanding becomes deeper with fun.\\nTurn the information into a comic-like visual explanation with a simple story.\\nUse panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions.\\nRelate the idea to a realistic situation so the audience can remember it.\\nThe style should be fun and explanatory, not childish or chaotic.\\nSTYLE TEMPLATE END\\n\\nDeck title: 把学习变成能力系统\\nDeck summary: 面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\\nDeck format: presenter_slides\\nChosen style: Manga Style\\n\\nSlide 9: 每周 30 分钟，复盘能力边界\\nMain message: 持续复盘，才能看清自己能独立完成什么、在哪些地方卡住。\\n\\nVisible text must be exactly these lines, with no extra words:\\n完成了什么？\\n哪里卡住？\\n下周补什么？\\n用什么任务练？\\n\\nLayout:\\n日历和复盘表漫画。主角周末坐下复盘，四个问题以便签形式围绕日历。\\n\\nVisual style:\\n温和收束，便签式问题，日历上标出 30 分钟和 4 周节奏。\\n\\nAdditional image direction:\\nA closing manga scene of the learner doing a weekly review at a desk. A calendar shows a repeating weekly rhythm and a small 30 分钟 marker. Four sticky notes around the calendar read: 完成了什么？ 哪里卡住？ 下周补什么？ 用什么任务练？ Show calm confidence, fewer speed lines, and a clear sense of progress over four weeks.\\n\\nRequirements:\\n- Canvas: 1536x864 landscape presentation slide.\\n- Generate one slide only, not a contact sheet and not multiple slide candidates.\\n- Make it presentation-ready, high-sensibility, and coherent with the chosen style template.\\n- Make it visually finished, not a sparse wireframe or placeholder diagram.\\n- Use the strongest signature elements of the style template: palette, typography, composition language, icon/illustration treatment, and editorial rhythm.\\n- Keep all visible text readable at presentation distance.\\n- Use low text density, strong hierarchy, and source-grounded diagrams, editorial composition, or template-specific visuals.\\n- Do not add markdown symbols, fake logos, watermarks, signatures, UI browser chrome, or extra captions.\\n- Match the source/user instruction language for all visible slide text and speaker-facing copy.\",\n      \"slideNumber\": 9,\n      \"visibleText\": [\n        \"完成了什么？\",\n        \"哪里卡住？\",\n        \"下周补什么？\",\n        \"用什么任务练？\"\n      ],\n      \"visualStyle\": \"温和收束，便签式问题，日历上标出 30 分钟和 4 周节奏。\",\n      \"speakerNotes\": \"材料给出一个每周复盘模板，建议每周花 30 分钟回答：本周完成了什么具体任务，哪些能独立完成，哪些地方卡住了，卡住是因为概念不懂、方法不会还是练习太少，收到了哪些反馈，下周最应该补哪一个短板，以及用什么真实任务来练习。持续 4 周，会更清楚自己的能力边界。\",\n      \"revisedPrompt\": \"Create exactly one complete presentation slide image in 16:9 landscape (1536x864), finished and presentation-ready, in a Manga Style inspired by NotebookLM / Kael style. Dominant art direction: understanding becomes deeper with fun; turn information into a comic-like visual explanation with a simple story; use panels, character reactions, manga framing, subtle speed lines, expressive close-ups, and concise captions; relate the idea to a realistic situation so it feels memorable; fun and explanatory, not childish or chaotic.\\n\\nDeck title theme: 把学习变成能力系统. Slide 9 topic: 每周 30 分钟，复盘能力边界. Main message: 持续复盘，才能看清自己能独立完成什么、在哪些地方卡住.\\n\\nScene/layout: a closing manga scene of a learner doing a weekly review at a desk on a calm weekend evening. The learner sits at a desk with a notebook, pen, and a calendar spread. A large calendar shows a repeating weekly rhythm and a small clear marker reading 30 分钟, plus a 4-week progression cue (e.g., four small weekly boxes or four weeks highlighted) indicating the review habit over four weeks. Around the calendar, four sticky notes placed in a balanced editorial composition, each with one exact visible text line. The four sticky notes must read exactly and only these lines, one per note, with no extra visible words anywhere else on the slide: 完成了什么？  哪里卡住？  下周补什么？  用什么任务练？\\n\\nImportant text constraints: visible text must be exactly these lines, with no extra words. No other readable text anywhere on the slide besides those four lines. Do not add title text, subtitles, labels, captions, watermarks, logos, UI chrome, or extra numbers/words. The calendar can show visual markers for 30 分钟 and 4-week rhythm using icons, bars, dots, brackets, or stylized marks, but no additional readable text. The sticky notes are the only readable text.\\n\\nVisual style: warm, calm, coherent manga editorial illustration with gentle cream paper background, muted ink lines, soft orange and teal accents, subtle halftone texture, clean composition, clear hierarchy. Use a desk scene with a sense of closure and progress. Show calm confidence, fewer speed lines than energetic slides, and a clear sense of progression over four weeks. Include manga framing and soft expressive character reaction but keep it mature and tidy. Make it visually finished with polished illustration, not a wireframe. Ensure all four note texts are legible at presentation distance.\\n\\nNo markdown, no fake logos, no watermark, no extra captions, no extra text blocks.\",\n      \"storageFileId\": \"qv6tiax8m7v5jri3kffciy1s-69c135520f937f8195a020ed3bfcd908\"\n    },\n    {\n      \"size\": \"1536x864\",\n      \"model\": \"gpt-image-2\",\n      \"title\": \"一句话建立学习系统\",\n      \"layout\": \"终章海报式漫画。主角站在完成的作品墙前，背后是从目标到作品的路径。\",\n      \"status\": \"completed\",\n      \"message\": \"用真实目标发现差距，用具体任务训练能力，用反馈修正方向，用作品证明自己。\",\n      \"imageUrl\": \"https://cdn.drillso.com/generated/drill/dbor32oeat876lytpnrpzxhn-8a313acde88f3a76e1e195ae46ecc9bc.png\",\n      \"storageKey\": \"generated/drill/dbor32oeat876lytpnrpzxhn-8a313acde88f3a76e1e195ae46ecc9bc.png\",\n      \"imagePrompt\": \"Create exactly one complete presentation slide image.\\n\\nUse the following NotebookLM / Kael style prompt as the dominant art direction. Follow it closely.\\n\\nSTYLE TEMPLATE START\\nUnderstanding becomes deeper with fun.\\nTurn the information into a comic-like visual explanation with a simple story.\\nUse panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions.\\nRelate the idea to a realistic situation so the audience can remember it.\\nThe style should be fun and explanatory, not childish or chaotic.\\nSTYLE TEMPLATE END\\n\\nDeck title: 把学习变成能力系统\\nDeck summary: 面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\\nDeck format: presenter_slides\\nChosen style: Manga Style\\n\\nSlide 10: 一句话建立学习系统\\nMain message: 用真实目标发现差距，用具体任务训练能力，用反馈修正方向，用作品证明自己。\\n\\nVisible text must be exactly these lines, with no extra words:\\n目标发现差距\\n任务训练能力\\n反馈修正方向\\n作品证明自己\\n\\nLayout:\\n终章海报式漫画。主角站在完成的作品墙前，背后是从目标到作品的路径。\\n\\nVisual style:\\n强收束海报构图，四个短句像宣言贴在作品墙上，主角表情坚定。\\n\\nAdditional image direction:\\nA final manga poster composition. The learner stands confidently in front of a wall of completed works: report, project board, tool, presentation, article. Behind them, a simple path connects goal, task, feedback, and work. Four bold Chinese phrases appear as clean labels: 目标发现差距, 任务训练能力, 反馈修正方向, 作品证明自己. Make it feel like a memorable ending slide.\\n\\nRequirements:\\n- Canvas: 1536x864 landscape presentation slide.\\n- Generate one slide only, not a contact sheet and not multiple slide candidates.\\n- Make it presentation-ready, high-sensibility, and coherent with the chosen style template.\\n- Make it visually finished, not a sparse wireframe or placeholder diagram.\\n- Use the strongest signature elements of the style template: palette, typography, composition language, icon/illustration treatment, and editorial rhythm.\\n- Keep all visible text readable at presentation distance.\\n- Use low text density, strong hierarchy, and source-grounded diagrams, editorial composition, or template-specific visuals.\\n- Do not add markdown symbols, fake logos, watermarks, signatures, UI browser chrome, or extra captions.\\n- Match the source/user instruction language for all visible slide text and speaker-facing copy.\",\n      \"slideNumber\": 10,\n      \"visibleText\": [\n        \"目标发现差距\",\n        \"任务训练能力\",\n        \"反馈修正方向\",\n        \"作品证明自己\"\n      ],\n      \"visualStyle\": \"强收束海报构图，四个短句像宣言贴在作品墙上，主角表情坚定。\",\n      \"speakerNotes\": \"最后用材料中的总结收束：如果想知道当前会什么，就看能独立完成什么；想知道不会什么，就看在哪些任务中卡住；想知道该学什么，就从目标和差距倒推；想知道怎么应用，就把知识放进真实问题里，做出可见成果。学习不是把知识堆进大脑，而是让自己解决问题的能力越来越强。\",\n      \"revisedPrompt\": \"Create a single 1536x864 landscape presentation slide in a polished manga / NotebookLM-Kael style. Final poster composition, strong closure, editorial manga framing, expressive but not chaotic, high sensibility, presentation-ready.\\n\\nScene: A confident learner stands in the foreground facing slightly left, in front of a completed works wall. The learner has a determined, calm expression, subtle manga shading, clean linework, dynamic but controlled pose. Behind them is a wall of completed outputs displayed like trophies / evidence: a report page, a project board with sticky-note shapes but NO readable text, a tool or app icon-like object, a presentation deck thumbnail, and an article page. Connect them with a simple visible path / flow from left to right or bottom to top: goal -> task -> feedback -> work, shown with arrows, panels, and motion cues, but without any extra words except the required four lines.\\n\\nTypography and visible text: ONLY these four Chinese lines, each as a bold clean label or poster caption, readable at presentation distance, exactly and only these lines:\\n目标发现差距\\n任务训练能力\\n反馈修正方向\\n作品证明自己\\nNo other visible text anywhere. No title, no footer, no watermark, no logos, no UI.\\n\\nLayout: One strong poster-like slide. The four phrases should appear as four bold declaration labels integrated into the wall or as floating comic placards. Strong hierarchy, low text density, clear visual rhythm. Use manga speed lines, expressive framing, and clean comic panel energy, but keep it elegant and coherent. The completed works wall should occupy the background right/center, the learner in foreground left/center. Use dramatic lighting, crisp outlines, warm-neutral palette with high contrast accent colors. Make it feel like the memorable ending slide of a learning system deck.\\n\\nImportant: Do not add any text beyond the four required lines. Do not add extra captions. Do not add watermarks, signatures, borders, browser chrome, or fake UI.\",\n      \"storageFileId\": \"dbor32oeat876lytpnrpzxhn-8a313acde88f3a76e1e195ae46ecc9bc\"\n    }\n  ],\n  \"styleId\": \"manga_story\",\n  \"summary\": \"面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\",\n  \"audience\": \"职场学习者、自我提升者、培训或工作坊参与者\",\n  \"imageSize\": \"1536x864\",\n  \"imageModel\": \"gpt-image-2\",\n  \"plannerModel\": \"gpt-5.5\"\n}","markdown":"# 怎么知道自己当前会什么、不会什么、该学什么、怎么应用？\n\n{\n  \"style\": \"Manga Style\",\n  \"title\": \"把学习变成能力系统\",\n  \"format\": \"presenter_slides\",\n  \"slides\": [\n    {\n      \"size\": \"1536x864\",\n      \"model\": \"gpt-image-2\",\n      \"title\": \"不是不努力，是不知道往哪努力\",\n      \"layout\": \"四格漫画开场。主角坐在书和课程图标堆里，头顶冒出四个问题气泡。最后一格出现一条清晰路径。\",\n      \"status\": \"completed\",\n      \"message\": \"学习的第一步，是把模糊焦虑变成清晰问题。\",\n      \"imageUrl\": \"https://cdn.drillso.com/generated/drill/x6euj249qdncoeyeva8qn63u-ab677975612bfd33a593cdc6a0deef6b.png\",\n      \"storageKey\": \"generated/drill/x6euj249qdncoeyeva8qn63u-ab677975612bfd33a593cdc6a0deef6b.png\",\n      \"imagePrompt\": \"Create exactly one complete presentation slide image.\\n\\nUse the following NotebookLM / Kael style prompt as the dominant art direction. Follow it closely.\\n\\nSTYLE TEMPLATE START\\nUnderstanding becomes deeper with fun.\\nTurn the information into a comic-like visual explanation with a simple story.\\nUse panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions.\\nRelate the idea to a realistic situation so the audience can remember it.\\nThe style should be fun and explanatory, not childish or chaotic.\\nSTYLE TEMPLATE END\\n\\nDeck title: 把学习变成能力系统\\nDeck summary: 面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\\nDeck format: presenter_slides\\nChosen style: Manga Style\\n\\nSlide 1: 不是不努力，是不知道往哪努力\\nMain message: 学习的第一步，是把模糊焦虑变成清晰问题。\\n\\nVisible text must be exactly these lines, with no extra words:\\n我到底会什么？\\n不会什么？\\n该学什么？\\n怎么用起来？\\n\\nLayout:\\n四格漫画开场。主角坐在书和课程图标堆里，头顶冒出四个问题气泡。最后一格出现一条清晰路径。\\n\\nVisual style:\\n漫画分镜，表情夸张但不幼稚，问题气泡大，路径线作为视觉主轴。\\n\\nAdditional image direction:\\nA four-panel comic opening scene. A young office learner sits at a desk surrounded by course cards, saved articles, books, and app icons, looking confused. Four speech bubbles contain the visible questions. In the final panel, the clutter transforms into a simple path with six small stepping stones. Use expressive manga reactions, clean panels, speed lines around the confusion, and concise Chinese text.\\n\\nRequirements:\\n- Canvas: 1536x864 landscape presentation slide.\\n- Generate one slide only, not a contact sheet and not multiple slide candidates.\\n- Make it presentation-ready, high-sensibility, and coherent with the chosen style template.\\n- Make it visually finished, not a sparse wireframe or placeholder diagram.\\n- Use the strongest signature elements of the style template: palette, typography, composition language, icon/illustration treatment, and editorial rhythm.\\n- Keep all visible text readable at presentation distance.\\n- Use low text density, strong hierarchy, and source-grounded diagrams, editorial composition, or template-specific visuals.\\n- Do not add markdown symbols, fake logos, watermarks, signatures, UI browser chrome, or extra captions.\\n- Match the source/user instruction language for all visible slide text and speaker-facing copy.\",\n      \"slideNumber\": 1,\n      \"visibleText\": [\n        \"我到底会什么？\",\n        \"不会什么？\",\n        \"该学什么？\",\n        \"怎么用起来？\"\n      ],\n      \"visualStyle\": \"漫画分镜，表情夸张但不幼稚，问题气泡大，路径线作为视觉主轴。\",\n      \"speakerNotes\": \"开场先点出常见困惑：很多人买了很多课、收藏了很多资料，却依然觉得能力模糊。真正的问题不是不努力，而是不知道该往哪里努力。本 deck 会用一条方法链来回答：盘点能力、识别差距、选择目标、设计学习、实践应用、复盘迭代。\",\n      \"revisedPrompt\": \"Create exactly one complete presentation slide image, landscape 16:9, polished and presentation-ready, in a Manga Style inspired by NotebookLM / Kael editorial comic explanation. Dominant art direction: understanding becomes deeper with fun; turn the information into a comic-like visual explanation with a simple story; use panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions; realistic situation, fun and explanatory, not childish or chaotic.\\n\\nSlide topic: Chinese learning-and-ability system. Title concept: 把学习变成能力系统. Slide 1 main message: 不是不努力，是不知道往哪努力. The slide should feel like the opening of a story: a young office learner is overwhelmed by ambiguity, then a simple path appears.