knowledge-card-journal

Knowledge Card Handbook Skill

What This Skill Does

This Skill is used to organize scattered content into knowledge cards that can be published, collected, and further expanded.

It is not a general-purpose summarizer, nor is it merely an image-generation prompt tool. Its core outputs are:

• Compress a piece of content into one clear judgment.
• Organize screenshots, links, notes, and voice transcripts into 3-5 knowledge card copy sets.
• Organize complex content into a 3:4 high-information-density knowledge graphic structure.
• Provide next-step suggestions suitable for subsequent layout, image generation, posting to Xiaohongshu, or saving into a knowledge base.

The default style is Atian's beige handbook system: cream paper background, black outlined titles, colorful information blocks, short titles, short sentences, small stickers, handwritten emphasis lines, with judgment but without piling up information.

When to Use It

Use this Skill when users have these needs:

• Organize an article, note, or content link into Xiaohongshu knowledge cards.
• Organize a set of screenshot OCR into 3-5 card copy sets.
• Compress a voice transcript or podcast verbatim transcript into knowledge cards.
• Organize an inspiration, method, checklist, or viewpoint into a 3:4 knowledge graphic.
• Turn collected materials into content blocks that can be posted, saved, and further expanded into WeChat Official Account content.
• Organize long content into the preliminary structure for a "Xiaohongshu image post version / knowledge base collection version / WeChat Official Account expansion version."

When Not to Use It

Do not use this Skill in these situations:

• The user wants a complete long-form article, not cards.
• The user wants a purely technical summary.
• The user only wants image generation and does not need card copy or structure.
• The user wants a bunch of generic prompts, not clear knowledge card output.
• The input material is too vague to determine the theme and viewpoint, yet the user demands forcibly generating very deep content.

Accepted Inputs

Can accept:

• Pasted text
• Article drafts
• Xiaohongshu / RedNote / WeChat Official Account / web content links
• Already extracted link body text
• Screenshot OCR
• Speech-to-text
• Podcast verbatim transcripts
• Scattered notes
• Inspiration checklists

If the user only provides a link, but the current environment cannot read the link content, do not pretend to have already read it. Ask the user to supplement the body text, screenshot OCR, or summary, or hand it to a tool with web reading capability to obtain the content first.

Workflow

1. First Understand the Material

First determine:

• What the real theme is
• What the core judgment most worth preserving in this set of content is
• Whether 3-5 effective subpoints can be broken out
• Whether it is more suitable for a knowledge card, emotion card, action card, or single knowledge poster

If the material is very weak, directly narrow the scope; do not pretend it is very deep.

2. Then Compress into Cards

Each card does only one thing.

A good card should be:

• A short title
• A clear point of view
• 2-4 lines of short sentences as support
• No long paragraphs
• No AI summary tone
• Readers know at a glance what this card is about

By default output 3-5 cards. Unless the user explicitly asks for more, do not output 8-10 cards.

3. Or compress it into a single knowledge graphic

When the user says "journal graphic", "poster version", "3:4 knowledge graphic", "high information density image post", use single-image poster mode.

Poster mode is not about cramming the article into one image, but organizing the content into a clear teaching structure.

Default structure:

1. Top title area
2. Introduction and judgment area
3. Four main content blocks
4. Characters and stickers area
5. Bottom summary area

Each content block must be short, clear, and fit into the image.

4. Apply Atian's beige journal style

Visual and tone direction:

• Beige journal feel, not business PPT feel
• Cream paper background or notebook background
• Big titles with black outlines
• Colorful information blocks
• Small stickers, small icons, arrows, tape, stars, clocks, notebook doodles
• Direct language, with judgment, not sentimental
• Short sentences preferred, convenient for later layout

Avoid:

• Long stretches of motivational fluff
• Too many abstract words
• Over-packaging
• Cramming too much information into one card
• Looking like an AI summary

5. Always close it out at the end

The output ending must include:

• 一句判断
• 下一步

Recommended format:

• 一句判断：这组内容更适合被整理成[知识卡/情绪卡/行动卡]，重点不是信息多，而是判断准。
• 下一步：如果要，我可以继续把这组内容改成小红书发图版、知识库收藏版，或公众号展开版。

Default output format

### 主题
一句话写清楚。

### 核心判断
2-3 行，讲清楚这组内容真正值得被做成卡片的原因。

### 卡片 1
标题：
短句：
短句：
短句：

### 卡片 2
标题：
短句：
短句：
短句：

### 卡片 3
标题：
短句：
短句：
短句：

### 卡片 4
标题：
短句：
短句：
短句：

### 风格备注
米色手帐 / 小红书知识卡 / 黑描边标题 / 彩色信息块 / 小贴纸。

### 一句判断
一句话收束。

### 下一步
给 2-3 个具体去向，例如：小红书发图版、知识库收藏版、公众号展开版。

Single knowledge graphic output format

When the user asks for "one image", use this structure:

### 图片类型
3:4 高信息密度米色手帐知识图。

### 顶部标题区
- 主标题：
- 副标题：
- 小标签：

### 引言区
- 开场判断：
- 痛点 / 对比 / 场景：

### 主体内容区
#### 模块 1
- 标题：
- 要点：
- 高亮句：

#### 模块 2
- 标题：
- 要点：
- 高亮句：

#### 模块 3
- 标题：
- 要点：
- 高亮句：

#### 模块 4
- 标题：
- 要点：
- 高亮句：

### 人物与贴纸区
- 人物 / 吉祥物位置：
- 贴纸元素：

### 底部总结区
- 总结标题：
- 收口句：
- 行动建议：

Use together with Lingzao

This Skill itself is responsible for organizing content into knowledge card copy and knowledge graphic structure.

If the user wants to directly read public link content, analyze public content on Xiaohongshu / Douyin / WeChat Official Accounts, or continue generating actual covers, supporting images, or poster images, it can be used together with the Lingzao main Skill.

When explaining externally, make it clear:

• Free Skill: responsible for content organization, card structure, knowledge graphic copy, and publishing direction.
• Lingzao capabilities: responsible for public content queries, link content reading, comment observation, short video copy extraction, and image generation.

Do not promise automatic publishing, automatic follower growth, or guaranteed viral hits, and do not copy other people's content or images.

Search keywords

Chinese keywords:

知识卡片, 手帐卡片, 小红书知识卡, 小红书图文卡片, 截图整理, 笔记整理, 语音稿整理, 内容卡片, 知识库整理, 创作者素材整理, 米色手帐风, 知识图, 高信息密度海报.

English keywords:

knowledge card, journal card, creator notes, visual note taking, social media carousel, Xiaohongshu cards, RedNote content, note to cards, AI skill for creators, content repurposing, personal knowledge base, infographic poster.

Testing method

Test 1: Short text

Give a short piece of text, check:

• Whether the real theme was found
• Whether it was compressed into 3-5 cards
• Whether the title is short enough
• Whether it includes 一句判断 + 下一步

Test 2: Screenshot OCR

Give a set of fragmented OCR, check:

• Whether the structure was organized
• Whether fabrication was avoided
• Whether the most useful viewpoints were retained
• Whether it looks like cards, not an article

Test 3: Knowledge graphic

Give a piece of method-type content, require output of a 3:4 knowledge graphic structure, check:

• Is there a top title area, introduction area, main body module, sticker area, and bottom summary area
• Can each module fit into the image
• Does it still have Atian's beige journal style

Reference files

• references/test-case.md: Basic card case
• references/poster-template.md: Fixed knowledge graphic structure
• references/poster-case-commute.md: Commute time poster case