Token导航 LogoToken导航TokenDH.com
研究检索需要联网github未标认证来源可访问许可证需确认审计通过

markdown-optimizationMarkdown optimization 搜索

Agent Skill

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。

总安装

303

周安装

13

GitHub Stars

21

下载量

106
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

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-optimization(Markdown optimization 搜索)
来源仓库:https://github.com/thapaliyabikendra/ai-artifacts
仓库路径:skills/markdown-optimization
安装命令:
npx skills add https://github.com/thapaliyabikendra/ai-artifacts --skill markdown-optimization
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。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/thapaliyabikendra/ai-artifacts --skill markdown-optimization

简介

用于 Markdown 文档的性能与可读性优化,提升信息传递效率。

  • 可自动精简冗余语句、优化标题层级或增强语义连贯性。
  • 建议结合具体项目上下文使用,避免破坏原有技术细节。
  • 涉及用户手册等正式文档时,应确保术语准确性和版本一致性。
  • markdown-optimization 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

Markdown Optimization

Patterns and techniques for analyzing and optimizing markdown documentation.

Quick Start

Size Check

# Check single file
wc -l docs/README.md

# Check all markdown in folder
find docs -name "*.md" -exec wc -l {} \; | sort -n

# Find files over 500 lines
find . -name "*.md" -exec sh -c \
  'lines=$(wc -l < "$1"); [ "$lines" -gt 500 ] && echo "$1: $lines"' _ {} \;

Structure Validation

Check for common issues:

  • Missing TOC (files > 100 lines)
  • Heading hierarchy violations (h1 → h3, skipping h2)
  • Sections over 200 lines

Document Profiles

Different document types have different optimization rules.

Authoritative limits: For Claude Code artifacts (agent, skill, command, guidelines), size limits are defined in GUIDELINES.md. This skill enforces those limits.
ProfileTargetKey Focus
claude-mdProject contextConcise, table-heavy
architectureSystem designDiagrams external, decisions linked
domainBusiness docsEntity tables, numbered rules
feature-specFeature docsSeparated concerns
readmeEntry pointsQuick start, links
agentAgent promptsLean, reference skills
skillSkill docsProgressive disclosure
commandSlash commandsFocused action
guidelinesMeta-knowledgeModular structure

Full profile definitions: See profiles.md

Core Analysis Patterns

1. Size Analysis

## Size Metrics
| File | Lines | Limit | Status |
|------|-------|-------|--------|
| README.md | 180 | 200 | ✅ OK |
| business-rules.md | 620 | 400 | ❌ OVER (+220) |

Actions for oversized files:

  • Split by topic/domain
  • Extract examples to separate files
  • Convert prose to tables
  • Move detailed content to references

2. Duplication Detection

Look for:

  • Identical paragraphs across files
  • Same concepts explained differently
  • Copy-pasted code blocks
  • Repeated tables/lists
## Duplication Findings
| Concept | Found In | Action |
|---------|----------|--------|
| "Entity base classes" | patterns.md:45, entities.md:12 | Consolidate to entities.md |
| "Build commands" | README.md:30, CLAUDE.md:15 | Single source in CLAUDE.md |

3. Structure Assessment

Heading Hierarchy:

# Title (h1) - Only one per file
## Section (h2)
### Subsection (h3)
#### Detail (h4) - Use sparingly

Issues to flag:

  • Multiple h1 headings
  • Skipped levels (h1 → h3)
  • Sections > 200 lines without subsections
  • Missing TOC for files > 100 lines

4. Link Validation

# Find all markdown links
grep -oP '\[.*?\]\(.*?\)' file.md

# Check if targets exist
# Internal: [text](path.md) or [text](#anchor)
# External: [text](https://...)

Common issues:

  • Broken relative paths
  • Missing anchor targets
  • Orphaned files (not linked anywhere)

5. Compaction Opportunities

Identify verbose patterns that can be condensed.

Full techniques: See compaction-patterns.md

Quick reference:

PatternBeforeAfterSavings
Prose listParagraph with itemsBullet list40-60%
Explanation tableMulti-paragraphTable50-70%
Code commentsHeavily commentedSelf-documenting30-50%
Repeated structureCopy-paste sectionsTemplate + instances60-80%

Optimization Techniques

Convert Prose to Tables

Before (12 lines):

The system supports three user roles. Administrators have full access
to all features including user management, system configuration, and
reporting. Managers can access reports and manage their team members
but cannot modify system settings. Regular users can only access their
own data and perform basic operations within their assigned scope.

After (6 lines):

| Role | Access |
|------|--------|
| Admin | Full access: user management, config, reporting |
| Manager | Reports, team management (no system settings) |
| User | Own data, basic operations |

Extract Examples

Before (inline, 50 lines):

## API Usage

Here's how to use the API:

40 lines of code

After (referenced, 5 lines):

## API Usage

See [examples/api-usage.md](examples/api-usage.md) for complete examples.

Quick start:

client.get("/api/resource")

Use Progressive Disclosure

Before (everything inline):

## Feature X
[200 lines of content]

After (layered):

## Feature X

[Core concept - 30 lines]

**Advanced configuration**: See [feature-x-advanced.md](references/feature-x-advanced.md)
**API reference**: See [feature-x-api.md](references/feature-x-api.md)

Custom Profiles

Define custom profiles in .claude/config/md-profiles.yaml:

profiles:
  # Custom profile for API docs
  api-docs:
    patterns:
      - "docs/api/**"
      - "**/api-*.md"
    max_lines: 400
    rules:
      - require_toc: true
      - require_examples: true
      - max_section_lines: 150
    compaction:
      - prose_to_table: true
      - extract_examples: true

  # Custom profile for internal docs
  internal:
    patterns:
      - "internal/**"
    max_lines: 800  # More lenient
    rules:
      - require_toc: false

Report Templates

Audit Report Structure

# Markdown Optimization Report

## Target: {path}
## Profile: {profile}
## Generated: {timestamp}

## Executive Summary
- Files analyzed: N
- Issues found: N (X critical, Y warnings)
- Estimated optimization potential: N lines (X%)

## Detailed Findings

### Size Issues
[Table of oversized files]

### Structure Issues
[Table of structural problems]

### Duplication
[Table of duplicated content]

### Broken Links
[Table of link issues]

## Recommendations
[Prioritized action items]

Integration with Commands

This skill supports:

  • /docs:optimize-md - Primary command (use --profile guidelines for GUIDELINES.md ecosystem)

Related

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

38.2%
按下载量换算40

Claude

30.07%
按下载量换算32

Cursor

20.16%
按下载量换算21

Gemini CLI

8.95%
按下载量换算9

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

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。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills