Token导航 LogoToken导航

Agent Skill · AI 工具

markdown-slides(Markdown slides 控制)

用于辅助文档、README、Markdown、说明文和内容稿件的整理与改写。它适合让 Agent 提炼结构、补齐章节、统一术语、检查链接或把零散材料整理成可读文档。使用时应保留项目已有事实、命令和路径,不要把未确认的信息写成确定结论;涉及对外文案时,还需要控制语气,避免过度营销或夸大能力。

markdown-slides 是AI 工具类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为只读;许可证为 MIT,第三方扫描已通过,按安装量在AI 工具类的 2,536 个 Skill 里排第 1,236。

总安装 948周安装 38GitHub Stars 43下载 307核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
MIT

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:markdown-slides(Markdown slides 控制)
来源仓库:https://github.com/jykim/claude-obsidian-skills
仓库路径:skills/markdown-slides
安装命令:
npx skills add https://github.com/jykim/claude-obsidian-skills --skill markdown-slides
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。不同来源提供的安装方式可能略有差异;本站展示可直接复制的安装命令,安装前请核对来源页面。

skills.shnpx skills
npx skills add https://github.com/jykim/claude-obsidian-skills --skill markdown-slides

这个 Skill 是做什么的

AI 工具类 Skill 用来扩展助手本身的能力,例如接模型、接检索、接外部服务。

将 Markdown 转换为演示文稿格式,自动生成幻灯片框架。适合快速产出会议简报或产品发布会的视觉材料。输出结果依赖原始文本质量,复杂图表建议手动调整布局。涉及版权素材引用时,必须标明来源并保持授权合规。markdown-slides 属于AI 工具类 Skill,可作为该场景下的辅助能力补充。

来源仓库内路径:skills/markdown-slides

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/jykim/claude-obsidian-skills --skill markdown-…;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它标注为只读,先用只读任务试一次,确认行为符合预期再扩大使用范围。

常见问题

6 条

markdown-slides 怎么安装?

在终端执行 npx skills add https://github.com/jykim/claude-obsidian-skills --skill markdown-slides。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

markdown-slides 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

markdown-slides 需要什么权限?

权限级别是只读。该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

markdown-slides 是免费的吗?

来源标注的许可证是 MIT。本站不代收任何费用,也不托管安装包;能否商用、要不要署名,以许可证原文为准。

markdown-slides 用的人多吗?

公开数据是累计安装 948 次,近一周安装 38 次,下载 307 次,GitHub 43 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

markdown-slides 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 3 个目录同时收录,说明这个 Skill 流传比较广。
  • 许可证明确为 MIT,商用和分发边界好判断。
  • 第三方扫描已通过,常见的高风险行为没有命中。

先想清楚的地方

  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理AI 工具类任务、又不想每次都手写提示词的人。
  • 需要许可证清楚、并且有第三方扫描结果的团队项目。

不太适合

  • 需要厂商级技术支持和 SLA 的生产系统。

AI 工具类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
executing-plans

执行计划 · Codex、Claude

1,515,336173,110只读MIT通过
verification-before-completion

完成前的验证 · Codex、Claude

1,319,472173,121权限未标注MIT通过
subagent-driven-development

子 Agent 驱动开发 · Codex、Claude

1,312,416173,138只读MIT通过
copy-editing

复制编辑 · Codex、Claude

1,209,57625,632只读MIT通过
paywall-upgrade-cro

付费专区升级 cro · Codex、Claude

1,030,17625,617只读MIT通过
markdown-slides

Markdown slides 控制 · 本页

94843只读MIT通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Markdown Slides Skill

Convert content to presentation slides in Markdown format compatible with Deckset and Marp presentation tools.

When to Use This Skill

Activate this skill when the user:

  • Asks to create slides or a presentation
  • Requests to convert a document to slide format
  • Mentions Deckset or Marp
  • Wants to generate speaker notes
  • Needs to format images for presentations

Input Requirements

  • Source content: Markdown file, outline, or structured content
  • Images: Files in _files_/ directory or paths to be resolved
  • Target platform: Deckset (default) or Marp
  • Options:

- Speaker notes (enabled by default) - Slide numbering (optional frontmatter) - Emoji enhancement (enabled by default)

Output Specifications

  • File location: Same folder as source
  • File naming: Original name with - slides suffix

- Example: document.md → document - slides.md

  • Format: Deckset-compatible markdown
  • Structure: Proper slide dividers, image tags, speaker notes
  • Enhancement: Emojis added for visual appeal

Main Process

Step 1: Slide Structure Setup

Objective: Create logical slide divisions with proper hierarchy

Actions:

  1. Read and analyze source content structure
  2. Insert slide dividers (---) at logical breaks:

- Between major sections - Between distinct topics - After section intro slides - Before/after major diagrams

  1. Maintain content hierarchy:

- H1 (#) for section titles - H2 (##) for main slide titles - H3 (###) for sub-topics

  1. Preserve original organization and flow
  2. Add optional frontmatter: slidenumbers: true

Example:

slidenumbers: true
# 1. Section Title 🎯

![](section-background.png)

^ Introduction to this section.

