image-skill-builder

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

批量生图 Skill 创建助手

Batch Image Generation Skill Creation Assistant

通过结构化问答，帮助零代码基础用户创建专属的批量生图 Skill。

Help users with zero coding experience create exclusive batch image generation Skills through structured Q&A.

核心理念

Core Concept

用户只需要回答问题，助手自动完成：

理解用户需求 → 2. 组装风格模板 → 3. 生成完整 Skill 文件 → 4. 验证可用性

Users only need to answer questions, and the assistant will automatically complete:

Understand user needs → 2. Assemble style templates → 3. Generate complete Skill files → 4. Verify usability

工作流程

Workflow

Phase 1: 场景理解（3 个问题）

Phase 1: Scenario Understanding (3 Questions)

逐个提问，不要一次性抛出所有问题。每次提问后等待回答。

Q1 - 配图场景:

你想创建什么类型的批量生图 Skill？

1. 📝 文章/博客配图 — 为长文的每个段落自动配图
2. 🎨 Logo/品牌设计 — 批量生成 Logo 变体
3. 🎬 分镜/故事板 — 将故事拆解为连续画面
4. 📱 社交媒体图 — 小红书/公众号/X 配图
5. 🖼️ 海报/封面设计 — 活动海报、课程封面、Banner
6. 📊 数据可视化图 — 信息图、流程图、架构图
7. 🎭 人物/角色设计 — IP 形象、头像、人设图
8. ✏️ 其他（请描述）

选一个数字，或直接描述你的场景。

Q2 - 使用场景:

这些图片主要用在哪里？

1. 📖 文章/文档中插图（需要自动插入 Markdown）
2. 📲 社交平台发布（需要特定尺寸）
3. 💼 品牌/商业用途（需要高分辨率+多变体）
4. 🎓 教学/演示（需要清晰易懂）
5. 🎮 娱乐/创作（追求视觉效果）

这会影响图片的分辨率、尺寸和输出格式。

Q3 - 批量数量:

每次使用这个 Skill 时，大约生成多少张图？

1. 1-3 张（精品路线，每张仔细打磨）
2. 5-10 张（标准批量，平衡质量和数量）
3. 10+ 张（大批量，快速产出）

这会影响 Skill 的工作流设计（是否需要分批确认）。

Ask questions one by one, do not throw all questions at once. Wait for the answer after each question.

Q1 - Image Scenario:

What type of batch image generation Skill do you want to create?

1. 📝 Article/Blog Illustrations — Automatically generate images for each paragraph of long-form text
2. 🎨 Logo/Brand Design — Batch generate Logo variants
3. 🎬 Storyboard/Story Sequence — Break down stories into consecutive frames
4. 📱 Social Media Images — Images for Xiaohongshu, Official Accounts, X
5. 🖼️ Poster/Cover Design — Event posters, course covers, Banners
6. 📊 Data Visualization — Infographics, flowcharts, architecture diagrams
7. 🎭 Character/IP Design — IP images, avatars, character settings
8. ✏️ Other (Please describe)

Select a number, or directly describe your scenario.

Q2 - Usage Scenario:

Where will these images be mainly used?

1. 📖 Illustrations in articles/documents (Need automatic insertion into Markdown)
2. 📲 Social platform publishing (Need specific dimensions)
3. 💼 Brand/commercial use (Need high resolution + multiple variants)
4. 🎓 Teaching/demonstration (Need clear and easy to understand)
5. 🎮 Entertainment/creation (Pursue visual effects)

This will affect the resolution, dimensions and output format of the images.

Q3 - Batch Quantity:

Approximately how many images will be generated each time this Skill is used?

1. 1-3 images (Premium route, polish each image carefully)
2. 5-10 images (Standard batch, balance quality and quantity)
3. 10+ images (Large batch, fast output)

This will affect the workflow design of the Skill (whether batch confirmation is required).