\\n\\nLayout: a clean four-panel comic opening scene across the slide, with strong editorial rhythm and clear hierarchy. Panel 1: the learner sits at a desk surrounded by course cards, saved articles, books, and app icons, looking confused, with dramatic manga speed lines and clutter. Panel 2: an expressive close-up reaction with a large speech bubble containing exactly this text: 我到底会什么？ Panel 3: another reaction close-up and question bubble containing exactly this text: 不会什么？ Panel 4: the clutter transforms into a simple clear path with six small stepping stones leading forward, with two more large speech bubbles containing exactly these texts, each on its own line: 该学什么？ and 怎么用起来？\\n\\nVisible text must be exactly these four lines and no extra words anywhere in the image:\\n我到底会什么？\\n不会什么？\\n该学什么？\\n怎么用起来？\\n\\nImportant constraints: no other visible text, no title text, no subtitles, no labels, no watermark, no logos, no UI chrome, no extra captions. Make the four speech bubbles the only text elements. Chinese text must be large, crisp, and readable at presentation distance. Use clean comic panel borders, bold linework, expressive but mature office-learner character design, and a palette of muted warm beige, off-white, charcoal, desaturated blue, and accent red/orange for emphasis. The final panel should visually resolve the clutter into a simple path as the visual main axis, with six small stepping stones and a forward direction. Make the composition feel finished, balanced, and coherent, with strong manga framing, motion lines, and a high-end slide design aesthetic.\",\n      \"storageFileId\": \"x6euj249qdncoeyeva8qn63u-ab677975612bfd33a593cdc6a0deef6b\"\n    },\n    {\n      \"size\": \"1536x864\",\n      \"model\": \"gpt-image-2\",\n      \"title\": \"先区分：知道，不等于会\",\n      \"layout\": \"竖向升级阶梯。主角从底层听讲，逐步走到顶层教别人，每层用一个简短动作表示。\",\n      \"status\": \"completed\",\n      \"message\": \"能力不是一个点，而是从听过到能教别人的层级。\",\n      \"imageUrl\": \"https://cdn.drillso.com/generated/drill/pdusfddf92up1ielh2rerlng-f2ea5ed653c402ef3317366352e59a05.png\",\n      \"storageKey\": \"generated/drill/pdusfddf92up1ielh2rerlng-f2ea5ed653c402ef3317366352e59a05.png\",\n      \"imagePrompt\": \"Create exactly one complete presentation slide image.\\n\\nUse the following NotebookLM / Kael style prompt as the dominant art direction. Follow it closely.\\n\\nSTYLE TEMPLATE START\\nUnderstanding becomes deeper with fun.\\nTurn the information into a comic-like visual explanation with a simple story.\\nUse panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions.\\nRelate the idea to a realistic situation so the audience can remember it.\\nThe style should be fun and explanatory, not childish or chaotic.\\nSTYLE TEMPLATE END\\n\\nDeck title: 把学习变成能力系统\\nDeck summary: 面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\\nDeck format: presenter_slides\\nChosen style: Manga Style\\n\\nSlide 2: 先区分：知道，不等于会\\nMain message: 能力不是一个点，而是从听过到能教别人的层级。\\n\\nVisible text must be exactly these lines, with no extra words:\\n听过\\n理解\\n模仿\\n独立完成\\n教别人\\n\\nLayout:\\n竖向升级阶梯。主角从底层听讲，逐步走到顶层教别人，每层用一个简短动作表示。\\n\\nVisual style:\\n升级闯关式漫画阶梯，层级清楚，顶层有聚光效果。\\n\\nAdditional image direction:\\nA manga-style vertical staircase with five labeled levels in Chinese: 听过, 理解, 模仿, 独立完成, 教别人. The same learner character climbs upward: hearing a term, explaining it, following a tutorial, completing a task alone, then teaching a colleague. Keep text sparse, use bold labels, action poses, and clean panel-like separators.\\n\\nRequirements:\\n- Canvas: 1536x864 landscape presentation slide.\\n- Generate one slide only, not a contact sheet and not multiple slide candidates.\\n- Make it presentation-ready, high-sensibility, and coherent with the chosen style template.\\n- Make it visually finished, not a sparse wireframe or placeholder diagram.\\n- Use the strongest signature elements of the style template: palette, typography, composition language, icon/illustration treatment, and editorial rhythm.\\n- Keep all visible text readable at presentation distance.\\n- Use low text density, strong hierarchy, and source-grounded diagrams, editorial composition, or template-specific visuals.\\n- Do not add markdown symbols, fake logos, watermarks, signatures, UI browser chrome, or extra captions.\\n- Match the source/user instruction language for all visible slide text and speaker-facing copy.\",\n      \"slideNumber\": 2,\n      \"visibleText\": [\n        \"听过\",\n        \"理解\",\n        \"模仿\",\n        \"独立完成\",\n        \"教别人\"\n      ],\n      \"visualStyle\": \"升级闯关式漫画阶梯，层级清楚，顶层有聚光效果。\",\n      \"speakerNotes\": \"很多人会把听过、看过误认为自己会了。材料把能力分成五层：听过、理解、模仿、独立完成、教别人。判断自己会不会，不要只问我学过吗，而要看能否不用资料独立完成，换场景还能不能用，出错时能不能排查，能不能教给别人，结果是否稳定可靠。\",\n      \"revisedPrompt\": \"Create one complete presentation slide image, 16:9 landscape, 1536x864, in a polished manga-style NotebookLM/Kael-inspired editorial illustration. Dominant art direction: understanding becomes deeper with fun; turn the information into a comic-like visual explanation with a simple story; use panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions; relate the idea to a realistic workplace learning situation; fun and explanatory, not childish or chaotic.\\n\\nSlide concept: a vertical upgrade staircase showing the progression from hearing a term to teaching it. The SAME learner character climbs upward through five clear levels, each level represented as a step/panel with a distinct action pose. Composition should feel like a clean manga challenge ladder with strong hierarchy and panel-like separators, dynamic angle, and a top spotlight at the highest level.\\n\\nVisible text must be EXACTLY these five Chinese labels and nothing else anywhere in the slide:\\n听过\\n理解\\n模仿\\n独立完成\\n教别人\\n\\nLayout details: a bold vertical staircase rising from bottom-left to top-right or straight upward within the slide, with five large steps. Each step contains the learner in a different state: bottom hears a term with a small speech bubble-like reaction icon (no text besides the label); next explains/understands with a thoughtful pose; next follows a tutorial with hands and eyes focused; next completes a task alone with confidence; top teaches a colleague under a bright spotlight. Use clean manga framing around each level, speed lines indicating upward progress, expressive facial expressions, and subtle office-learning context such as a desk, laptop, sticky notes, document, whiteboard, and colleague silhouette. Keep the slide elegant and readable at presentation distance. Use bold Chinese labels placed clearly on each step, large enough to read.\\n\\nColor and rendering: manga style with crisp black ink lines, high-contrast shading, tasteful limited palette, editorial composition, subtle warm accent colors and blue-gray shadows, white/light background with dynamic diagonal shapes, polished presentation-ready finish. Avoid clutter, avoid extra captions, avoid any text beyond the five labels, avoid logos, watermarks, UI, or decorative paragraphs. Make it look like a finished single keynote slide, not a wireframe or contact sheet.\",\n      \"storageFileId\": \"pdusfddf92up1ielh2rerlng-f2ea5ed653c402ef3317366352e59a05\"\n    },\n    {\n      \"size\": \"1536x864\",\n      \"model\": \"gpt-image-2\",\n      \"title\": \"能力必须有证据\",\n      \"layout\": \"桌面俯视漫画。一本打开的能力清单笔记本占主体，四个栏目像表格贴纸一样排布。\",\n      \"status\": \"completed\",\n      \"message\": \"不要写“我会”，要写“我做成过什么”。\",\n      \"imageUrl\": \"https://cdn.drillso.com/generated/drill/c9nu33ayugtpvkkfey1cxq9g-442834a751b795f701ca120cf9eff850.png\",\n      \"storageKey\": \"generated/drill/c9nu33ayugtpvkkfey1cxq9g-442834a751b795f701ca120cf9eff850.png\",\n      \"imagePrompt\": \"Create exactly one complete presentation slide image.\\n\\nUse the following NotebookLM / Kael style prompt as the dominant art direction. Follow it closely.\\n\\nSTYLE TEMPLATE START\\nUnderstanding becomes deeper with fun.\\nTurn the information into a comic-like visual explanation with a simple story.\\nUse panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions.\\nRelate the idea to a realistic situation so the audience can remember it.\\nThe style should be fun and explanatory, not childish or chaotic.\\nSTYLE TEMPLATE END\\n\\nDeck title: 把学习变成能力系统\\nDeck summary: 面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\\nDeck format: presenter_slides\\nChosen style: Manga Style\\n\\nSlide 3: 能力必须有证据\\nMain message: 不要写“我会”，要写“我做成过什么”。\\n\\nVisible text must be exactly these lines, with no extra words:\\n当前水平\\n证据\\n问题\\n下一步\\n\\nLayout:\\n桌面俯视漫画。一本打开的能力清单笔记本占主体，四个栏目像表格贴纸一样排布。\\n\\nVisual style:\\n手账式漫画信息图，证据栏用放大镜强调，旁边有主角认真记录。\\n\\nAdditional image direction:\\nTop-down manga desk scene with an open notebook titled 能力清单. Four large clean fields are visible: 当前水平, 证据, 问题, 下一步. A magnifying glass highlights the 证据 field. Around the notebook, small icons represent writing, Excel, communication, project delivery, and AI tools. Use clear hierarchy and minimal Chinese text.\\n\\nRequirements:\\n- Canvas: 1536x864 landscape presentation slide.\\n- Generate one slide only, not a contact sheet and not multiple slide candidates.\\n- Make it presentation-ready, high-sensibility, and coherent with the chosen style template.\\n- Make it visually finished, not a sparse wireframe or placeholder diagram.\\n- Use the strongest signature elements of the style template: palette, typography, composition language, icon/illustration treatment, and editorial rhythm.\\n- Keep all visible text readable at presentation distance.\\n- Use low text density, strong hierarchy, and source-grounded diagrams, editorial composition, or template-specific visuals.\\n- Do not add markdown symbols, fake logos, watermarks, signatures, UI browser chrome, or extra captions.\\n- Match the source/user instruction language for all visible slide text and speaker-facing copy.\",\n      \"slideNumber\": 3,\n      \"visibleText\": [\n        \"当前水平\",\n        \"证据\",\n        \"问题\",\n        \"下一步\"\n      ],\n      \"visualStyle\": \"手账式漫画信息图，证据栏用放大镜强调，旁边有主角认真记录。\",\n      \"speakerNotes\": \"盘点能力时，要建立能力清单，把笼统能力拆成具体项目。材料建议可以按专业技能、通用能力、工具能力、业务理解、交付能力来拆。关键是能力必须有证据。不要只写我会演讲，而要写能在 20 人会议上做 10 分钟汇报，并回答 3 个问题。\",\n      \"revisedPrompt\": \"Create exactly one complete presentation slide image, 1536x864 landscape, in a NotebookLM / Kael inspired manga style with a fun explanatory comic-like visual language. Top-down desk scene, editorial and polished, not childish, not chaotic. Dominant art direction: understanding becomes deeper with fun; simple story, panels, character reactions, manga framing, speed lines, expressive close-ups, concise captions. \\n\\nSlide 3 topic: 能力必须有证据. Main message should be visually communicated: 不要写“我会”，要写“我做成过什么”。 \\n\\nVisible text must be exactly these lines and no other text anywhere:\\n当前水平\\n证据\\n问题\\n下一步\\n\\nLayout: desktop top-down manga infographic. An open notebook dominates the center and is titled 能力清单 in a clean handwritten manga style. Inside the notebook, four large clean field areas are arranged like sticker-like table sections, with the visible text exactly: 当前水平, 证据, 问题, 下一步. The 证据 field is emphasized by a magnifying glass placed over it, with strong graphic focus, highlight ring, and subtle speed lines. A thoughtful protagonist hand/character is nearby, seriously noting things down. Around the notebook, small icons illustrate writing, Excel, communication, project delivery, and AI tools, arranged neatly like desk items. \\n\\nArt direction: hand-drawn manga editorial infographic, clean composition, strong hierarchy, minimal Chinese text, presentation-ready, coherent, high sensibility, realistic desk objects combined with comic elements. Use palette and typography inspired by Japanese manga notes: off-white paper, warm beige desk, black ink lines, muted teal and orange accents, clean shadows, crisp legibility. Add subtle panel-like framing and expressive close-up feel without splitting into multiple slide candidates. Make the evidence area the visual focal point. \\n\\nImportant constraints: no extra captions, no watermarks, no logos, no browser chrome, no UI mockups, no markdown symbols, no extra words beyond the four required visible text lines; ensure those text lines are readable at presentation distance. The notebook title 能力清单 can appear as part of the notebook cover/title. Make the slide feel finished and coherent, not sparse or placeholder-like.\",\n      \"storageFileId\": \"c9nu33ayugtpvkkfey1cxq9g-442834a751b795f701ca120cf9eff850\"\n    },\n    {\n      \"size\": \"1536x864\",\n      \"model\": \"gpt-image-2\",\n      \"title\": \"不会的地方，藏在任务里\",\n      \"layout\": \"四宫格侦探漫画。主角像侦探一样寻找盲区，每格对应一种方法。\",\n      \"status\": \"completed\",\n      \"message\": \"真正的差距，会在真实任务中暴露出来。\",\n      \"imageUrl\": \"https://cdn.drillso.com/generated/drill/h1er9ut8gu1bakkopkd0flmk-fdf0f5f5eeffbbff9435605789265fb2.png\",\n      \"storageKey\": \"generated/drill/h1er9ut8gu1bakkopkd0flmk-fdf0f5f5eeffbbff9435605789265fb2.png\",\n      \"imagePrompt\": \"Create exactly one complete presentation slide image.\\n\\nUse the following NotebookLM / Kael style prompt as the dominant art direction. Follow it closely.\\n\\nSTYLE TEMPLATE START\\nUnderstanding becomes deeper with fun.\\nTurn the information into a comic-like visual explanation with a simple story.\\nUse panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions.\\nRelate the idea to a realistic situation so the audience can remember it.\\nThe style should be fun and explanatory, not childish or chaotic.\\nSTYLE TEMPLATE END\\n\\nDeck title: 把学习变成能力系统\\nDeck summary: 面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\\nDeck format: presenter_slides\\nChosen style: Manga Style\\n\\nSlide 4: 不会的地方，藏在任务里\\nMain message: 真正的差距，会在真实任务中暴露出来。\\n\\nVisible text must be exactly these lines, with no extra words:\\n对标\\n任务\\n提问\\n反馈\\n\\nLayout:\\n四宫格侦探漫画。主角像侦探一样寻找盲区，每格对应一种方法。\\n\\nVisual style:\\n侦探调查感，放大镜、脚印、问号和评分牌作为符号。\\n\\nAdditional image direction:\\nA four-panel manga detective layout. The learner wears a simple detective cap and investigates four clues labeled 对标, 任务, 提问, 反馈. Panel 1 compares their work with an expert article. Panel 2 shows a sales data task with a stuck expression. Panel 3 shows question marks around core concepts. Panel 4 shows a colleague giving a score card. Keep labels large and text minimal.\\n\\nRequirements:\\n- Canvas: 1536x864 landscape presentation slide.\\n- Generate one slide only, not a contact sheet and not multiple slide candidates.\\n- Make it presentation-ready, high-sensibility, and coherent with the chosen style template.\\n- Make it visually finished, not a sparse wireframe or placeholder diagram.\\n- Use the strongest signature elements of the style template: palette, typography, composition language, icon/illustration treatment, and editorial rhythm.