---

## Main Topic 📊

Content here...

---

### Detailed Subtopic

More details...

Step 2: Image Format Conversion

Objective: Convert images to Deckset format with proper positioning

Critical Requirements: ⚠️ MUST resolve relative paths per-image basis ⚠️ MUST URL-encode spaces and escape special characters ⚠️ MUST verify image files exist

Image Position Formats:

FormatUsageExample
![]()Section intro backgrounds![](background.png)
![inline]()Diagrams within text flow![inline](diagram.svg)
![right fit]()PRIMARY: Content slides![right fit](chart.png)
![right 80%]()Specific sizing needed![right 80%](image.png)
![inline fill]()Full-width inline![inline fill](wide.png)

Actions:

  1. Identify all image references in source
  2. For each image: a. Determine appropriate position format b. Resolve relative path from slide file location c. URL-encode spaces: " " → "%20" d. Escape parentheses: "(" → "%28", ")" → "%29" e. Verify file exists at resolved path f. Use original images when available g. Copy missing images to _files_/ directory
  3. Apply positioning based on context:

- Section intros: Use ![]() for full background - Diagrams in text: Use ![inline]() - Default for content slides: Use ![right fit]() ⭐ - Custom sizing: Use ![right XX%]() when needed - Full-width inline: Use ![inline fill]() for wide images

Examples:

# Section Title
![](background-image.png)

---

## Content Slide
Key points here:
- Point 1
- Point 2

![right fit](diagram.png)

---

## Inline Diagram
Text before diagram.

![inline](process-flow.svg)

Text after diagram.

Step 3: Speaker Notes Conversion

Objective: Convert appropriate content to speaker notes

Format: Lines starting with ^ (caret + space)

Placement: At the end of each slide

Convert to speaker notes: ✅ Multi-sentence paragraphs ✅ Explanatory text ✅ Additional context ✅ Talking points

DO NOT convert: ❌ Single-sentence paragraphs ❌ Lists (bullet or numbered) ❌ Block quotes ❌ Sentences ending with colon (:) ❌ Headers (H1, H2, H3) ❌ Table content

For Marp platform: Use HTML comments instead

<!-- This is a speaker note for Marp -->

Examples:

## Slide Title

Main content visible to audience.

^ This paragraph becomes a speaker note because it's multi-sentence explanatory text. It provides context that the speaker needs but the audience doesn't need to read on the slide.

---

## Another Slide

- Bullet points stay visible
- Not converted to notes

^ Additional speaking points go here.

Step 4: Content Enhancement

Objective: Polish the presentation for maximum impact

Actions:

  1. Add emojis to section and slide titles:

- Use relevant emojis that enhance meaning - Don't overuse - keep it professional - Examples: 🎯📊🤖💡🚀📚🔧⚡

  1. Clean up comments:

- Remove markdown comments not meant for slides - Remove TODO items - Remove internal notes

  1. Ensure clean format:

- Consistent spacing - Proper header hierarchy - Clear slide breaks

  1. Add frontmatter (if requested): slidenumbers: true
  2. Final review:

- Each slide has clear focus - Images positioned correctly - Speaker notes placed at end - No orphaned content

Critical Guidelines

Image Handling

⚠️ CRITICAL RULES:

  1. NEVER invent images - Only reference images that exist in the source document or _files_/ folder. Do NOT create placeholder references like background-xxx.png for images that don't exist.
  2. Use original images from source when available
  3. Copy missing images to _files_/ directory rather than substituting
  4. Resolve paths per-image - don't assume same directory
  5. URL-encode spaces: my file.png → my%20file.png
  6. Escape special characters:

- Parentheses: ( → %28, ) → %29 - Other special chars as needed

  1. Verify existence before referencing - run ls or glob to confirm file exists
  2. Maintain semantic relevance - image should match slide content
  3. Section intros without images are OK - If no background image exists for a section intro slide, just use the title and speaker notes without an image reference

Image Position Standards

Decision tree for image positioning:

Is it a section intro slide with just title?
  → Does a background image exist?
    → YES: Use ![]() for full background
    → NO: Skip image, use title + speaker notes only

Is it a diagram embedded in flowing text?
  → YES: Use ![inline]() or ![inline fill]()

Is it the main visual for the slide with bullet points?
  → YES: Use ![right fit]() ⭐ PRIMARY FORMAT

Does it need specific sizing?
  → YES: Use ![right 80%]() or other percentage

Speaker Notes Rules

When to convert to speaker notes:

  • Multi-sentence explanatory paragraphs
  • Contextual information not needed on slide
  • Talking points for the speaker
  • Additional details for verbal explanation

When NOT to convert:

  • Single-sentence paragraphs (might be slide content)
  • Bullet or numbered lists (usually slide content)
  • Block quotes (usually featured content)
  • Sentences ending with : (usually introducing lists)
  • Any header level
  • Table content

Format:

Slide content here.

^ Speaker note paragraph one. Can be multiple sentences providing context.

^ Speaker note paragraph two. Each note paragraph gets its own ^ prefix.

Content Standards

  1. File Naming:

- Always use - slides suffix - Preserve original filename otherwise - Example: AI for PKM.md → AI for PKM - slides.md

  1. Emoji Usage:

- Add to H1 and H2 headers - Choose relevant, professional emojis - Don't overuse - quality over quantity - Examples: - 🎯 Goals/Objectives - 📊 Data/Charts - 🤖 AI/Technology - 💡 Ideas/Insights - 🚀 Future/Launch - 📚 Learning/Knowledge

  1. Hierarchy:

- H1 for major sections (usually numbered) - H2 for main slide titles - H3 for subtopics within slides - Consistent numbering scheme

  1. Slide Breaks:

- Use --- (three dashes) on its own line - Blank line before and after recommended - Logical breaks between topics

Platform-Specific Notes

Deckset (Default)

Format:

  • Speaker notes: ^ prefix on each line
  • Image positioning: All formats fully supported
  • Frontmatter: Simple key-value pairs

Features:

  • Automatic slide numbering with slidenumbers: true
  • Full control over image positioning
  • Rich presenter notes support

Example Frontmatter:

slidenumbers: true
autoscale: true
theme: Plain Jane, 3

Marp (Optional)

Format:

  • Speaker notes: HTML comments <!-- speaker note -->
  • Image positioning: May use different syntax
  • Frontmatter: YAML with marp: true

Conversion for Marp:

---
marp: true
paginate: true
---

<!-- This is a speaker note in Marp -->

Differences:

  • HTML comments instead of ^ prefix
  • Different frontmatter structure
  • May need ![bg]() for backgrounds instead of ![]()

Quality Checklist

Before marking task complete, verify:

  • No invented images - every image reference points to a real file
  • All images have valid, URL-encoded paths
  • Image files exist at specified locations (run ls to confirm)
  • Spaces in paths converted to %20
  • Special characters properly escaped
  • Speaker notes use ^ prefix (or HTML for Marp)
  • Speaker notes placed at END of each slide
  • Slide dividers (---) at logical breaks
  • Consistent header hierarchy (H1 → H2 → H3)
  • Emojis added appropriately to titles
  • File saved with '- slides' suffix
  • Content hierarchy maintained from source
  • No orphaned content or broken sections
  • Internal comments removed
  • Frontmatter added if requested

Error Handling

Missing Images

Problem: Source references image that doesn't exist

Solution:

  1. Check source _files_/ directory
  2. Check parent directory _files_/
  3. Search for image by name in project
  4. If found: Copy to presentation _files_/ directory
  5. If not found: Note in speaker notes and skip image

Broken Relative Paths

Problem: Path doesn't resolve correctly

Solution:

  1. Determine slide file location
  2. Calculate relative path from slide to image
  3. Test path resolution
  4. URL-encode the working path

Invalid Characters in Paths

Problem: Special characters break image links

Solution:

  • Spaces: Convert to %20
  • Parentheses: ( → %28, ) → %29
  • Brackets: [ → %5B, ] → %5D
  • Other special chars: Use URL encoding

Tips for Best Results

  1. Start with outline: Create clear section and slide structure first
  2. One idea per slide: Each slide should have single clear message
  3. Visual hierarchy: Use H1 for sections, H2 for slides, H3 for sub-points
  4. Image positioning: Default to ![right fit]() for content slides
  5. Speaker notes: Add context that helps speaker but clutters slide
  6. Emoji consistency: Use similar emojis for similar concepts
  7. Test in Deckset: Preview the slides to verify formatting
  8. Iterate: First pass for structure, second for polish

平台分布

Claude Code

31.05%
按下载量换算95

Antigravity

24.22%
按下载量换算74

Codex

17.36%
按下载量换算53

Gemini CLI

11.22%
按下载量换算34

OpenCode

8.49%
按下载量换算26

windsurf

3.68%
按下载量换算11

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。

继续浏览同类 Skills