Phase 2: 风格定义（3-4 个问题）

Phase 2: Style Definition (3-4 Questions)

Q4 - 视觉风格:

根据 Q1 的回答，从 references/style-library.md 中筛选出 5-8 个最相关的风格选项展示。

展示格式（每个风格一行描述）：

根据你的场景，推荐以下风格：

1. [风格名] — [一句话特征] | 适合: [场景]
2. [风格名] — [一句话特征] | 适合: [场景]
...

你喜欢哪个？也可以描述你想要的风格，我来匹配。

如果用户选了某个风格 → 记录并继续 如果用户描述了自定义风格 → 提炼关键词，构建自定义风格前缀

Q5 - 参考图片（可选）:

有参考图片吗？上传一张你喜欢的风格图，我来拆解：

- 图形元素（形状/构图/布局）
- 色彩方案（主色/辅色/氛围）
- 视觉风格（写实/扁平/手绘/3D...）
- 情感调性（专业/温暖/酷炫/可爱...）
- 文字处理（有无文字/字体风格）

没有也可以，直接说"跳过"。

如果用户上传了图片 → 分析图片并生成风格描述，与 Q4 的选择融合 如果用户跳过 → 继续

Q6 - 色彩偏好:

色彩偏好？

1. ⬛ 黑白/灰度 — 极简、专业
2. 🔵 单色调 — 一个主色+黑白（如蓝色系、绿色系）
3. 🌈 多彩 — 丰富配色，视觉活泼
4. 🎯 品牌色 — 指定具体颜色（给我色号或描述）
5. 🤷 随意 — 根据风格自动搭配

选一个，或直接告诉我你想要的颜色。

Q7 - 图片比例:

图片比例？

1. 📺 16:9 横版 — 文章配图、演示文稿、Banner
2. ⬜ 1:1 正方形 — Logo、社交头像、Instagram
3. 📱 9:16 竖版 — 小红书、手机壁纸、Story
4. 📄 3:4 竖版 — 海报、封面
5. 🖥️ 21:9 超宽 — Banner、网站头图
6. 其他（告诉我具体比例）

最常用的是 16:9 和 1:1。

Q4 - Visual Style:

Based on the answer to Q1, filter and display 5-8 most relevant style options from references/style-library.md.

Display format (one style description per line):

Based on your scenario, the following styles are recommended:

1. [Style Name] — [One-sentence feature] | Suitable for: [Scenario]
2. [Style Name] — [One-sentence feature] | Suitable for: [Scenario]
...

Which one do you like? You can also describe the style you want, and I will match it for you.

If the user selects a style → Record and continue If the user describes a custom style → Extract keywords and build a custom style prefix

Q5 - Reference Images (Optional):

Do you have reference images? Upload a style image you like, and I will analyze it:

- Graphic elements (Shape/composition/layout)
- Color scheme (Main color/secondary color/atmosphere)
- Visual style (Realistic/flat/hand-drawn/3D...)
- Emotional tone (Professional/warm/cool/cute...)
- Text processing (With/without text/font style)

You can skip this if you don't have any, just say "Skip".

If the user uploads an image → Analyze the image and generate a style description, integrate with the selection in Q4 If the user skips → Continue

Q6 - Color Preference:

Color preference?

1. ⬛ Black and white/grayscale — Minimalist, professional
2. 🔵 Monochrome — One main color + black and white (e.g., blue series, green series)
3. 🌈 Colorful — Rich color scheme, visually lively
4. 🎯 Brand color — Specify a specific color (give me the color code or description)
5. 🤷 Random — Automatically matched according to style

Select one, or directly tell me the color you want.

Q7 - Image Aspect Ratio:

Image aspect ratio?

1. 📺 16:9 Landscape — Article illustrations, presentations, Banners
2. ⬜ 1:1 Square — Logos, social avatars, Instagram
3. 📱 9:16 Portrait — Xiaohongshu, mobile wallpapers, Stories
4. 📄 3:4 Portrait — Posters, covers
5. 🖥️ 21:9 Ultra-wide — Banners, website headers
6. Other (Tell me the specific ratio)

The most commonly used are 16:9 and 1:1.

Phase 3: 内容规则（2-3 个问题）

Phase 3: Content Rules (2-3 Questions)

Q8 - 文字规则:

图片中需要包含文字吗？

1. ❌ 无文字 — 纯视觉图
2. 📌 少量关键词 — 1-3 个核心词汇
3. 📝 标题+关键词 — 有明显的文字信息
4. 🌐 中英文混合 — 中文为主，专业术语英文

如果需要文字，默认使用中文。

Q9 - 内容来源:

图片的内容从哪里来？

1. ✍️ 每次手动描述 — 我告诉 AI 画什么
2. 📄 从文章/文档自动提取 — AI 分析文章生成配图
3. 📋 从模板/列表批量生成 — 我提供一个清单批量出图
4. 🖼️ 从参考图延伸 — 上传图片+描述来改编

这决定了 Skill 的输入方式。

Q10 - 一致性要求:

多张图片之间需要保持风格一致吗？

1. 🔒 必须一致 — 同一套视觉语言（如系列配图、故事板）
2. 🔓 允许变化 — 可以有风格差异（如 Logo 变体探索）
3. 🎯 部分一致 — 整体风格一致但细节可变

一致性越强，提示词中的风格锚定越严格。

Q8 - Text Rules:

Does the image need to include text?

1. ❌ No text — Pure visual image
2. 📌 A few keywords — 1-3 core vocabulary
3. 📝 Title + keywords — Obvious text information
4. 🌐 Mixed Chinese and English — Mainly Chinese, professional terms in English

If text is needed, Chinese is used by default.

Q9 - Content Source:

Where does the image content come from?

1. ✍️ Manual description each time — I tell the AI what to draw
2. 📄 Automatically extract from articles/documents — AI analyzes the article to generate illustrations
3. 📋 Batch generate from templates/lists — I provide a list for batch image generation
4. 🖼️ Extend from reference images — Upload images + descriptions to adapt

This determines the input method of the Skill.

Q10 - Consistency Requirement:

Do multiple images need to maintain a consistent style?

1. 🔒 Must be consistent — Same visual language (e.g., series illustrations, storyboards)
2. 🔓 Allow changes — Style differences are allowed (e.g., Logo variant exploration)
3. 🎯 Partially consistent — Overall style is consistent but details can vary

The stronger the consistency, the stricter the style anchoring in the prompt.

Phase 4: 技术配置（快速确认）

Phase 4: Technical Configuration (Quick Confirmation)

Q11 - 技术参数确认（一次性展示，用户修改或确认即可）:

最后确认技术参数：

| 配置项 | 当前值 | 说明 |
|-------|--------|------|
| 分辨率 | 2K | 可选 1K/2K/4K |
| API | Gemini (gemini-3-pro-image-preview) | 默认 |
| 保存位置 | obsidian/09image/MMDD-主题名/ | 自动按日期分类 |
| 插入文档 | [是/否，根据Q2判断] | 自动插入 Markdown |
| Skill 名称 | [根据Q1自动生成] | 你也可以起个名字 |

有需要修改的吗？没有就直接回复"确认"。

Q11 - Technical Parameter Confirmation (Display all at once, users can modify or confirm directly):

Final confirmation of technical parameters:

| Configuration Item | Current Value | Description |
|-------|--------|------|
| Resolution | 2K | Optional 1K/2K/4K |
| API | Gemini (gemini-3-pro-image-preview) | Default |
| Save Location | obsidian/09image/MMDD-Theme Name/ | Automatically categorized by date |
| Insert into Document | [Yes/No, determined based on Q2] | Automatically insert into Markdown |
| Skill Name | [Automatically generated based on Q1] | You can also name it yourself |

Do you need to modify anything? If not, just reply "Confirm".

Phase 5: 自动生成 Skill

Phase 5: Automatic Skill Generation

收集完所有答案后，自动执行以下步骤：

After collecting all answers, automatically execute the following steps:

5.1 组装 Skill 结构

5.1 Assemble Skill Structure

根据用户回答，决定 Skill 的结构：

用户选择	生成的 Skill 结构
文章配图	类似 article-batch-illustration: 分析文章 → 拆段落 → 生成提示词 → 确认 → 批量出图 → 插入文档
Logo 设计	类似 logo-batch-generator: 收集品牌信息 → 选风格 → 10个变体 → 出图 → 迭代
分镜故事	类似 storyboard-generator: 润色故事 → 拆分镜 → 选风格 → 批量出图
社交媒体图	新模式: 输入主题/文案 → 选模板 → 批量出图 → 导出特定尺寸
海报/封面	新模式: 输入主题 → 选风格 → 生成多个变体 → 迭代
数据可视化	类似 article-batch-illustration 但专注图表: 输入数据/概念 → 选图表类型 → 出图
人物/角色	新模式: 描述人设 → 选风格 → 多角度/表情变体 → 出图

Determine the Skill structure based on user answers:

User Selection	Generated Skill Structure
Article illustrations	Similar to article-batch-illustration: Analyze article → Split paragraphs → Generate prompts → Confirm → Batch generate images → Insert into document
Logo design	Similar to logo-batch-generator: Collect brand information → Select style → 10 variants → Generate images → Iterate
Storyboard	Similar to storyboard-generator: Polish story → Split into storyboards → Select style → Batch generate images
Social media images	New mode: Input theme/copy → Select template → Batch generate images → Export specific dimensions
Poster/cover	New mode: Input theme → Select style → Generate multiple variants → Iterate
Data visualization	Similar to article-batch-illustration but focused on charts: Input data/concepts → Select chart type → Generate images
Character/IP design	New mode: Describe character settings → Select style → Multi-angle/expression variants → Generate images

5.2 生成文件

5.2 Generate Files

为用户创建完整的 Skill 目录：

skills/[skill-name]/
├── SKILL.md              ← 根据回答自动填充
├── agents/
│   └── openai.yaml       ← 自动生成 UI 元数据
├── scripts/
│   └── generate_image.py ← 复制自 references/generate_image.py
└── references/
    └── style-guide.md    ← 根据风格选择自动组装

Create a complete Skill directory for the user:

skills/[skill-name]/
├── SKILL.md              ← Automatically filled based on answers
├── agents/
│   └── openai.yaml       ← Automatically generate UI metadata
├── scripts/
│   └── generate_image.py ← Copied from references/generate_image.py
└── references/
    └── style-guide.md    ← Automatically assembled based on style selection

5.3 生成 SKILL.md

5.3 Generate SKILL.md

SKILL.md 模板结构（读取 references/skill-template.md 获取完整模板）：

markdown

---
name: [skill-name]
description: [根据场景+风格自动生成描述，包含触发词]
---

SKILL.md template structure (Read references/skill-template.md for the complete template):

markdown

---
name: [skill-name]
description: [Automatically generated based on scenario + style, including trigger words]
---

[Skill 显示名]

[Skill Display Name]

[一句话描述功能]

[One-sentence function description]

工作流程

Workflow

[根据场景类型填充对应的工作流]

[Fill in the corresponding workflow based on scenario type]

风格配置

Style Configuration

[用户选定的风格前缀和参数]

[User-selected style prefix and parameters]

API 配置

API Configuration

[固定 API 信息]

[Fixed API information]

保存路径规范

Save Path Specification

[路径模板]

[Path template]

常见用法

Common Usage

[触发词和使用示例]

undefined

[Trigger words and usage examples]

undefined

5.4 生成 style-guide.md

5.4 Generate style-guide.md

根据用户选择的风格 + 色彩 + 比例 + 文字规则，从 references/style-library.md 中提取对应风格的完整提示词模板，组装成该 Skill 专属的 style-guide.md。

Based on the user's selected style + color + aspect ratio + text rules, extract the corresponding complete prompt template from references/style-library.md and assemble it into a style-guide.md exclusive to this Skill.

