本文へ移動
cccskills
無料GitHub で公開

mermaid-tools

Extracts Mermaid diagrams from markdown files and generates high-quality PNG images using bundled scripts. Activates when working with Mermaid diagrams, converting diagrams to PNG, extracting diagrams from markdown, or processing markdown files with embedded Mermaid code.

インストール方法を見る

含まれるファイル(5)

  • SKILL.md5.4 KB
  • references/setup_and_troubleshooting.md5.1 KB
  • scripts/extract_diagrams.py4.9 KB
  • scripts/extract-and-generate.sh4.9 KB
  • scripts/puppeteer-config.json356 B

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

Mermaid Tools

Overview

This skill enables extraction of Mermaid diagrams from markdown files and generation of high-quality PNG images. The skill bundles all necessary scripts (extract-and-generate.sh, extract_diagrams.py, and puppeteer-config.json) in the scripts/ directory for portability and reliability.

Core Workflow

Standard Diagram Extraction and Generation

Extract Mermaid diagrams from a markdown file and generate PNG images using the bundled extract-and-generate.sh script:

cd "${CLAUDE_SKILL_DIR}/scripts"
./extract-and-generate.sh "<markdown_file>" "<output_directory>"

Parameters:

  • <markdown_file>: Path to the markdown file containing Mermaid diagrams
  • <output_directory>: (Optional) Directory for output files. Defaults to <markdown_file_directory>/diagrams

Example:

cd "${CLAUDE_SKILL_DIR}/scripts"
./extract-and-generate.sh "<markdown_file>" "<output_directory>"

What the Script Does

  1. Extracts all Mermaid code blocks from the markdown file
  2. Numbers them sequentially (01, 02, 03, etc.) in order of appearance
  3. Generates .mmd files for each diagram
  4. Creates high-resolution PNG images with smart sizing
  5. Validates all generated PNG files

Output Files

For each diagram, the script generates:

  • 01-diagram-name.mmd - Extracted Mermaid code
  • 01-diagram-name.png - High-resolution PNG image

The numbering ensures diagrams maintain their order from the source document.

Advanced Usage

Custom Dimensions and Scaling

Override default dimensions using environment variables:

cd "${CLAUDE_SKILL_DIR}/scripts"
MERMAID_WIDTH=1600 MERMAID_HEIGHT=1200 ./extract-and-generate.sh "<markdown_file>" "<output_directory>"

Available variables:

  • MERMAID_WIDTH (default: 1200) - Base width in pixels
  • MERMAID_HEIGHT (default: 800) - Base height in pixels
  • MERMAID_SCALE (default: 2) - Scale factor for high-resolution output

High-Resolution Output for Presentations

cd "${CLAUDE_SKILL_DIR}/scripts"
MERMAID_WIDTH=2400 MERMAID_HEIGHT=1800 MERMAID_SCALE=4 ./extract-and-generate.sh "<markdown_file>" "<output_directory>"

Print-Quality Output

cd "${CLAUDE_SKILL_DIR}/scripts"
MERMAID_SCALE=5 ./extract-and-generate.sh "<markdown_file>" "<output_directory>"

Smart Sizing Feature

The script automatically adjusts dimensions based on diagram type (detected from filename):

  • Timeline/Gantt: 2400×400 (wide and short)
  • Architecture/System/Caching: 2400×1600 (large and detailed)
  • Monitoring/Workflow/Sequence/API: 2400×800 (wide for process flows)
  • Default: 1200×800 (standard size)

Context-aware naming in the extraction process helps trigger appropriate smart sizing.

Important Principles

Use Bundled Scripts

CRITICAL: Use the bundled extract-and-generate.sh script from this skill's scripts/ directory. All necessary dependencies are bundled together.

Change to Script Directory

Run the script from its own directory to properly locate dependencies (extract_diagrams.py and puppeteer-config.json):

cd "${CLAUDE_SKILL_DIR}/scripts"
./extract-and-generate.sh "<markdown_file>" "<output_directory>"

Running the script without changing to the scripts directory first may fail due to missing dependencies.

Prerequisites Verification

Before running the script, verify dependencies are installed:

  1. mermaid-cli: mmdc --version
  2. Google Chrome: google-chrome-stable --version
  3. Python 3: python3 --version

If any are missing, consult references/setup_and_troubleshooting.md for installation instructions.

Troubleshooting

For detailed troubleshooting guidance, refer to references/setup_and_troubleshooting.md, which covers:

  • Browser launch failures
  • Permission issues
  • No diagrams found
  • Python extraction failures
  • Output quality issues
  • Diagram-specific sizing problems

Quick fixes for common issues:

Permission denied:

chmod +x "${CLAUDE_SKILL_DIR}/scripts/extract-and-generate.sh"

Low quality output:

MERMAID_SCALE=3 ./extract-and-generate.sh "<markdown_file>" "<output_directory>"

Chrome/Puppeteer errors: Verify all WSL2 dependencies are installed (see references for full list).

Bundled Resources

scripts/

This skill bundles all necessary scripts for Mermaid diagram generation:

  • extract-and-generate.sh - Main script that orchestrates extraction and PNG generation
  • extract_diagrams.py - Python script for extracting Mermaid code blocks from markdown
  • puppeteer-config.json - Chrome/Puppeteer configuration for WSL2 environment

All scripts must be run from the scripts/ directory to properly locate dependencies.

references/setup_and_troubleshooting.md

Comprehensive reference documentation including:

  • Complete prerequisite installation instructions
  • Detailed environment variable reference
  • Extensive troubleshooting guide
  • WSL2-specific Chrome dependency setup
  • Validation procedures

Load this reference when dealing with setup issues, installation problems, or advanced customization needs.

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

Fixes web search on an agent whose model backend can't run it: a relay/reseller proxying Claude or Codex returns empty instead of failing. Use when web search returns nothing, a model insists a shipped product doesn't exist, someone wants to give an agent internet access, or the user is on a third-party base URL, relay, or 中转站. Diagnoses which built-in tools are dead, removes them, and installs a working replacement.

日本語の概要は準備中です。原文の説明を表示しています。

daymade/claude-code-skills1,4522026年10月11日 更新

抓取 A 股消息面情报:从财联社、华尔街见闻、金十、新浪 7x24、东财快讯、 证监会/央行/上交所/财政部政策公告、东方财富股吧等公开来源抓取与股票相关的 新闻、政策、情绪,输出结构化 JSON 或 Markdown。 当用户提到“A 股消息面”、“抓新闻”、“个股消息”、“政策监管”、“股吧情绪”、 “财联社”、“东财快讯”、“市场情绪”或需要把某只股票相关的公开情报聚合出来时 触发。也适用于“帮我看看 000001 最近有什么消息”这类口语化请求。

日本語の概要は準備中です。原文の説明を表示しています。

daymade/claude-code-skills1,4522026年10月11日 更新

Transcribes audio or video to speaker-labeled, timestamped text, locally with MLX on Apple Silicon or remotely. Use for 转录 / 录音转文字 / 说话人分离 / 字幕, and also for preparing audio for ASR without transcribing: 转格式, 降采样到 16kHz, merging recorder segments, or compressing and speeding up audio before 飞书妙记 — even when it looks like a one-line ffmpeg job.

日本語の概要は準備中です。原文の説明を表示しています。

daymade/claude-code-skills1,4522026年10月11日 更新

Routes audio: StepFun ASR/语音识别, StepFun TTS/配音, transcript/妙记→会议纪要, merge/review minutes. Reads one bundled specialist; generic ASR and correction stay direct.

日本語の概要は準備中です。原文の説明を表示しています。

daymade/claude-code-skills1,4522026年10月11日 更新

Diagnoses and repairs repository setup and guarded Git workflows for Claude Code or Codex — environment repair, startup sync, hook auditing, collaborator handoff. Use when a repo won't run, a teammate onboards, hook output duplicates, or commit/push/conflict needs guarding. Not for lost-commit recovery (use git-safety-net), GitHub ops (use github-ops), or history scrubbing (use github-sensitive-data-cleanup).

日本語の概要は準備中です。原文の説明を表示しています。

daymade/claude-code-skills1,4522026年10月11日 更新

Runs adversarial due-diligence on a benchmark the user envies — a founder, KOL, company, or product whose success looks inflated — splitting marketing bubble from real signal, then mapping the validated playbook onto the user's own resources. Use for 尽调/对标/拆解 a competitor, 抄/偷师 their playbook, or suspecting 水分/泡沫 in claims. Prefer over deep-research when debunking inflated claims, not a neutral briefing.

日本語の概要は準備中です。原文の説明を表示しています。

daymade/claude-code-skills1,4522026年10月11日 更新

daymade のスキルをすべて見る

このスキルの問題を報告する