\\n- Keep all visible text readable at presentation distance.\\n- Use low text density, strong hierarchy, and source-grounded diagrams, editorial composition, or template-specific visuals.\\n- Do not add markdown symbols, fake logos, watermarks, signatures, UI browser chrome, or extra captions.\\n- Match the source/user instruction language for all visible slide text and speaker-facing copy.\",\n      \"slideNumber\": 4,\n      \"visibleText\": [\n        \"对标\",\n        \"任务\",\n        \"提问\",\n        \"反馈\"\n      ],\n      \"visualStyle\": \"侦探调查感，放大镜、脚印、问号和评分牌作为符号。\",\n      \"speakerNotes\": \"识别不会什么，可以用四种方法。对标法是看高手的成果、流程和思考方式。任务法是用真实任务测试，比如拿销售数据找出下降原因并提出建议。提问法是用核心概念、关系、异常处理等问题暴露盲区。反馈法是请同事、老师或朋友指出哪里好、哪里不清楚、哪里效率低。\",\n      \"revisedPrompt\": \"Create exactly one complete presentation slide image, landscape 1536x864, in a polished Manga Style / NotebookLM / Kael style. Dominant art direction: understanding becomes deeper with fun; comic-like explanatory visual with a simple story, panels, character reactions, manga framing, speed lines, expressive close-ups, concise captions. Serious, fun, explanatory, not childish. \\n\\nDeck slide topic: \\\"不会的地方，藏在任务里\\\" with main message: truly real gaps are exposed in real tasks.\\n\\nDesign a single finished presentation slide, not a contact sheet, not multiple candidate slides. Four-panel detective manga layout, coherent editorial composition. The learner is a detective-like main character wearing a simple detective cap, investigating four clues. Visual style: detective investigation vibe, magnifying glass, footprints, question marks, and score card symbols. Include a realistic office/work-learning setting. \\n\\nVisible text must be exactly these four Chinese lines and nothing else, large and readable, each as a label in one panel: \\n对标\\n任务\\n提问\\n反馈\\n\\nNo other visible text anywhere. No title, no subtitle, no extra captions, no watermark, no UI. Keep text minimal and clean.\\n\\nPanel 1: learner compares their work with an expert article / benchmark material, using a magnifying glass over a document, subtle comparison cues. Label: 对标.\\nPanel 2: learner confronts a sales data task, stuck expression, evidence of uncertainty, files/spreadsheet vibe, detective examining task clues. Label: 任务.\\nPanel 3: question marks swirl around core concepts, learner thinking and asking precise questions, manga close-up reaction. Label: 提问.\\nPanel 4: colleague gives a score card / feedback card, learner receiving constructive feedback, modestly relieved expression. Label: 反馈.\\n\\nUse a four-panel comic grid with strong manga framing, varied camera angles, expressive faces, speed lines, inked outlines, halftone shading, dynamic but clean composition. Presentation-ready, high-sensibility, visually finished. Use a restrained palette: off-white paper, black ink, muted teal, warm orange accents, and red highlights for emphasis. Ensure the four labels are clearly legible and the overall slide feels cohesive and professional.\",\n      \"storageFileId\": \"h1er9ut8gu1bakkopkd0flmk-fdf0f5f5eeffbbff9435605789265fb2\"\n    },\n    {\n      \"size\": \"1536x864\",\n      \"model\": \"gpt-image-2\",\n      \"title\": \"该学什么，从目标倒推\",\n      \"layout\": \"横向流程漫画。主角从左到右推着一个目标牌，经过六个节点，最后进入真实应用场景。\",\n      \"status\": \"completed\",\n      \"message\": \"不要被热门牵着走，要从真实目标拆出关键差距。\",\n      \"imageUrl\": \"https://cdn.drillso.com/generated/drill/pigjfvkup3iucez8bovjfolb-8df686416354c784ce6a5d2632631f54.png\",\n      \"storageKey\": \"generated/drill/pigjfvkup3iucez8bovjfolb-8df686416354c784ce6a5d2632631f54.png\",\n      \"imagePrompt\": \"Create exactly one complete presentation slide image.\\n\\nUse the following NotebookLM / Kael style prompt as the dominant art direction. Follow it closely.\\n\\nSTYLE TEMPLATE START\\nUnderstanding becomes deeper with fun.\\nTurn the information into a comic-like visual explanation with a simple story.\\nUse panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions.\\nRelate the idea to a realistic situation so the audience can remember it.\\nThe style should be fun and explanatory, not childish or chaotic.\\nSTYLE TEMPLATE END\\n\\nDeck title: 把学习变成能力系统\\nDeck summary: 面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\\nDeck format: presenter_slides\\nChosen style: Manga Style\\n\\nSlide 5: 该学什么，从目标倒推\\nMain message: 不要被热门牵着走，要从真实目标拆出关键差距。\\n\\nVisible text must be exactly these lines, with no extra words:\\n目标\\n任务\\n能力\\n差距\\n计划\\n应用\\n\\nLayout:\\n横向流程漫画。主角从左到右推着一个目标牌，经过六个节点，最后进入真实应用场景。\\n\\nVisual style:\\n流程图和漫画动作结合，节点像游戏关卡，箭头清晰。\\n\\nAdditional image direction:\\nA horizontal manga flow diagram with the learner pushing a sign labeled 目标 through six connected nodes: 目标, 任务, 能力, 差距, 计划, 应用. Avoid dense text. The background hints at a workplace project report scenario. Use dynamic arrows, speed lines, and clear game-level checkpoints.\\n\\nRequirements:\\n- Canvas: 1536x864 landscape presentation slide.\\n- Generate one slide only, not a contact sheet and not multiple slide candidates.\\n- Make it presentation-ready, high-sensibility, and coherent with the chosen style template.\\n- Make it visually finished, not a sparse wireframe or placeholder diagram.\\n- Use the strongest signature elements of the style template: palette, typography, composition language, icon/illustration treatment, and editorial rhythm.\\n- Keep all visible text readable at presentation distance.\\n- Use low text density, strong hierarchy, and source-grounded diagrams, editorial composition, or template-specific visuals.\\n- Do not add markdown symbols, fake logos, watermarks, signatures, UI browser chrome, or extra captions.\\n- Match the source/user instruction language for all visible slide text and speaker-facing copy.\",\n      \"slideNumber\": 5,\n      \"visibleText\": [\n        \"目标\",\n        \"任务\",\n        \"能力\",\n        \"差距\",\n        \"计划\",\n        \"应用\"\n      ],\n      \"visualStyle\": \"流程图和漫画动作结合，节点像游戏关卡，箭头清晰。\",\n      \"speakerNotes\": \"材料指出，常见误区是看到什么热门就学什么，今天学 AI，明天学剪辑，后天学投资，最后没有形成能力闭环。更有效的方法是从目标倒推：明确目标，拆解任务，识别所需能力，评估当前水平，找出关键差距，制定学习计划，再真实应用并复盘改进。\",\n      \"revisedPrompt\": \"Create a single presentation slide image, 1536x864 landscape, in a manga-style NotebookLM / Kael editorial comic format. Dominant art direction: understanding becomes deeper with fun; turn the information into a comic-like visual explanation with a simple story; use panels, expressive reactions, manga framing, speed lines, concise captions, but keep it clean, fun, explanatory, not childish or chaotic.\\n\\nSlide topic: '该学什么，从目标倒推'. Main message:不要被热门牵着走，要从真实目标拆出关键差距.\\n\\nComposition: a horizontal manga flow diagram from left to right. A learner character pushes a large sign labeled exactly '目标' through six connected game-like checkpoint nodes, then into a realistic workplace project report scene at the far right. The six nodes must be clearly visible and the ONLY visible text on the slide must be these exact six lines, each appearing once and nothing else:\\n目标\\n任务\\n能力\\n差距\\n计划\\n应用\\n\\nLayout details: The learner begins on the left with determined expression, pushing the '目标' sign forward through connected nodes like level checkpoints. Each node is a stylized rounded rectangle or game stage badge with manga framing and dynamic arrows. Show the flow progression: 目标 → 任务 → 能力 → 差距 → 计划 → 应用. Use clear arrows, speed lines, motion effects, and expressive close-ups of the learner at a few stages. At the rightmost end, the flow enters a realistic workplace project report application setting: desk, laptop, charts, meeting board, documents, office environment hinted in the background, showing real-world use. Background should subtly hint at a work project report scenario without adding extra text.\\n\\nVisual style: manga style with editorial composition, bold but tasteful black linework, vivid yet restrained colors, strong hierarchy, game-level checkpoint visuals, clean readable typography for the Chinese node labels, presentation-ready, polished, high contrast, coherent, no clutter, no extra words, no watermark, no logo, no browser chrome. Ensure text is legible at presentation distance. Keep low text density and strong visual rhythm. Make it look like a finished slide, not a wireframe. Avoid any additional captions or labels beyond the six required words.\",\n      \"storageFileId\": \"pigjfvkup3iucez8bovjfolb-8df686416354c784ce6a5d2632631f54\"\n    },\n    {\n      \"size\": \"1536x864\",\n      \"model\": \"gpt-image-2\",\n      \"title\": \"优先级，看三件事\",\n      \"layout\": \"漫画仪表盘。三个大旋钮分别代表重要性、紧急性、杠杆性，主角把学习项目卡片放入评分机。\",\n      \"status\": \"completed\",\n      \"message\": \"不是所有不会的东西都要马上学。\",\n      \"imageUrl\": \"https://cdn.drillso.com/generated/drill/bskvd1s25xzz10oiu38tsxh7-e01f9c917a2b1fcce79a9a65e6f5fc81.png\",\n      \"storageKey\": \"generated/drill/bskvd1s25xzz10oiu38tsxh7-e01f9c917a2b1fcce79a9a65e6f5fc81.png\",\n      \"imagePrompt\": \"Create exactly one complete presentation slide image.\\n\\nUse the following NotebookLM / Kael style prompt as the dominant art direction. Follow it closely.\\n\\nSTYLE TEMPLATE START\\nUnderstanding becomes deeper with fun.\\nTurn the information into a comic-like visual explanation with a simple story.\\nUse panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions.\\nRelate the idea to a realistic situation so the audience can remember it.\\nThe style should be fun and explanatory, not childish or chaotic.\\nSTYLE TEMPLATE END\\n\\nDeck title: 把学习变成能力系统\\nDeck summary: 面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\\nDeck format: presenter_slides\\nChosen style: Manga Style\\n\\nSlide 6: 优先级，看三件事\\nMain message: 不是所有不会的东西都要马上学。\\n\\nVisible text must be exactly these lines, with no extra words:\\n重要性\\n紧急性\\n杠杆性\\n\\nLayout:\\n漫画仪表盘。三个大旋钮分别代表重要性、紧急性、杠杆性，主角把学习项目卡片放入评分机。\\n\\nVisual style:\\n轻量信息仪表盘，三个维度占画面中心，项目卡片作为辅助元素。\\n\\nAdditional image direction:\\nA manga-style priority scoring machine. Three large dials are labeled 重要性, 紧急性, 杠杆性. The learner feeds skill cards into the machine, such as Excel 数据清洗 and 汇报表达, while low-priority cards sit to the side. Use bold dial shapes, minimal visible text, and a clear sense of ranking without adding extra numbers.\\n\\nRequirements:\\n- Canvas: 1536x864 landscape presentation slide.\\n- Generate one slide only, not a contact sheet and not multiple slide candidates.\\n- Make it presentation-ready, high-sensibility, and coherent with the chosen style template.\\n- Make it visually finished, not a sparse wireframe or placeholder diagram.\\n- Use the strongest signature elements of the style template: palette, typography, composition language, icon/illustration treatment, and editorial rhythm.\\n- Keep all visible text readable at presentation distance.\\n- Use low text density, strong hierarchy, and source-grounded diagrams, editorial composition, or template-specific visuals.\\n- Do not add markdown symbols, fake logos, watermarks, signatures, UI browser chrome, or extra captions.\\n- Match the source/user instruction language for all visible slide text and speaker-facing copy.\",\n      \"slideNumber\": 6,\n      \"visibleText\": [\n        \"重要性\",\n        \"紧急性\",\n        \"杠杆性\"\n      ],\n      \"visualStyle\": \"轻量信息仪表盘，三个维度占画面中心，项目卡片作为辅助元素。\",\n      \"speakerNotes\": \"判断学习优先级，可以用三维评分法。第一是重要性：是否影响核心目标。第二是紧急性：近期是否马上会用到。第三是杠杆性：学会后是否能带来长期收益。总分越高，越应该优先学。材料中的例子显示，Excel 数据清洗和汇报表达因为得分高，更应该优先学习。\",\n      \"revisedPrompt\": \"Create exactly one complete presentation slide image, landscape 1536x864, in a manga-style NotebookLM / Kael explanatory comic style. Theme: priority scoring machine for learning choices. The composition is a polished presentation slide, not a wireframe, not a contact sheet, not multiple panels. Use a fun, explanatory, editorial manga look with bold shapes, expressive reactions, speed lines, clean panel framing hints, and a realistic office-learning situation.\\n\\nScene: A learner at a desk feeds skill cards into a large priority-scoring machine in the center of the slide. The machine has THREE huge bold dials taking up the visual center, each clearly labeled with exactly one of the following Chinese words only: 重要性, 紧急性, 杠杆性. These are the ONLY visible words anywhere in the slide. No title, no subtitle, no extra captions, no numbers, no labels on cards, no watermarks, no UI chrome.\\n\\nVisual structure: Center the three dials prominently like a comic dashboard. The learner character is in a clear manga pose, slightly tense but thoughtful, inserting cards into the machine. The cards are symbolic only, with simple icon-like visuals to suggest work skills, but NO text on the cards. One or two low-priority cards sit aside in a dimmer area, also without text. Show the machine ranking the cards by visual emphasis: one card is highlighted going into the machine, another sits aside in muted tone. Use strong hierarchy and a sense of sorting/decision-making.\\n\\nArt direction: Light information dashboard, manga framing, expressive close-up feel, clean and presentation-ready. High contrast, crisp line art, modern comic shading, subtle motion lines, dynamic but not chaotic. Color palette: warm cream background, ink black lines, muted teal and orange accent colors, with bright highlights for the active card and dials. Keep it elegant, educational, and easy to read at presentation distance.\\n\\nImportant constraints: Only the three Chinese dial labels may appear as text. Absolutely no other visible text. Do not add extra words in any language. Do not include logos, signatures, or watermarks. Make it visually finished and coherent, with a strong sense of ranking and prioritization.\",\n      \"storageFileId\": \"bskvd1s25xzz10oiu38tsxh7-e01f9c917a2b1fcce79a9a65e6f5fc81\"\n    },\n    {\n      \"size\": \"1536x864\",\n      \"model\": \"gpt-image-2\",\n      \"title\": \"有效学习，是一个闭环\",\n      \"layout\": \"环形闭环漫画图。四个节点围成循环，中间是一个逐渐成形的作品。\",\n      \"status\": \"completed\",\n      \"message\": \"只看资料不够，要让知识变成成果。\",\n      \"imageUrl\": \"https://cdn.drillso.com/generated/drill/x194b29yzfoalt2z4pfn81mp-b0d8e1614c3cb6dfd8a24f12a5b17f98.png\",\n      \"storageKey\": \"generated/drill/x194b29yzfoalt2z4pfn81mp-b0d8e1614c3cb6dfd8a24f12a5b17f98.png\",\n      \"imagePrompt\": \"Create exactly one complete presentation slide image.\\n\\nUse the following NotebookLM / Kael style prompt as the dominant art direction. Follow it closely.\\n\\nSTYLE TEMPLATE START\\nUnderstanding becomes deeper with fun.\\nTurn the information into a comic-like visual explanation with a simple story.\\nUse panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions.\\nRelate the idea to a realistic situation so the audience can remember it.\\nThe style should be fun and explanatory, not childish or chaotic.\\nSTYLE TEMPLATE END\\n\\nDeck title: 把学习变成能力系统\\nDeck summary: 面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\\nDeck format: presenter_slides\\nChosen style: Manga Style\\n\\nSlide 7: 有效学习，是一个闭环\\nMain message: 只看资料不够，要让知识变成成果。\\n\\nVisible text must be exactly these lines, with no extra words:\\n输入\\n练习\\n反馈\\n输出\\n\\nLayout:\\n环形闭环漫画图。四个节点围成循环，中间是一个逐渐成形的作品。\\n\\nVisual style:\\n循环动作分镜，中心作品发光，强调从资料到成果的转化。\\n\\nAdditional image direction:\\nA circular manga infographic with four connected nodes: 输入, 练习, 反馈, 输出. In the center, a rough idea transforms into a finished work artifact, such as a report or project board. Show books and courses at input, pen and keyboard at practice, comment bubbles at feedback, and a completed deliverable at output. Keep text sparse and hierarchy strong.\\n\\nRequirements:\\n- Canvas: 1536x864 landscape presentation slide.\\n- Generate one slide only, not a contact sheet and not multiple slide candidates.\\n- Make it presentation-ready, high-sensibility, and coherent with the chosen style template.\\n- Make it visually finished, not a sparse wireframe or placeholder diagram.\\n- Use the strongest signature elements of the style template: palette, typography, composition language, icon/illustration treatment, and editorial rhythm.\\n- Keep all visible text readable at presentation distance.\\n- Use low text density, strong hierarchy, and source-grounded diagrams, editorial composition, or template-specific visuals.\\n- Do not add markdown symbols, fake logos, watermarks, signatures, UI browser chrome, or extra captions.\\n- Match the source/user instruction language for all visible slide text and speaker-facing copy.\",\n      \"slideNumber\": 7,\n      \"visibleText\": [\n        \"输入\",\n        \"练习\",\n        \"反馈\",\n        \"输出\"\n      ],\n      \"visualStyle\": \"循环动作分镜，中心作品发光，强调从资料到成果的转化。\",\n      \"speakerNotes\": \"材料把有效学习概括为输入、练习、反馈、输出的闭环。输入包括看书、听课、读文章、观察案例、请教他人，但要有边界，先学够用的最小知识。练习是把知识变成动作。反馈用来发现偏差。输出是最可靠的检验方式，可以是一篇文章、一个项目、一次汇报、一个工具或一个真实问题的解决方案。\",\n      \"revisedPrompt\": \"Create exactly one complete presentation slide image in 16:9 landscape, 1536x864. Dominant art direction: NotebookLM / Kael style, manga-style explanatory comic infographic, fun but polished, not childish. Strong editorial composition, clear hierarchy, low text density, high sensibility, presentation-ready.\\n\\nSlide concept: a circular closed-loop learning system. Four connected nodes arranged in a ring around a glowing center. Center shows a rough idea transforming into a finished work artifact, like a polished report or project board, with a bright glow and subtle motion lines to show transformation from vague draft to completed output.\\n\\nVisual structure:\\n- Circular loop / cycle infographic in the middle of the slide.\\n- Four manga comic nodes evenly spaced around the circle with arrows or curved connectors forming a loop.\\n- Node 1: books, course cards, and reading materials to represent input.\\n- Node 2: pen, notebook, and keyboard to represent practice.\\n- Node 3: comment bubbles, review marks, and mentor-like feedback to represent feedback.\\n- Node 4: completed deliverable, clean report/project board to represent output.\\n- Center: an in-progress idea becoming a finished artifact, luminous, with a subtle burst effect.\\n\\nStyle details:\\n- Manga framing, expressive motion lines, dynamic but clean composition.\\n- Use comic panel energy within a single infographic, with connected vignette-like nodes.\\n- Palette: high-contrast editorial manga colors, mostly warm cream background with black inks and selective accent colors like teal, orange, and gold glow.\\n- Use crisp, readable Chinese typography integrated into the node labels.\\n- Keep the slide uncluttered and visually cohesive.\\n\\nVisible text must be exactly these four lines and nothing else:\\n输入\\n练习\\n反馈\\n输出\\n\\nNo extra text, no title, no footer, no watermark, no UI chrome, no logos, no captions beyond the four required labels. Make the four labels clearly readable and the only text on the slide.\",\n      \"storageFileId\": \"x194b29yzfoalt2z4pfn81mp-b0d8e1614c3cb6dfd8a24f12a5b17f98\"\n    },\n    {\n      \"size\": \"1536x864\",\n      \"model\": \"gpt-image-2\",\n      \"title\": \"把学习绑到真实场景\",\n      \"layout\": \"三段式漫画。左边是问题，中间是作品，右边是用户在场景中使用作品。\",\n      \"status\": \"completed\",\n      \"message\": \"应用不是记笔记，而是解决一个看得见的问题。\",\n      \"imageUrl\": \"https://cdn.drillso.com/generated/drill/e7wzw8m3ktcode3m8xg3p68s-a566d46672a8045db28f95956b376cfc.png\",\n      \"storageKey\": \"generated/drill/e7wzw8m3ktcode3m8xg3p68s-a566d46672a8045db28f95956b376cfc.png\",\n      \"imagePrompt\": \"Create exactly one complete presentation slide image.\\n\\nUse the following NotebookLM / Kael style prompt as the dominant art direction. Follow it closely.\\n\\nSTYLE TEMPLATE START\\nUnderstanding becomes deeper with fun.\\nTurn the information into a comic-like visual explanation with a simple story.\\nUse panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions.\\nRelate the idea to a realistic situation so the audience can remember it.\\nThe style should be fun and explanatory, not childish or chaotic.\\nSTYLE TEMPLATE END\\n\\nDeck title: 把学习变成能力系统\\nDeck summary: 面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\\nDeck format: presenter_slides\\nChosen style: Manga Style\\n\\nSlide 8: 把学习绑到真实场景\\nMain message: 应用不是记笔记，而是解决一个看得见的问题。\\n\\nVisible text must be exactly these lines, with no extra words:\\n真实问题\\n具体作品\\n使用场景\\n\\nLayout:\\n三段式漫画。左边是问题，中间是作品，右边是用户在场景中使用作品。\\n\\nVisual style:\\n产品小剧场感，三个绑定像三张卡片，用线连接成因果链。\\n\\nAdditional image direction:\\nA three-part manga scene connected by arrows. Card 1 labeled 真实问题 shows messy weekly sales data. Card 2 labeled 具体作品 shows an automation script or clean report. Card 3 labeled 使用场景 shows someone using the report in a meeting. Include a small AI tool icon in the background but no brand logos. Minimal Chinese text, expressive characters, clear cause-and-effect.\\n\\nRequirements:\\n- Canvas: 1536x864 landscape presentation slide.\\n- Generate one slide only, not a contact sheet and not multiple slide candidates.\\n- Make it presentation-ready, high-sensibility, and coherent with the chosen style template.\\n- Make it visually finished, not a sparse wireframe or placeholder diagram.\\n- Use the strongest signature elements of the style template: palette, typography, composition language, icon/illustration treatment, and editorial rhythm.\\n- Keep all visible text readable at presentation distance.\\n- Use low text density, strong hierarchy, and source-grounded diagrams, editorial composition, or template-specific visuals.\\n- Do not add markdown symbols, fake logos, watermarks, signatures, UI browser chrome, or extra captions.\\n- Match the source/user instruction language for all visible slide text and speaker-facing copy.\",\n      \"slideNumber\": 8,\n      \"visibleText\": [\n        \"真实问题\",\n        \"具体作品\",\n        \"使用场景\"\n      ],\n      \"visualStyle\": \"产品小剧场感，三个绑定像三张卡片，用线连接成因果链。\",\n      \"speakerNotes\": \"学完不用，很快会忘。材料建议应用能力时做三个绑定：绑定一个真实问题，绑定一个具体作品，绑定一个使用场景。比如不要只说我要学 Python，而是说我要用 Python 自动整理每周销售数据，减少手工复制时间。学习 AI 工具也不要只研究提示词技巧，而要放到写周报、做调研、学英语或写代码等具体场景里。\",\n      \"revisedPrompt\": \"Create exactly one complete presentation slide image in 1536x864 landscape, manga-style educational editorial slide, inspired by NotebookLM / Kael style: fun explanatory comic with simple story, panels, character reactions, manga framing, speed lines, expressive close-ups, concise captions, polished and presentation-ready. The slide topic is: 把学习绑到真实场景. Main message visualized: 应用不是记笔记，而是解决一个看得见的问题.\\n\\nLayout: three-part manga scene arranged left to right as three connected cards, with arrows/line connectors showing a cause-and-effect chain. Product theater feeling, like three linked cards. Left card = messy weekly sales problem, middle card = clean output/automation script/report, right card = person using the report in a meeting. Add a small AI tool icon in the background, abstract and generic, no brand logos.\\n\\nVisible text must be exactly these three lines and nothing else anywhere on the slide:\\n真实问题\\n具体作品\\n使用场景\\n\\nPlace the text as the card headings only, large, readable, and centered or top-left on each card. No other Chinese or English text, no extra captions, no watermarks, no UI chrome, no logo.\\n\\nArt direction: manga-style but not childish or chaotic; modern editorial palette with strong contrast, clean composition, subtle halftone textures, speed lines, expressive character faces. Left card should show a realistic messy weekly sales data situation: scattered numbers, charts, papers, and a stressed office worker. Middle card should show a neat artifact: a clean report dashboard or automation script flowing into a tidy summary sheet, with a focused character producing it. Right card should show a confident person presenting/using the report in a meeting, with colleagues reacting positively. Use arrows connecting the cards. Keep low text density, strong hierarchy, and clear cause-and-effect. Make it visually finished, balanced, and high-sensibility, suitable for a deck slide.\\n\\nDo not include any other visible text beyond the exact three headings listed above.\",\n      \"storageFileId\": \"e7wzw8m3ktcode3m8xg3p68s-a566d46672a8045db28f95956b376cfc\"\n    },\n    {\n      \"size\": \"1536x864\",\n      \"model\": \"gpt-image-2\",\n      \"title\": \"每周 30 分钟，复盘能力边界\",\n      \"layout\": \"日历和复盘表漫画。主角周末坐下复盘，四个问题以便签形式围绕日历。\",\n      \"status\": \"completed\",\n      \"message\": \"持续复盘，才能看清自己能独立完成什么、在哪些地方卡住。\",\n      \"imageUrl\": \"https://cdn.drillso.com/generated/drill/qv6tiax8m7v5jri3kffciy1s-69c135520f937f8195a020ed3bfcd908.png\",\n      \"storageKey\": \"generated/drill/qv6tiax8m7v5jri3kffciy1s-69c135520f937f8195a020ed3bfcd908.png\",\n      \"imagePrompt\": \"Create exactly one complete presentation slide image.\\n\\nUse the following NotebookLM / Kael style prompt as the dominant art direction. Follow it closely.\\n\\nSTYLE TEMPLATE START\\nUnderstanding becomes deeper with fun.\\nTurn the information into a comic-like visual explanation with a simple story.\\nUse panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions.\\nRelate the idea to a realistic situation so the audience can remember it.\\nThe style should be fun and explanatory, not childish or chaotic.\\nSTYLE TEMPLATE END\\n\\nDeck title: 把学习变成能力系统\\nDeck summary: 面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\\nDeck format: presenter_slides\\nChosen style: Manga Style\\n\\nSlide 9: 每周 30 分钟，复盘能力边界\\nMain message: 持续复盘，才能看清自己能独立完成什么、在哪些地方卡住。\\n\\nVisible text must be exactly these lines, with no extra words:\\n完成了什么？\\n哪里卡住？\\n下周补什么？\\n用什么任务练？\\n\\nLayout:\\n日历和复盘表漫画。主角周末坐下复盘，四个问题以便签形式围绕日历。\\n\\nVisual style:\\n温和收束，便签式问题，日历上标出 30 分钟和 4 周节奏。\\n\\nAdditional image direction:\\nA closing manga scene of the learner doing a weekly review at a desk. A calendar shows a repeating weekly rhythm and a small 30 分钟 marker. Four sticky notes around the calendar read: 完成了什么？ 哪里卡住？ 下周补什么？ 用什么任务练？ Show calm confidence, fewer speed lines, and a clear sense of progress over four weeks.\\n\\nRequirements:\\n- Canvas: 1536x864 landscape presentation slide.\\n- Generate one slide only, not a contact sheet and not multiple slide candidates.\\n- Make it presentation-ready, high-sensibility, and coherent with the chosen style template.\\n- Make it visually finished, not a sparse wireframe or placeholder diagram.\\n- Use the strongest signature elements of the style template: palette, typography, composition language, icon/illustration treatment, and editorial rhythm.\\n- Keep all visible text readable at presentation distance.\\n- Use low text density, strong hierarchy, and source-grounded diagrams, editorial composition, or template-specific visuals.\\n- Do not add markdown symbols, fake logos, watermarks, signatures, UI browser chrome, or extra captions.\\n- Match the source/user instruction language for all visible slide text and speaker-facing copy.\",\n      \"slideNumber\": 9,\n      \"visibleText\": [\n        \"完成了什么？\",\n        \"哪里卡住？\",\n        \"下周补什么？\",\n        \"用什么任务练？\"\n      ],\n      \"visualStyle\": \"温和收束，便签式问题，日历上标出 30 分钟和 4 周节奏。\",\n      \"speakerNotes\": \"材料给出一个每周复盘模板，建议每周花 30 分钟回答：本周完成了什么具体任务，哪些能独立完成，哪些地方卡住了，卡住是因为概念不懂、方法不会还是练习太少，收到了哪些反馈，下周最应该补哪一个短板，以及用什么真实任务来练习。持续 4 周，会更清楚自己的能力边界。\",\n      \"revisedPrompt\": \"Create exactly one complete presentation slide image in 16:9 landscape (1536x864), finished and presentation-ready, in a Manga Style inspired by NotebookLM / Kael style. Dominant art direction: understanding becomes deeper with fun; turn information into a comic-like visual explanation with a simple story; use panels, character reactions, manga framing, subtle speed lines, expressive close-ups, and concise captions; relate the idea to a realistic situation so it feels memorable; fun and explanatory, not childish or chaotic.\\n\\nDeck title theme: 把学习变成能力系统. Slide 9 topic: 每周 30 分钟，复盘能力边界. Main message: 持续复盘，才能看清自己能独立完成什么、在哪些地方卡住.\\n\\nScene/layout: a closing manga scene of a learner doing a weekly review at a desk on a calm weekend evening. The learner sits at a desk with a notebook, pen, and a calendar spread. A large calendar shows a repeating weekly rhythm and a small clear marker reading 30 分钟, plus a 4-week progression cue (e.g., four small weekly boxes or four weeks highlighted) indicating the review habit over four weeks. Around the calendar, four sticky notes placed in a balanced editorial composition, each with one exact visible text line. The four sticky notes must read exactly and only these lines, one per note, with no extra visible words anywhere else on the slide: 完成了什么？  哪里卡住？  下周补什么？  用什么任务练？\\n\\nImportant text constraints: visible text must be exactly these lines, with no extra words. No other readable text anywhere on the slide besides those four lines. Do not add title text, subtitles, labels, captions, watermarks, logos, UI chrome, or extra numbers/words. The calendar can show visual markers for 30 分钟 and 4-week rhythm using icons, bars, dots, brackets, or stylized marks, but no additional readable text. The sticky notes are the only readable text.