5.5 复制脚本

5.5 Copy Script

将

scripts/generate_image.py

复制到新 Skill 的 scripts/ 目录。

Copy

scripts/generate_image.py

to the scripts/ directory of the new Skill.

5.6 验证

5.6 Verification

运行 quick_validate.py 确认 Skill 结构正确。

Run quick_validate.py to confirm the Skill structure is correct.

Phase 6: 展示与试用

Phase 6: Display and Trial

生成完成后，展示给用户：

你的 Skill 已创建完成！

📁 位置: skills/[skill-name]/
📄 文件:
  - SKILL.md (主文件)
  - scripts/generate_image.py (图片生成脚本)
  - references/style-guide.md (风格指南)
  - agents/openai.yaml (UI 配置)

🔧 核心配置:
  - 场景: [xxx]
  - 风格: [xxx]
  - 比例: [xxx]
  - 分辨率: [xxx]

想现在试用一下这个 Skill 吗？给我发一段内容，我用刚创建的 Skill 帮你生成图片。

After generation is complete, show to the user:

Your Skill has been created!

📁 Location: skills/[skill-name]/
📄 Files:
  - SKILL.md (Main file)
  - scripts/generate_image.py (Image generation script)
  - references/style-guide.md (Style guide)
  - agents/openai.yaml (UI configuration)

🔧 Core Configuration:
  - Scenario: [xxx]
  - Style: [xxx]
  - Aspect Ratio: [xxx]
  - Resolution: [xxx]

Do you want to try this Skill now? Send me a piece of content, and I will generate images for you using the newly created Skill.

关键设计原则

Key Design Principles

每次只问一个问题 — 不要同时抛多个问题，降低用户认知负担
提供明确选项+自定义入口 — 数字选择快速，同时允许自由描述
智能默认值 — 根据前面的回答推断后续默认值（如选了文章配图 → 默认16:9、自动插入文档）
即时反馈 — 每个回答后简短确认理解，让用户有掌控感
允许跳过和回退 — 用户可以说"跳过"或"回去改第X个"

Ask only one question at a time — Do not throw multiple questions at once to reduce user cognitive load
Provide clear options + custom entry — Number selection is fast, and free description is allowed at the same time
Smart default values — Infer subsequent default values based on previous answers (e.g., if article illustrations are selected → default 16:9, automatic document insertion)
Instant feedback — Briefly confirm understanding after each answer to give users a sense of control
Allow skipping and backtracking — Users can say "Skip" or "Go back to modify the Xth question"

参考文件

Reference Files

文件	用途	何时读取
references/style-library.md	所有可用风格的完整提示词模板库	Phase 2 展示风格选项时
references/skill-template.md	Skill 文件生成模板	Phase 5 组装 Skill 时
scripts/generate_image.py	图片生成脚本（复制到新 Skill）	Phase 5 复制脚本时

File	Purpose	When to Read
references/style-library.md	Complete prompt template library for all available styles	When displaying style options in Phase 2
references/skill-template.md	Skill file generation template	When assembling the Skill in Phase 5
scripts/generate_image.py	Image generation script (copied to new Skill)	When copying the script in Phase 5

常见用法

Common Usage

用户说	操作
"帮我创建一个生图 Skill"	从 Q1 开始完整问答流程
"我想做一个小红书配图的 Skill"	跳过 Q1（已知场景=社交媒体），从 Q2 开始
"创建一个和文章配图类似的 Skill，但用赛博朋克风"	跳过 Q1+Q4（已知场景+风格），快速确认其他参数
"帮我改一下刚创建的 Skill 的风格"	读取已有 Skill，只重新走 Phase 2

User Says	Operation
"Help me create an image generation Skill"	Start the complete Q&A process from Q1
"I want to make a Xiaohongshu image matching Skill"	Skip Q1 (Known scenario = social media) and start from Q2
"Create a Skill similar to article illustration, but in cyberpunk style"	Skip Q1+Q4 (Known scenario + style) and quickly confirm other parameters
"Help me modify the style of the Skill I just created"	Read the existing Skill and only re-run Phase 2