\\n\\nVisual style: warm, calm, coherent manga editorial illustration with gentle cream paper background, muted ink lines, soft orange and teal accents, subtle halftone texture, clean composition, clear hierarchy. Use a desk scene with a sense of closure and progress. Show calm confidence, fewer speed lines than energetic slides, and a clear sense of progression over four weeks. Include manga framing and soft expressive character reaction but keep it mature and tidy. Make it visually finished with polished illustration, not a wireframe. Ensure all four note texts are legible at presentation distance.\\n\\nNo markdown, no fake logos, no watermark, no extra captions, no extra text blocks.\",\n      \"storageFileId\": \"qv6tiax8m7v5jri3kffciy1s-69c135520f937f8195a020ed3bfcd908\"\n    },\n    {\n      \"size\": \"1536x864\",\n      \"model\": \"gpt-image-2\",\n      \"title\": \"一句话建立学习系统\",\n      \"layout\": \"终章海报式漫画。主角站在完成的作品墙前，背后是从目标到作品的路径。\",\n      \"status\": \"completed\",\n      \"message\": \"用真实目标发现差距，用具体任务训练能力，用反馈修正方向，用作品证明自己。\",\n      \"imageUrl\": \"https://cdn.drillso.com/generated/drill/dbor32oeat876lytpnrpzxhn-8a313acde88f3a76e1e195ae46ecc9bc.png\",\n      \"storageKey\": \"generated/drill/dbor32oeat876lytpnrpzxhn-8a313acde88f3a76e1e195ae46ecc9bc.png\",\n      \"imagePrompt\": \"Create exactly one complete presentation slide image.\\n\\nUse the following NotebookLM / Kael style prompt as the dominant art direction. Follow it closely.\\n\\nSTYLE TEMPLATE START\\nUnderstanding becomes deeper with fun.\\nTurn the information into a comic-like visual explanation with a simple story.\\nUse panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions.\\nRelate the idea to a realistic situation so the audience can remember it.\\nThe style should be fun and explanatory, not childish or chaotic.\\nSTYLE TEMPLATE END\\n\\nDeck title: 把学习变成能力系统\\nDeck summary: 面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\\nDeck format: presenter_slides\\nChosen style: Manga Style\\n\\nSlide 10: 一句话建立学习系统\\nMain message: 用真实目标发现差距，用具体任务训练能力，用反馈修正方向，用作品证明自己。\\n\\nVisible text must be exactly these lines, with no extra words:\\n目标发现差距\\n任务训练能力\\n反馈修正方向\\n作品证明自己\\n\\nLayout:\\n终章海报式漫画。主角站在完成的作品墙前，背后是从目标到作品的路径。\\n\\nVisual style:\\n强收束海报构图，四个短句像宣言贴在作品墙上，主角表情坚定。\\n\\nAdditional image direction:\\nA final manga poster composition. The learner stands confidently in front of a wall of completed works: report, project board, tool, presentation, article. Behind them, a simple path connects goal, task, feedback, and work. Four bold Chinese phrases appear as clean labels: 目标发现差距, 任务训练能力, 反馈修正方向, 作品证明自己. Make it feel like a memorable ending slide.\\n\\nRequirements:\\n- Canvas: 1536x864 landscape presentation slide.\\n- Generate one slide only, not a contact sheet and not multiple slide candidates.\\n- Make it presentation-ready, high-sensibility, and coherent with the chosen style template.\\n- Make it visually finished, not a sparse wireframe or placeholder diagram.\\n- Use the strongest signature elements of the style template: palette, typography, composition language, icon/illustration treatment, and editorial rhythm.\\n- Keep all visible text readable at presentation distance.\\n- Use low text density, strong hierarchy, and source-grounded diagrams, editorial composition, or template-specific visuals.\\n- Do not add markdown symbols, fake logos, watermarks, signatures, UI browser chrome, or extra captions.\\n- Match the source/user instruction language for all visible slide text and speaker-facing copy.\",\n      \"slideNumber\": 10,\n      \"visibleText\": [\n        \"目标发现差距\",\n        \"任务训练能力\",\n        \"反馈修正方向\",\n        \"作品证明自己\"\n      ],\n      \"visualStyle\": \"强收束海报构图，四个短句像宣言贴在作品墙上，主角表情坚定。\",\n      \"speakerNotes\": \"最后用材料中的总结收束：如果想知道当前会什么，就看能独立完成什么；想知道不会什么，就看在哪些任务中卡住；想知道该学什么，就从目标和差距倒推；想知道怎么应用，就把知识放进真实问题里，做出可见成果。学习不是把知识堆进大脑，而是让自己解决问题的能力越来越强。\",\n      \"revisedPrompt\": \"Create a single 1536x864 landscape presentation slide in a polished manga / NotebookLM-Kael style. Final poster composition, strong closure, editorial manga framing, expressive but not chaotic, high sensibility, presentation-ready.\\n\\nScene: A confident learner stands in the foreground facing slightly left, in front of a completed works wall. The learner has a determined, calm expression, subtle manga shading, clean linework, dynamic but controlled pose. Behind them is a wall of completed outputs displayed like trophies / evidence: a report page, a project board with sticky-note shapes but NO readable text, a tool or app icon-like object, a presentation deck thumbnail, and an article page. Connect them with a simple visible path / flow from left to right or bottom to top: goal -> task -> feedback -> work, shown with arrows, panels, and motion cues, but without any extra words except the required four lines.\\n\\nTypography and visible text: ONLY these four Chinese lines, each as a bold clean label or poster caption, readable at presentation distance, exactly and only these lines:\\n目标发现差距\\n任务训练能力\\n反馈修正方向\\n作品证明自己\\nNo other visible text anywhere. No title, no footer, no watermark, no logos, no UI.\\n\\nLayout: One strong poster-like slide. The four phrases should appear as four bold declaration labels integrated into the wall or as floating comic placards. Strong hierarchy, low text density, clear visual rhythm. Use manga speed lines, expressive framing, and clean comic panel energy, but keep it elegant and coherent. The completed works wall should occupy the background right/center, the learner in foreground left/center. Use dramatic lighting, crisp outlines, warm-neutral palette with high contrast accent colors. Make it feel like the memorable ending slide of a learning system deck.\\n\\nImportant: Do not add any text beyond the four required lines. Do not add extra captions. Do not add watermarks, signatures, borders, browser chrome, or fake UI.\",\n      \"storageFileId\": \"dbor32oeat876lytpnrpzxhn-8a313acde88f3a76e1e195ae46ecc9bc\"\n    }\n  ],\n  \"styleId\": \"manga_story\",\n  \"summary\": \"面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\",\n  \"audience\": \"职场学习者、自我提升者、培训或工作坊参与者\",\n  \"imageSize\": \"1536x864\",\n  \"imageModel\": \"gpt-image-2\",\n  \"plannerModel\": \"gpt-5.5\"\n}","structured":{"style":"Manga Style","title":"把学习变成能力系统","format":"presenter_slides","slides":[{"size":"1536x864","model":"gpt-image-2","title":"不是不努力，是不知道往哪努力","layout":"四格漫画开场。主角坐在书和课程图标堆里，头顶冒出四个问题气泡。最后一格出现一条清晰路径。","status":"completed","message":"学习的第一步，是把模糊焦虑变成清晰问题。","imageUrl":"https://cdn.drillso.com/generated/drill/x6euj249qdncoeyeva8qn63u-ab677975612bfd33a593cdc6a0deef6b.png","storageKey":"generated/drill/x6euj249qdncoeyeva8qn63u-ab677975612bfd33a593cdc6a0deef6b.png","imagePrompt":"Create exactly one complete presentation slide image.\n\nUse the following NotebookLM / Kael style prompt as the dominant art direction. Follow it closely.\n\nSTYLE TEMPLATE START\nUnderstanding becomes deeper with fun.\nTurn the information into a comic-like visual explanation with a simple story.\nUse panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions.\nRelate the idea to a realistic situation so the audience can remember it.\nThe style should be fun and explanatory, not childish or chaotic.\nSTYLE TEMPLATE END\n\nDeck title: 把学习变成能力系统\nDeck summary: 面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\nDeck format: presenter_slides\nChosen style: Manga Style\n\nSlide 1: 不是不努力，是不知道往哪努力\nMain message: 学习的第一步，是把模糊焦虑变成清晰问题。\n\nVisible text must be exactly these lines, with no extra words:\n我到底会什么？\n不会什么？\n该学什么？\n怎么用起来？\n\nLayout:\n四格漫画开场。主角坐在书和课程图标堆里，头顶冒出四个问题气泡。最后一格出现一条清晰路径。\n\nVisual style:\n漫画分镜，表情夸张但不幼稚，问题气泡大，路径线作为视觉主轴。\n\nAdditional image direction:\nA four-panel comic opening scene. A young office learner sits at a desk surrounded by course cards, saved articles, books, and app icons, looking confused. Four speech bubbles contain the visible questions. In the final panel, the clutter transforms into a simple path with six small stepping stones. Use expressive manga reactions, clean panels, speed lines around the confusion, and concise Chinese text.\n\nRequirements:\n- Canvas: 1536x864 landscape presentation slide.\n- Generate one slide only, not a contact sheet and not multiple slide candidates.\n- Make it presentation-ready, high-sensibility, and coherent with the chosen style template.\n- Make it visually finished, not a sparse wireframe or placeholder diagram.\n- Use the strongest signature elements of the style template: palette, typography, composition language, icon/illustration treatment, and editorial rhythm.\n- Keep all visible text readable at presentation distance.\n- Use low text density, strong hierarchy, and source-grounded diagrams, editorial composition, or template-specific visuals.\n- Do not add markdown symbols, fake logos, watermarks, signatures, UI browser chrome, or extra captions.\n- Match the source/user instruction language for all visible slide text and speaker-facing copy.","slideNumber":1,"visibleText":["我到底会什么？","不会什么？","该学什么？","怎么用起来？"],"visualStyle":"漫画分镜，表情夸张但不幼稚，问题气泡大，路径线作为视觉主轴。","speakerNotes":"开场先点出常见困惑：很多人买了很多课、收藏了很多资料，却依然觉得能力模糊。真正的问题不是不努力，而是不知道该往哪里努力。本 deck 会用一条方法链来回答：盘点能力、识别差距、选择目标、设计学习、实践应用、复盘迭代。","revisedPrompt":"Create exactly one complete presentation slide image, landscape 16:9, polished and presentation-ready, in a Manga Style inspired by NotebookLM / Kael editorial comic explanation. Dominant art direction: understanding becomes deeper with fun; turn the information into a comic-like visual explanation with a simple story; use panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions; realistic situation, fun and explanatory, not childish or chaotic.\n\nSlide topic: Chinese learning-and-ability system. Title concept: 把学习变成能力系统. Slide 1 main message: 不是不努力，是不知道往哪努力. The slide should feel like the opening of a story: a young office learner is overwhelmed by ambiguity, then a simple path appears.\n\nLayout: a clean four-panel comic opening scene across the slide, with strong editorial rhythm and clear hierarchy. Panel 1: the learner sits at a desk surrounded by course cards, saved articles, books, and app icons, looking confused, with dramatic manga speed lines and clutter. Panel 2: an expressive close-up reaction with a large speech bubble containing exactly this text: 我到底会什么？ Panel 3: another reaction close-up and question bubble containing exactly this text: 不会什么？ Panel 4: the clutter transforms into a simple clear path with six small stepping stones leading forward, with two more large speech bubbles containing exactly these texts, each on its own line: 该学什么？ and 怎么用起来？\n\nVisible text must be exactly these four lines and no extra words anywhere in the image:\n我到底会什么？\n不会什么？\n该学什么？\n怎么用起来？\n\nImportant constraints: no other visible text, no title text, no subtitles, no labels, no watermark, no logos, no UI chrome, no extra captions. Make the four speech bubbles the only text elements. Chinese text must be large, crisp, and readable at presentation distance. Use clean comic panel borders, bold linework, expressive but mature office-learner character design, and a palette of muted warm beige, off-white, charcoal, desaturated blue, and accent red/orange for emphasis. The final panel should visually resolve the clutter into a simple path as the visual main axis, with six small stepping stones and a forward direction. Make the composition feel finished, balanced, and coherent, with strong manga framing, motion lines, and a high-end slide design aesthetic.","storageFileId":"x6euj249qdncoeyeva8qn63u-ab677975612bfd33a593cdc6a0deef6b"},{"size":"1536x864","model":"gpt-image-2","title":"先区分：知道，不等于会","layout":"竖向升级阶梯。主角从底层听讲，逐步走到顶层教别人，每层用一个简短动作表示。","status":"completed","message":"能力不是一个点，而是从听过到能教别人的层级。","imageUrl":"https://cdn.drillso.com/generated/drill/pdusfddf92up1ielh2rerlng-f2ea5ed653c402ef3317366352e59a05.png","storageKey":"generated/drill/pdusfddf92up1ielh2rerlng-f2ea5ed653c402ef3317366352e59a05.png","imagePrompt":"Create exactly one complete presentation slide image.\n\nUse the following NotebookLM / Kael style prompt as the dominant art direction. Follow it closely.\n\nSTYLE TEMPLATE START\nUnderstanding becomes deeper with fun.\nTurn the information into a comic-like visual explanation with a simple story.\nUse panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions.\nRelate the idea to a realistic situation so the audience can remember it.\nThe style should be fun and explanatory, not childish or chaotic.\nSTYLE TEMPLATE END\n\nDeck title: 把学习变成能力系统\nDeck summary: 面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\nDeck format: presenter_slides\nChosen style: Manga Style\n\nSlide 2: 先区分：知道，不等于会\nMain message: 能力不是一个点，而是从听过到能教别人的层级。\n\nVisible text must be exactly these lines, with no extra words:\n听过\n理解\n模仿\n独立完成\n教别人\n\nLayout:\n竖向升级阶梯。主角从底层听讲，逐步走到顶层教别人，每层用一个简短动作表示。\n\nVisual style:\n升级闯关式漫画阶梯，层级清楚，顶层有聚光效果。\n\nAdditional image direction:\nA manga-style vertical staircase with five labeled levels in Chinese: 听过, 理解, 模仿, 独立完成, 教别人. The same learner character climbs upward: hearing a term, explaining it, following a tutorial, completing a task alone, then teaching a colleague. Keep text sparse, use bold labels, action poses, and clean panel-like separators.\n\nRequirements:\n- Canvas: 1536x864 landscape presentation slide.\n- Generate one slide only, not a contact sheet and not multiple slide candidates.\n- Make it presentation-ready, high-sensibility, and coherent with the chosen style template.\n- Make it visually finished, not a sparse wireframe or placeholder diagram.\n- Use the strongest signature elements of the style template: palette, typography, composition language, icon/illustration treatment, and editorial rhythm.\n- Keep all visible text readable at presentation distance.\n- Use low text density, strong hierarchy, and source-grounded diagrams, editorial composition, or template-specific visuals.\n- Do not add markdown symbols, fake logos, watermarks, signatures, UI browser chrome, or extra captions.\n- Match the source/user instruction language for all visible slide text and speaker-facing copy.","slideNumber":2,"visibleText":["听过","理解","模仿","独立完成","教别人"],"visualStyle":"升级闯关式漫画阶梯，层级清楚，顶层有聚光效果。","speakerNotes":"很多人会把听过、看过误认为自己会了。材料把能力分成五层：听过、理解、模仿、独立完成、教别人。判断自己会不会，不要只问我学过吗，而要看能否不用资料独立完成，换场景还能不能用，出错时能不能排查，能不能教给别人，结果是否稳定可靠。","revisedPrompt":"Create one complete presentation slide image, 16:9 landscape, 1536x864, in a polished manga-style NotebookLM/Kael-inspired editorial illustration. Dominant art direction: understanding becomes deeper with fun; turn the information into a comic-like visual explanation with a simple story; use panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions; relate the idea to a realistic workplace learning situation; fun and explanatory, not childish or chaotic.\n\nSlide concept: a vertical upgrade staircase showing the progression from hearing a term to teaching it. The SAME learner character climbs upward through five clear levels, each level represented as a step/panel with a distinct action pose. Composition should feel like a clean manga challenge ladder with strong hierarchy and panel-like separators, dynamic angle, and a top spotlight at the highest level.\n\nVisible text must be EXACTLY these five Chinese labels and nothing else anywhere in the slide:\n听过\n理解\n模仿\n独立完成\n教别人\n\nLayout details: a bold vertical staircase rising from bottom-left to top-right or straight upward within the slide, with five large steps. Each step contains the learner in a different state: bottom hears a term with a small speech bubble-like reaction icon (no text besides the label); next explains/understands with a thoughtful pose; next follows a tutorial with hands and eyes focused; next completes a task alone with confidence; top teaches a colleague under a bright spotlight. Use clean manga framing around each level, speed lines indicating upward progress, expressive facial expressions, and subtle office-learning context such as a desk, laptop, sticky notes, document, whiteboard, and colleague silhouette. Keep the slide elegant and readable at presentation distance. Use bold Chinese labels placed clearly on each step, large enough to read.\n\nColor and rendering: manga style with crisp black ink lines, high-contrast shading, tasteful limited palette, editorial composition, subtle warm accent colors and blue-gray shadows, white/light background with dynamic diagonal shapes, polished presentation-ready finish. Avoid clutter, avoid extra captions, avoid any text beyond the five labels, avoid logos, watermarks, UI, or decorative paragraphs. Make it look like a finished single keynote slide, not a wireframe or contact sheet.","storageFileId":"pdusfddf92up1ielh2rerlng-f2ea5ed653c402ef3317366352e59a05"},{"size":"1536x864","model":"gpt-image-2","title":"能力必须有证据","layout":"桌面俯视漫画。一本打开的能力清单笔记本占主体，四个栏目像表格贴纸一样排布。","status":"completed","message":"不要写“我会”，要写“我做成过什么”。","imageUrl":"https://cdn.drillso.com/generated/drill/c9nu33ayugtpvkkfey1cxq9g-442834a751b795f701ca120cf9eff850.png","storageKey":"generated/drill/c9nu33ayugtpvkkfey1cxq9g-442834a751b795f701ca120cf9eff850.png","imagePrompt":"Create exactly one complete presentation slide image.\n\nUse the following NotebookLM / Kael style prompt as the dominant art direction. Follow it closely.\n\nSTYLE TEMPLATE START\nUnderstanding becomes deeper with fun.\nTurn the information into a comic-like visual explanation with a simple story.\nUse panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions.\nRelate the idea to a realistic situation so the audience can remember it.\nThe style should be fun and explanatory, not childish or chaotic.\nSTYLE TEMPLATE END\n\nDeck title: 把学习变成能力系统\nDeck summary: 面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\nDeck format: presenter_slides\nChosen style: Manga Style\n\nSlide 3: 能力必须有证据\nMain message: 不要写“我会”，要写“我做成过什么”。\n\nVisible text must be exactly these lines, with no extra words:\n当前水平\n证据\n问题\n下一步\n\nLayout:\n桌面俯视漫画。一本打开的能力清单笔记本占主体，四个栏目像表格贴纸一样排布。\n\nVisual style:\n手账式漫画信息图，证据栏用放大镜强调，旁边有主角认真记录。\n\nAdditional image direction:\nTop-down manga desk scene with an open notebook titled 能力清单. Four large clean fields are visible: 当前水平, 证据, 问题, 下一步. A magnifying glass highlights the 证据 field. Around the notebook, small icons represent writing, Excel, communication, project delivery, and AI tools. Use clear hierarchy and minimal Chinese text.\n\nRequirements:\n- Canvas: 1536x864 landscape presentation slide.\n- Generate one slide only, not a contact sheet and not multiple slide candidates.\n- Make it presentation-ready, high-sensibility, and coherent with the chosen style template.\n- Make it visually finished, not a sparse wireframe or placeholder diagram.\n- Use the strongest signature elements of the style template: palette, typography, composition language, icon/illustration treatment, and editorial rhythm.\n- Keep all visible text readable at presentation distance.\n- Use low text density, strong hierarchy, and source-grounded diagrams, editorial composition, or template-specific visuals.\n- Do not add markdown symbols, fake logos, watermarks, signatures, UI browser chrome, or extra captions.\n- Match the source/user instruction language for all visible slide text and speaker-facing copy.","slideNumber":3,"visibleText":["当前水平","证据","问题","下一步"],"visualStyle":"手账式漫画信息图，证据栏用放大镜强调，旁边有主角认真记录。","speakerNotes":"盘点能力时，要建立能力清单，把笼统能力拆成具体项目。材料建议可以按专业技能、通用能力、工具能力、业务理解、交付能力来拆。关键是能力必须有证据。不要只写我会演讲，而要写能在 20 人会议上做 10 分钟汇报，并回答 3 个问题。","revisedPrompt":"Create exactly one complete presentation slide image, 1536x864 landscape, in a NotebookLM / Kael inspired manga style with a fun explanatory comic-like visual language. Top-down desk scene, editorial and polished, not childish, not chaotic. Dominant art direction: understanding becomes deeper with fun; simple story, panels, character reactions, manga framing, speed lines, expressive close-ups, concise captions. \n\nSlide 3 topic: 能力必须有证据. Main message should be visually communicated: 不要写“我会”，要写“我做成过什么”。 \n\nVisible text must be exactly these lines and no other text anywhere:\n当前水平\n证据\n问题\n下一步\n\nLayout: desktop top-down manga infographic. An open notebook dominates the center and is titled 能力清单 in a clean handwritten manga style. Inside the notebook, four large clean field areas are arranged like sticker-like table sections, with the visible text exactly: 当前水平, 证据, 问题, 下一步. The 证据 field is emphasized by a magnifying glass placed over it, with strong graphic focus, highlight ring, and subtle speed lines. A thoughtful protagonist hand/character is nearby, seriously noting things down. Around the notebook, small icons illustrate writing, Excel, communication, project delivery, and AI tools, arranged neatly like desk items. \n\nArt direction: hand-drawn manga editorial infographic, clean composition, strong hierarchy, minimal Chinese text, presentation-ready, coherent, high sensibility, realistic desk objects combined with comic elements. Use palette and typography inspired by Japanese manga notes: off-white paper, warm beige desk, black ink lines, muted teal and orange accents, clean shadows, crisp legibility. Add subtle panel-like framing and expressive close-up feel without splitting into multiple slide candidates. Make the evidence area the visual focal point. \n\nImportant constraints: no extra captions, no watermarks, no logos, no browser chrome, no UI mockups, no markdown symbols, no extra words beyond the four required visible text lines; ensure those text lines are readable at presentation distance. The notebook title 能力清单 can appear as part of the notebook cover/title. Make the slide feel finished and coherent, not sparse or placeholder-like.","storageFileId":"c9nu33ayugtpvkkfey1cxq9g-442834a751b795f701ca120cf9eff850"},{"size":"1536x864","model":"gpt-image-2","title":"不会的地方，藏在任务里","layout":"四宫格侦探漫画。主角像侦探一样寻找盲区，每格对应一种方法。","status":"completed","message":"真正的差距，会在真实任务中暴露出来。","imageUrl":"https://cdn.drillso.com/generated/drill/h1er9ut8gu1bakkopkd0flmk-fdf0f5f5eeffbbff9435605789265fb2.png","storageKey":"generated/drill/h1er9ut8gu1bakkopkd0flmk-fdf0f5f5eeffbbff9435605789265fb2.png","imagePrompt":"Create exactly one complete presentation slide image.\n\nUse the following NotebookLM / Kael style prompt as the dominant art direction. Follow it closely.\n\nSTYLE TEMPLATE START\nUnderstanding becomes deeper with fun.\nTurn the information into a comic-like visual explanation with a simple story.\nUse panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions.\nRelate the idea to a realistic situation so the audience can remember it.\nThe style should be fun and explanatory, not childish or chaotic.\nSTYLE TEMPLATE END\n\nDeck title: 把学习变成能力系统\nDeck summary: 面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\nDeck format: presenter_slides\nChosen style: Manga Style\n\nSlide 4: 不会的地方，藏在任务里\nMain message: 真正的差距，会在真实任务中暴露出来。\n\nVisible text must be exactly these lines, with no extra words:\n对标\n任务\n提问\n反馈\n\nLayout:\n四宫格侦探漫画。主角像侦探一样寻找盲区，每格对应一种方法。\n\nVisual style:\n侦探调查感，放大镜、脚印、问号和评分牌作为符号。\n\nAdditional image direction:\nA four-panel manga detective layout. The learner wears a simple detective cap and investigates four clues labeled 对标, 任务, 提问, 反馈. Panel 1 compares their work with an expert article. Panel 2 shows a sales data task with a stuck expression. Panel 3 shows question marks around core concepts. Panel 4 shows a colleague giving a score card. Keep labels large and text minimal.\n\nRequirements:\n- Canvas: 1536x864 landscape presentation slide.\n- Generate one slide only, not a contact sheet and not multiple slide candidates.\n- Make it presentation-ready, high-sensibility, and coherent with the chosen style template.\n- Make it visually finished, not a sparse wireframe or placeholder diagram.\n- Use the strongest signature elements of the style template: palette, typography, composition language, icon/illustration treatment, and editorial rhythm.\n- Keep all visible text readable at presentation distance.\n- Use low text density, strong hierarchy, and source-grounded diagrams, editorial composition, or template-specific visuals.\n- Do not add markdown symbols, fake logos, watermarks, signatures, UI browser chrome, or extra captions.\n- Match the source/user instruction language for all visible slide text and speaker-facing copy.","slideNumber":4,"visibleText":["对标","任务","提问","反馈"],"visualStyle":"侦探调查感，放大镜、脚印、问号和评分牌作为符号。","speakerNotes":"识别不会什么，可以用四种方法。对标法是看高手的成果、流程和思考方式。任务法是用真实任务测试，比如拿销售数据找出下降原因并提出建议。提问法是用核心概念、关系、异常处理等问题暴露盲区。反馈法是请同事、老师或朋友指出哪里好、哪里不清楚、哪里效率低。","revisedPrompt":"Create exactly one complete presentation slide image, landscape 1536x864, in a polished Manga Style / NotebookLM / Kael style. Dominant art direction: understanding becomes deeper with fun; comic-like explanatory visual with a simple story, panels, character reactions, manga framing, speed lines, expressive close-ups, concise captions. Serious, fun, explanatory, not childish. \n\nDeck slide topic: \"不会的地方，藏在任务里\" with main message: truly real gaps are exposed in real tasks.\n\nDesign a single finished presentation slide, not a contact sheet, not multiple candidate slides. Four-panel detective manga layout, coherent editorial composition. The learner is a detective-like main character wearing a simple detective cap, investigating four clues. Visual style: detective investigation vibe, magnifying glass, footprints, question marks, and score card symbols. Include a realistic office/work-learning setting. \n\nVisible text must be exactly these four Chinese lines and nothing else, large and readable, each as a label in one panel: \n对标\n任务\n提问\n反馈\n\nNo other visible text anywhere. No title, no subtitle, no extra captions, no watermark, no UI. Keep text minimal and clean.\n\nPanel 1: learner compares their work with an expert article / benchmark material, using a magnifying glass over a document, subtle comparison cues. Label: 对标.\nPanel 2: learner confronts a sales data task, stuck expression, evidence of uncertainty, files/spreadsheet vibe, detective examining task clues. Label: 任务.\nPanel 3: question marks swirl around core concepts, learner thinking and asking precise questions, manga close-up reaction. Label: 提问.\nPanel 4: colleague gives a score card / feedback card, learner receiving constructive feedback, modestly relieved expression. Label: 反馈.\n\nUse a four-panel comic grid with strong manga framing, varied camera angles, expressive faces, speed lines, inked outlines, halftone shading, dynamic but clean composition. Presentation-ready, high-sensibility, visually finished. Use a restrained palette: off-white paper, black ink, muted teal, warm orange accents, and red highlights for emphasis. Ensure the four labels are clearly legible and the overall slide feels cohesive and professional.","storageFileId":"h1er9ut8gu1bakkopkd0flmk-fdf0f5f5eeffbbff9435605789265fb2"},{"size":"1536x864","model":"gpt-image-2","title":"该学什么，从目标倒推","layout":"横向流程漫画。主角从左到右推着一个目标牌，经过六个节点，最后进入真实应用场景。","status":"completed","message":"不要被热门牵着走，要从真实目标拆出关键差距。","imageUrl":"https://cdn.drillso.com/generated/drill/pigjfvkup3iucez8bovjfolb-8df686416354c784ce6a5d2632631f54.png","storageKey":"generated/drill/pigjfvkup3iucez8bovjfolb-8df686416354c784ce6a5d2632631f54.png","imagePrompt":"Create exactly one complete presentation slide image.\n\nUse the following NotebookLM / Kael style prompt as the dominant art direction. Follow it closely.\n\nSTYLE TEMPLATE START\nUnderstanding becomes deeper with fun.\nTurn the information into a comic-like visual explanation with a simple story.\nUse panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions.\nRelate the idea to a realistic situation so the audience can remember it.\nThe style should be fun and explanatory, not childish or chaotic.\nSTYLE TEMPLATE END\n\nDeck title: 把学习变成能力系统\nDeck summary: 面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\nDeck format: presenter_slides\nChosen style: Manga Style\n\nSlide 5: 该学什么，从目标倒推\nMain message: 不要被热门牵着走，要从真实目标拆出关键差距。\n\nVisible text must be exactly these lines, with no extra words:\n目标\n任务\n能力\n差距\n计划\n应用\n\nLayout:\n横向流程漫画。主角从左到右推着一个目标牌，经过六个节点，最后进入真实应用场景。\n\nVisual style:\n流程图和漫画动作结合，节点像游戏关卡，箭头清晰。\n\nAdditional image direction:\nA horizontal manga flow diagram with the learner pushing a sign labeled 目标 through six connected nodes: 目标, 任务, 能力, 差距, 计划, 应用. Avoid dense text. The background hints at a workplace project report scenario. Use dynamic arrows, speed lines, and clear game-level checkpoints.\n\nRequirements:\n- Canvas: 1536x864 landscape presentation slide.\n- Generate one slide only, not a contact sheet and not multiple slide candidates.\n- Make it presentation-ready, high-sensibility, and coherent with the chosen style template.\n- Make it visually finished, not a sparse wireframe or placeholder diagram.\n- Use the strongest signature elements of the style template: palette, typography, composition language, icon/illustration treatment, and editorial rhythm.\n- Keep all visible text readable at presentation distance.\n- Use low text density, strong hierarchy, and source-grounded diagrams, editorial composition, or template-specific visuals.\n- Do not add markdown symbols, fake logos, watermarks, signatures, UI browser chrome, or extra captions.\n- Match the source/user instruction language for all visible slide text and speaker-facing copy.","slideNumber":5,"visibleText":["目标","任务","能力","差距","计划","应用"],"visualStyle":"流程图和漫画动作结合，节点像游戏关卡，箭头清晰。","speakerNotes":"材料指出，常见误区是看到什么热门就学什么，今天学 AI，明天学剪辑，后天学投资，最后没有形成能力闭环。更有效的方法是从目标倒推：明确目标，拆解任务，识别所需能力，评估当前水平，找出关键差距，制定学习计划，再真实应用并复盘改进。","revisedPrompt":"Create a single presentation slide image, 1536x864 landscape, in a manga-style NotebookLM / Kael editorial comic format. Dominant art direction: understanding becomes deeper with fun; turn the information into a comic-like visual explanation with a simple story; use panels, expressive reactions, manga framing, speed lines, concise captions, but keep it clean, fun, explanatory, not childish or chaotic.\n\nSlide topic: '该学什么，从目标倒推'. Main message:不要被热门牵着走，要从真实目标拆出关键差距.\n\nComposition: a horizontal manga flow diagram from left to right. A learner character pushes a large sign labeled exactly '目标' through six connected game-like checkpoint nodes, then into a realistic workplace project report scene at the far right. The six nodes must be clearly visible and the ONLY visible text on the slide must be these exact six lines, each appearing once and nothing else:\n目标\n任务\n能力\n差距\n计划\n应用\n\nLayout details: The learner begins on the left with determined expression, pushing the '目标' sign forward through connected nodes like level checkpoints. Each node is a stylized rounded rectangle or game stage badge with manga framing and dynamic arrows. Show the flow progression: 目标 → 任务 → 能力 → 差距 → 计划 → 应用. Use clear arrows, speed lines, motion effects, and expressive close-ups of the learner at a few stages. At the rightmost end, the flow enters a realistic workplace project report application setting: desk, laptop, charts, meeting board, documents, office environment hinted in the background, showing real-world use. Background should subtly hint at a work project report scenario without adding extra text.\n\nVisual style: manga style with editorial composition, bold but tasteful black linework, vivid yet restrained colors, strong hierarchy, game-level checkpoint visuals, clean readable typography for the Chinese node labels, presentation-ready, polished, high contrast, coherent, no clutter, no extra words, no watermark, no logo, no browser chrome. Ensure text is legible at presentation distance. Keep low text density and strong visual rhythm. Make it look like a finished slide, not a wireframe. Avoid any additional captions or labels beyond the six required words.","storageFileId":"pigjfvkup3iucez8bovjfolb-8df686416354c784ce6a5d2632631f54"},{"size":"1536x864","model":"gpt-image-2","title":"优先级，看三件事","layout":"漫画仪表盘。三个大旋钮分别代表重要性、紧急性、杠杆性，主角把学习项目卡片放入评分机。","status":"completed","message":"不是所有不会的东西都要马上学。","imageUrl":"https://cdn.drillso.com/generated/drill/bskvd1s25xzz10oiu38tsxh7-e01f9c917a2b1fcce79a9a65e6f5fc81.png","storageKey":"generated/drill/bskvd1s25xzz10oiu38tsxh7-e01f9c917a2b1fcce79a9a65e6f5fc81.png","imagePrompt":"Create exactly one complete presentation slide image.\n\nUse the following NotebookLM / Kael style prompt as the dominant art direction. Follow it closely.\n\nSTYLE TEMPLATE START\nUnderstanding becomes deeper with fun.\nTurn the information into a comic-like visual explanation with a simple story.\nUse panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions.\nRelate the idea to a realistic situation so the audience can remember it.\nThe style should be fun and explanatory, not childish or chaotic.\nSTYLE TEMPLATE END\n\nDeck title: 把学习变成能力系统\nDeck summary: 面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\nDeck format: presenter_slides\nChosen style: Manga Style\n\nSlide 6: 优先级，看三件事\nMain message: 不是所有不会的东西都要马上学。\n\nVisible text must be exactly these lines, with no extra words:\n重要性\n紧急性\n杠杆性\n\nLayout:\n漫画仪表盘。三个大旋钮分别代表重要性、紧急性、杠杆性，主角把学习项目卡片放入评分机。\n\nVisual style:\n轻量信息仪表盘，三个维度占画面中心，项目卡片作为辅助元素。\n\nAdditional image direction:\nA manga-style priority scoring machine. Three large dials are labeled 重要性, 紧急性, 杠杆性. The learner feeds skill cards into the machine, such as Excel 数据清洗 and 汇报表达, while low-priority cards sit to the side. Use bold dial shapes, minimal visible text, and a clear sense of ranking without adding extra numbers.\n\nRequirements:\n- Canvas: 1536x864 landscape presentation slide.\n- Generate one slide only, not a contact sheet and not multiple slide candidates.\n- Make it presentation-ready, high-sensibility, and coherent with the chosen style template.\n- Make it visually finished, not a sparse wireframe or placeholder diagram.\n- Use the strongest signature elements of the style template: palette, typography, composition language, icon/illustration treatment, and editorial rhythm.\n- Keep all visible text readable at presentation distance.\n- Use low text density, strong hierarchy, and source-grounded diagrams, editorial composition, or template-specific visuals.\n- Do not add markdown symbols, fake logos, watermarks, signatures, UI browser chrome, or extra captions.\n- Match the source/user instruction language for all visible slide text and speaker-facing copy.","slideNumber":6,"visibleText":["重要性","紧急性","杠杆性"],"visualStyle":"轻量信息仪表盘，三个维度占画面中心，项目卡片作为辅助元素。","speakerNotes":"判断学习优先级，可以用三维评分法。第一是重要性：是否影响核心目标。第二是紧急性：近期是否马上会用到。第三是杠杆性：学会后是否能带来长期收益。总分越高，越应该优先学。材料中的例子显示，Excel 数据清洗和汇报表达因为得分高，更应该优先学习。","revisedPrompt":"Create exactly one complete presentation slide image, landscape 1536x864, in a manga-style NotebookLM / Kael explanatory comic style. Theme: priority scoring machine for learning choices. The composition is a polished presentation slide, not a wireframe, not a contact sheet, not multiple panels. Use a fun, explanatory, editorial manga look with bold shapes, expressive reactions, speed lines, clean panel framing hints, and a realistic office-learning situation.\n\nScene: A learner at a desk feeds skill cards into a large priority-scoring machine in the center of the slide. The machine has THREE huge bold dials taking up the visual center, each clearly labeled with exactly one of the following Chinese words only: 重要性, 紧急性, 杠杆性. These are the ONLY visible words anywhere in the slide. No title, no subtitle, no extra captions, no numbers, no labels on cards, no watermarks, no UI chrome.\n\nVisual structure: Center the three dials prominently like a comic dashboard. The learner character is in a clear manga pose, slightly tense but thoughtful, inserting cards into the machine. The cards are symbolic only, with simple icon-like visuals to suggest work skills, but NO text on the cards. One or two low-priority cards sit aside in a dimmer area, also without text. Show the machine ranking the cards by visual emphasis: one card is highlighted going into the machine, another sits aside in muted tone. Use strong hierarchy and a sense of sorting/decision-making.\n\nArt direction: Light information dashboard, manga framing, expressive close-up feel, clean and presentation-ready. High contrast, crisp line art, modern comic shading, subtle motion lines, dynamic but not chaotic. Color palette: warm cream background, ink black lines, muted teal and orange accent colors, with bright highlights for the active card and dials. Keep it elegant, educational, and easy to read at presentation distance.\n\nImportant constraints: Only the three Chinese dial labels may appear as text. Absolutely no other visible text. Do not add extra words in any language. Do not include logos, signatures, or watermarks. Make it visually finished and coherent, with a strong sense of ranking and prioritization.","storageFileId":"bskvd1s25xzz10oiu38tsxh7-e01f9c917a2b1fcce79a9a65e6f5fc81"},{"size":"1536x864","model":"gpt-image-2","title":"有效学习，是一个闭环","layout":"环形闭环漫画图。四个节点围成循环，中间是一个逐渐成形的作品。","status":"completed","message":"只看资料不够，要让知识变成成果。","imageUrl":"https://cdn.drillso.com/generated/drill/x194b29yzfoalt2z4pfn81mp-b0d8e1614c3cb6dfd8a24f12a5b17f98.png","storageKey":"generated/drill/x194b29yzfoalt2z4pfn81mp-b0d8e1614c3cb6dfd8a24f12a5b17f98.png","imagePrompt":"Create exactly one complete presentation slide image.\n\nUse the following NotebookLM / Kael style prompt as the dominant art direction. Follow it closely.\n\nSTYLE TEMPLATE START\nUnderstanding becomes deeper with fun.\nTurn the information into a comic-like visual explanation with a simple story.\nUse panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions.\nRelate the idea to a realistic situation so the audience can remember it.\nThe style should be fun and explanatory, not childish or chaotic.\nSTYLE TEMPLATE END\n\nDeck title: 把学习变成能力系统\nDeck summary: 面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\nDeck format: presenter_slides\nChosen style: Manga Style\n\nSlide 7: 有效学习，是一个闭环\nMain message: 只看资料不够，要让知识变成成果。\n\nVisible text must be exactly these lines, with no extra words:\n输入\n练习\n反馈\n输出\n\nLayout:\n环形闭环漫画图。四个节点围成循环，中间是一个逐渐成形的作品。\n\nVisual style:\n循环动作分镜，中心作品发光，强调从资料到成果的转化。\n\nAdditional image direction:\nA circular manga infographic with four connected nodes: 输入, 练习, 反馈, 输出. In the center, a rough idea transforms into a finished work artifact, such as a report or project board. Show books and courses at input, pen and keyboard at practice, comment bubbles at feedback, and a completed deliverable at output. Keep text sparse and hierarchy strong.\n\nRequirements:\n- Canvas: 1536x864 landscape presentation slide.\n- Generate one slide only, not a contact sheet and not multiple slide candidates.\n- Make it presentation-ready, high-sensibility, and coherent with the chosen style template.\n- Make it visually finished, not a sparse wireframe or placeholder diagram.\n- Use the strongest signature elements of the style template: palette, typography, composition language, icon/illustration treatment, and editorial rhythm.\n- Keep all visible text readable at presentation distance.\n- Use low text density, strong hierarchy, and source-grounded diagrams, editorial composition, or template-specific visuals.\n- Do not add markdown symbols, fake logos, watermarks, signatures, UI browser chrome, or extra captions.\n- Match the source/user instruction language for all visible slide text and speaker-facing copy.","slideNumber":7,"visibleText":["输入","练习","反馈","输出"],"visualStyle":"循环动作分镜，中心作品发光，强调从资料到成果的转化。","speakerNotes":"材料把有效学习概括为输入、练习、反馈、输出的闭环。输入包括看书、听课、读文章、观察案例、请教他人，但要有边界，先学够用的最小知识。练习是把知识变成动作。反馈用来发现偏差。输出是最可靠的检验方式，可以是一篇文章、一个项目、一次汇报、一个工具或一个真实问题的解决方案。","revisedPrompt":"Create exactly one complete presentation slide image in 16:9 landscape, 1536x864. Dominant art direction: NotebookLM / Kael style, manga-style explanatory comic infographic, fun but polished, not childish. Strong editorial composition, clear hierarchy, low text density, high sensibility, presentation-ready.\n\nSlide concept: a circular closed-loop learning system. Four connected nodes arranged in a ring around a glowing center. Center shows a rough idea transforming into a finished work artifact, like a polished report or project board, with a bright glow and subtle motion lines to show transformation from vague draft to completed output.\n\nVisual structure:\n- Circular loop / cycle infographic in the middle of the slide.\n- Four manga comic nodes evenly spaced around the circle with arrows or curved connectors forming a loop.\n- Node 1: books, course cards, and reading materials to represent input.\n- Node 2: pen, notebook, and keyboard to represent practice.\n- Node 3: comment bubbles, review marks, and mentor-like feedback to represent feedback.\n- Node 4: completed deliverable, clean report/project board to represent output.\n- Center: an in-progress idea becoming a finished artifact, luminous, with a subtle burst effect.\n\nStyle details:\n- Manga framing, expressive motion lines, dynamic but clean composition.\n- Use comic panel energy within a single infographic, with connected vignette-like nodes.\n- Palette: high-contrast editorial manga colors, mostly warm cream background with black inks and selective accent colors like teal, orange, and gold glow.\n- Use crisp, readable Chinese typography integrated into the node labels.\n- Keep the slide uncluttered and visually cohesive.\n\nVisible text must be exactly these four lines and nothing else:\n输入\n练习\n反馈\n输出\n\nNo extra text, no title, no footer, no watermark, no UI chrome, no logos, no captions beyond the four required labels. Make the four labels clearly readable and the only text on the slide.","storageFileId":"x194b29yzfoalt2z4pfn81mp-b0d8e1614c3cb6dfd8a24f12a5b17f98"},{"size":"1536x864","model":"gpt-image-2","title":"把学习绑到真实场景","layout":"三段式漫画。左边是问题，中间是作品，右边是用户在场景中使用作品。","status":"completed","message":"应用不是记笔记，而是解决一个看得见的问题。","imageUrl":"https://cdn.drillso.com/generated/drill/e7wzw8m3ktcode3m8xg3p68s-a566d46672a8045db28f95956b376cfc.png","storageKey":"generated/drill/e7wzw8m3ktcode3m8xg3p68s-a566d46672a8045db28f95956b376cfc.png","imagePrompt":"Create exactly one complete presentation slide image.\n\nUse the following NotebookLM / Kael style prompt as the dominant art direction. Follow it closely.\n\nSTYLE TEMPLATE START\nUnderstanding becomes deeper with fun.\nTurn the information into a comic-like visual explanation with a simple story.\nUse panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions.\nRelate the idea to a realistic situation so the audience can remember it.\nThe style should be fun and explanatory, not childish or chaotic.\nSTYLE TEMPLATE END\n\nDeck title: 把学习变成能力系统\nDeck summary: 面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\nDeck format: presenter_slides\nChosen style: Manga Style\n\nSlide 8: 把学习绑到真实场景\nMain message: 应用不是记笔记，而是解决一个看得见的问题。\n\nVisible text must be exactly these lines, with no extra words:\n真实问题\n具体作品\n使用场景\n\nLayout:\n三段式漫画。左边是问题，中间是作品，右边是用户在场景中使用作品。\n\nVisual style:\n产品小剧场感，三个绑定像三张卡片，用线连接成因果链。\n\nAdditional image direction:\nA three-part manga scene connected by arrows. Card 1 labeled 真实问题 shows messy weekly sales data. Card 2 labeled 具体作品 shows an automation script or clean report. Card 3 labeled 使用场景 shows someone using the report in a meeting. Include a small AI tool icon in the background but no brand logos. Minimal Chinese text, expressive characters, clear cause-and-effect.\n\nRequirements:\n- Canvas: 1536x864 landscape presentation slide.\n- Generate one slide only, not a contact sheet and not multiple slide candidates.\n- Make it presentation-ready, high-sensibility, and coherent with the chosen style template.\n- Make it visually finished, not a sparse wireframe or placeholder diagram.\n- Use the strongest signature elements of the style template: palette, typography, composition language, icon/illustration treatment, and editorial rhythm.\n- Keep all visible text readable at presentation distance.\n- Use low text density, strong hierarchy, and source-grounded diagrams, editorial composition, or template-specific visuals.\n- Do not add markdown symbols, fake logos, watermarks, signatures, UI browser chrome, or extra captions.\n- Match the source/user instruction language for all visible slide text and speaker-facing copy.","slideNumber":8,"visibleText":["真实问题","具体作品","使用场景"],"visualStyle":"产品小剧场感，三个绑定像三张卡片，用线连接成因果链。","speakerNotes":"学完不用，很快会忘。材料建议应用能力时做三个绑定：绑定一个真实问题，绑定一个具体作品，绑定一个使用场景。比如不要只说我要学 Python，而是说我要用 Python 自动整理每周销售数据，减少手工复制时间。学习 AI 工具也不要只研究提示词技巧，而要放到写周报、做调研、学英语或写代码等具体场景里。","revisedPrompt":"Create exactly one complete presentation slide image in 1536x864 landscape, manga-style educational editorial slide, inspired by NotebookLM / Kael style: fun explanatory comic with simple story, panels, character reactions, manga framing, speed lines, expressive close-ups, concise captions, polished and presentation-ready. The slide topic is: 把学习绑到真实场景. Main message visualized: 应用不是记笔记，而是解决一个看得见的问题.\n\nLayout: three-part manga scene arranged left to right as three connected cards, with arrows/line connectors showing a cause-and-effect chain. Product theater feeling, like three linked cards. Left card = messy weekly sales problem, middle card = clean output/automation script/report, right card = person using the report in a meeting. Add a small AI tool icon in the background, abstract and generic, no brand logos.\n\nVisible text must be exactly these three lines and nothing else anywhere on the slide:\n真实问题\n具体作品\n使用场景\n\nPlace the text as the card headings only, large, readable, and centered or top-left on each card. No other Chinese or English text, no extra captions, no watermarks, no UI chrome, no logo.\n\nArt direction: manga-style but not childish or chaotic; modern editorial palette with strong contrast, clean composition, subtle halftone textures, speed lines, expressive character faces. Left card should show a realistic messy weekly sales data situation: scattered numbers, charts, papers, and a stressed office worker. Middle card should show a neat artifact: a clean report dashboard or automation script flowing into a tidy summary sheet, with a focused character producing it. Right card should show a confident person presenting/using the report in a meeting, with colleagues reacting positively. Use arrows connecting the cards. Keep low text density, strong hierarchy, and clear cause-and-effect. Make it visually finished, balanced, and high-sensibility, suitable for a deck slide.\n\nDo not include any other visible text beyond the exact three headings listed above.","storageFileId":"e7wzw8m3ktcode3m8xg3p68s-a566d46672a8045db28f95956b376cfc"},{"size":"1536x864","model":"gpt-image-2","title":"每周 30 分钟，复盘能力边界","layout":"日历和复盘表漫画。主角周末坐下复盘，四个问题以便签形式围绕日历。","status":"completed","message":"持续复盘，才能看清自己能独立完成什么、在哪些地方卡住。","imageUrl":"https://cdn.drillso.com/generated/drill/qv6tiax8m7v5jri3kffciy1s-69c135520f937f8195a020ed3bfcd908.png","storageKey":"generated/drill/qv6tiax8m7v5jri3kffciy1s-69c135520f937f8195a020ed3bfcd908.png","imagePrompt":"Create exactly one complete presentation slide image.\n\nUse the following NotebookLM / Kael style prompt as the dominant art direction. Follow it closely.\n\nSTYLE TEMPLATE START\nUnderstanding becomes deeper with fun.\nTurn the information into a comic-like visual explanation with a simple story.\nUse panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions.\nRelate the idea to a realistic situation so the audience can remember it.\nThe style should be fun and explanatory, not childish or chaotic.\nSTYLE TEMPLATE END\n\nDeck title: 把学习变成能力系统\nDeck summary: 面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\nDeck format: presenter_slides\nChosen style: Manga Style\n\nSlide 9: 每周 30 分钟，复盘能力边界\nMain message: 持续复盘，才能看清自己能独立完成什么、在哪些地方卡住。\n\nVisible text must be exactly these lines, with no extra words:\n完成了什么？\n哪里卡住？\n下周补什么？\n用什么任务练？\n\nLayout:\n日历和复盘表漫画。主角周末坐下复盘，四个问题以便签形式围绕日历。\n\nVisual style:\n温和收束，便签式问题，日历上标出 30 分钟和 4 周节奏。\n\nAdditional image direction:\nA closing manga scene of the learner doing a weekly review at a desk. A calendar shows a repeating weekly rhythm and a small 30 分钟 marker. Four sticky notes around the calendar read: 完成了什么？ 哪里卡住？ 下周补什么？ 用什么任务练？ Show calm confidence, fewer speed lines, and a clear sense of progress over four weeks.\n\nRequirements:\n- Canvas: 1536x864 landscape presentation slide.\n- Generate one slide only, not a contact sheet and not multiple slide candidates.\n- Make it presentation-ready, high-sensibility, and coherent with the chosen style template.\n- Make it visually finished, not a sparse wireframe or placeholder diagram.\n- Use the strongest signature elements of the style template: palette, typography, composition language, icon/illustration treatment, and editorial rhythm.\n- Keep all visible text readable at presentation distance.\n- Use low text density, strong hierarchy, and source-grounded diagrams, editorial composition, or template-specific visuals.\n- Do not add markdown symbols, fake logos, watermarks, signatures, UI browser chrome, or extra captions.\n- Match the source/user instruction language for all visible slide text and speaker-facing copy.","slideNumber":9,"visibleText":["完成了什么？","哪里卡住？","下周补什么？","用什么任务练？"],"visualStyle":"温和收束，便签式问题，日历上标出 30 分钟和 4 周节奏。","speakerNotes":"材料给出一个每周复盘模板，建议每周花 30 分钟回答：本周完成了什么具体任务，哪些能独立完成，哪些地方卡住了，卡住是因为概念不懂、方法不会还是练习太少，收到了哪些反馈，下周最应该补哪一个短板，以及用什么真实任务来练习。持续 4 周，会更清楚自己的能力边界。","revisedPrompt":"Create exactly one complete presentation slide image in 16:9 landscape (1536x864), finished and presentation-ready, in a Manga Style inspired by NotebookLM / Kael style. Dominant art direction: understanding becomes deeper with fun; turn information into a comic-like visual explanation with a simple story; use panels, character reactions, manga framing, subtle speed lines, expressive close-ups, and concise captions; relate the idea to a realistic situation so it feels memorable; fun and explanatory, not childish or chaotic.\n\nDeck title theme: 把学习变成能力系统. Slide 9 topic: 每周 30 分钟，复盘能力边界. Main message: 持续复盘，才能看清自己能独立完成什么、在哪些地方卡住.\n\nScene/layout: a closing manga scene of a learner doing a weekly review at a desk on a calm weekend evening. The learner sits at a desk with a notebook, pen, and a calendar spread. A large calendar shows a repeating weekly rhythm and a small clear marker reading 30 分钟, plus a 4-week progression cue (e.g., four small weekly boxes or four weeks highlighted) indicating the review habit over four weeks. Around the calendar, four sticky notes placed in a balanced editorial composition, each with one exact visible text line. The four sticky notes must read exactly and only these lines, one per note, with no extra visible words anywhere else on the slide: 完成了什么？  哪里卡住？  下周补什么？  用什么任务练？\n\nImportant text constraints: visible text must be exactly these lines, with no extra words. No other readable text anywhere on the slide besides those four lines. Do not add title text, subtitles, labels, captions, watermarks, logos, UI chrome, or extra numbers/words. The calendar can show visual markers for 30 分钟 and 4-week rhythm using icons, bars, dots, brackets, or stylized marks, but no additional readable text. The sticky notes are the only readable text.\n\nVisual style: warm, calm, coherent manga editorial illustration with gentle cream paper background, muted ink lines, soft orange and teal accents, subtle halftone texture, clean composition, clear hierarchy. Use a desk scene with a sense of closure and progress. Show calm confidence, fewer speed lines than energetic slides, and a clear sense of progression over four weeks. Include manga framing and soft expressive character reaction but keep it mature and tidy. Make it visually finished with polished illustration, not a wireframe. Ensure all four note texts are legible at presentation distance.\n\nNo markdown, no fake logos, no watermark, no extra captions, no extra text blocks.","storageFileId":"qv6tiax8m7v5jri3kffciy1s-69c135520f937f8195a020ed3bfcd908"},{"size":"1536x864","model":"gpt-image-2","title":"一句话建立学习系统","layout":"终章海报式漫画。主角站在完成的作品墙前，背后是从目标到作品的路径。","status":"completed","message":"用真实目标发现差距，用具体任务训练能力，用反馈修正方向，用作品证明自己。","imageUrl":"https://cdn.drillso.com/generated/drill/dbor32oeat876lytpnrpzxhn-8a313acde88f3a76e1e195ae46ecc9bc.png","storageKey":"generated/drill/dbor32oeat876lytpnrpzxhn-8a313acde88f3a76e1e195ae46ecc9bc.png","imagePrompt":"Create exactly one complete presentation slide image.\n\nUse the following NotebookLM / Kael style prompt as the dominant art direction. Follow it closely.\n\nSTYLE TEMPLATE START\nUnderstanding becomes deeper with fun.\nTurn the information into a comic-like visual explanation with a simple story.\nUse panels, character reactions, manga framing, speed lines, expressive close-ups, and concise captions.\nRelate the idea to a realistic situation so the audience can remember it.\nThe style should be fun and explanatory, not childish or chaotic.\nSTYLE TEMPLATE END\n\nDeck title: 把学习变成能力系统\nDeck summary: 面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。\nDeck format: presenter_slides\nChosen style: Manga Style\n\nSlide 10: 一句话建立学习系统\nMain message: 用真实目标发现差距，用具体任务训练能力，用反馈修正方向，用作品证明自己。\n\nVisible text must be exactly these lines, with no extra words:\n目标发现差距\n任务训练能力\n反馈修正方向\n作品证明自己\n\nLayout:\n终章海报式漫画。主角站在完成的作品墙前，背后是从目标到作品的路径。\n\nVisual style:\n强收束海报构图，四个短句像宣言贴在作品墙上，主角表情坚定。\n\nAdditional image direction:\nA final manga poster composition. The learner stands confidently in front of a wall of completed works: report, project board, tool, presentation, article. Behind them, a simple path connects goal, task, feedback, and work. Four bold Chinese phrases appear as clean labels: 目标发现差距, 任务训练能力, 反馈修正方向, 作品证明自己. Make it feel like a memorable ending slide.\n\nRequirements:\n- Canvas: 1536x864 landscape presentation slide.\n- Generate one slide only, not a contact sheet and not multiple slide candidates.\n- Make it presentation-ready, high-sensibility, and coherent with the chosen style template.\n- Make it visually finished, not a sparse wireframe or placeholder diagram.\n- Use the strongest signature elements of the style template: palette, typography, composition language, icon/illustration treatment, and editorial rhythm.\n- Keep all visible text readable at presentation distance.\n- Use low text density, strong hierarchy, and source-grounded diagrams, editorial composition, or template-specific visuals.\n- Do not add markdown symbols, fake logos, watermarks, signatures, UI browser chrome, or extra captions.\n- Match the source/user instruction language for all visible slide text and speaker-facing copy.","slideNumber":10,"visibleText":["目标发现差距","任务训练能力","反馈修正方向","作品证明自己"],"visualStyle":"强收束海报构图，四个短句像宣言贴在作品墙上，主角表情坚定。","speakerNotes":"最后用材料中的总结收束：如果想知道当前会什么，就看能独立完成什么；想知道不会什么，就看在哪些任务中卡住；想知道该学什么，就从目标和差距倒推；想知道怎么应用，就把知识放进真实问题里，做出可见成果。学习不是把知识堆进大脑，而是让自己解决问题的能力越来越强。","revisedPrompt":"Create a single 1536x864 landscape presentation slide in a polished manga / NotebookLM-Kael style. Final poster composition, strong closure, editorial manga framing, expressive but not chaotic, high sensibility, presentation-ready.\n\nScene: A confident learner stands in the foreground facing slightly left, in front of a completed works wall. The learner has a determined, calm expression, subtle manga shading, clean linework, dynamic but controlled pose. Behind them is a wall of completed outputs displayed like trophies / evidence: a report page, a project board with sticky-note shapes but NO readable text, a tool or app icon-like object, a presentation deck thumbnail, and an article page. Connect them with a simple visible path / flow from left to right or bottom to top: goal -> task -> feedback -> work, shown with arrows, panels, and motion cues, but without any extra words except the required four lines.\n\nTypography and visible text: ONLY these four Chinese lines, each as a bold clean label or poster caption, readable at presentation distance, exactly and only these lines:\n目标发现差距\n任务训练能力\n反馈修正方向\n作品证明自己\nNo other visible text anywhere. No title, no footer, no watermark, no logos, no UI.\n\nLayout: One strong poster-like slide. The four phrases should appear as four bold declaration labels integrated into the wall or as floating comic placards. Strong hierarchy, low text density, clear visual rhythm. Use manga speed lines, expressive framing, and clean comic panel energy, but keep it elegant and coherent. The completed works wall should occupy the background right/center, the learner in foreground left/center. Use dramatic lighting, crisp outlines, warm-neutral palette with high contrast accent colors. Make it feel like the memorable ending slide of a learning system deck.\n\nImportant: Do not add any text beyond the four required lines. Do not add extra captions. Do not add watermarks, signatures, borders, browser chrome, or fake UI.","storageFileId":"dbor32oeat876lytpnrpzxhn-8a313acde88f3a76e1e195ae46ecc9bc"}],"styleId":"manga_story","summary":"面向想提升职场或个人能力的学习者。叙事从“能力模糊”的困惑开始，依次建立判断会不会、盘点能力、识别差距、倒推学习、闭环练习、真实应用和每周复盘的完整方法。","audience":"职场学习者、自我提升者、培训或工作坊参与者","imageSize":"1536x864","imageModel":"gpt-image-2","plannerModel":"gpt-5.5"},"children":[]},"breadcrumbs":[{"id":"b9fc08c6-83c3-4d17-b5ef-841327f6deec","slug":"怎么知道自己当前会什么？不会什么？该学什么？怎么应用？-b9fc08c6","title":"怎么知道自己当前会什么？不会什么？该学什么？怎么应用？","type":"page","url":"https://drillso.com/zh/share/sessions/tbhNiIW8RWXL/%E6%80%8E%E4%B9%88%E7%9F%A5%E9%81%93%E8%87%AA%E5%B7%B1%E5%BD%93%E5%89%8D%E4%BC%9A%E4%BB%80%E4%B9%88%EF%BC%9F%E4%B8%8D%E4%BC%9A%E4%BB%80%E4%B9%88%EF%BC%9F%E8%AF%A5%E5%AD%A6%E4%BB%80%E4%B9%88%EF%BC%9F%E6%80%8E%E4%B9%88%E5%BA%94%E7%94%A8%EF%BC%9F-b9fc08c6","agentUrl":"https://drillso.com/zh/share/sessions/tbhNiIW8RWXL/agent.json?node=%E6%80%8E%E4%B9%88%E7%9F%A5%E9%81%93%E8%87%AA%E5%B7%B1%E5%BD%93%E5%89%8D%E4%BC%9A%E4%BB%80%E4%B9%88%EF%BC%9F%E4%B8%8D%E4%BC%9A%E4%BB%80%E4%B9%88%EF%BC%9F%E8%AF%A5%E5%AD%A6%E4%BB%80%E4%B9%88%EF%BC%9F%E6%80%8E%E4%B9%88%E5%BA%94%E7%94%A8%EF%BC%9F-b9fc08c6"}],"parent":{"id":"b9fc08c6-83c3-4d17-b5ef-841327f6deec","slug":"怎么知道自己当前会什么？不会什么？该学什么？怎么应用？-b9fc08c6","title":"怎么知道自己当前会什么？不会什么？该学什么？怎么应用？","type":"page","url":"https://drillso.com/zh/share/sessions/tbhNiIW8RWXL/%E6%80%8E%E4%B9%88%E7%9F%A5%E9%81%93%E8%87%AA%E5%B7%B1%E5%BD%93%E5%89%8D%E4%BC%9A%E4%BB%80%E4%B9%88%EF%BC%9F%E4%B8%8D%E4%BC%9A%E4%BB%80%E4%B9%88%EF%BC%9F%E8%AF%A5%E5%AD%A6%E4%BB%80%E4%B9%88%EF%BC%9F%E6%80%8E%E4%B9%88%E5%BA%94%E7%94%A8%EF%BC%9F-b9fc08c6","agentUrl":"https://drillso.com/zh/share/sessions/tbhNiIW8RWXL/agent.json?node=%E6%80%8E%E4%B9%88%E7%9F%A5%E9%81%93%E8%87%AA%E5%B7%B1%E5%BD%93%E5%89%8D%E4%BC%9A%E4%BB%80%E4%B9%88%EF%BC%9F%E4%B8%8D%E4%BC%9A%E4%BB%80%E4%B9%88%EF%BC%9F%E8%AF%A5%E5%AD%A6%E4%BB%80%E4%B9%88%EF%BC%9F%E6%80%8E%E4%B9%88%E5%BA%94%E7%94%A8%EF%BC%9F-b9fc08c6"},"children":[],"fullTree":null,"warnings":[],"truncated":false}