markdown-helper_skill

Overview

This skill provides token-efficient markdown parsing, editing, and diagram generation using native CLI tools on Windows, Mac, and Linux. It performs structure-focused operations (headers, tables, lists, stats, linting, bulk replace, and Mermaid diagram generation) without loading full file contents into context, saving significant token costs. Use it to speed common markdown workflows and reduce AI usage overhead.

How this skill works

The skill invokes small Node.js CLI utilities to parse files into compact outputs (headers, tables, lists, stats) or to run linters and Mermaid generators. Instead of embedding full files in the AI context, it runs targeted commands and returns concise results, minimizing tokens. Commands support JSON output, filters, format options, and safe bulk operations (dry-run, backups).

When to use it

• Extract headings, tables, lists, or task lists from large markdown files
• Generate Mermaid flowcharts, sequence diagrams, or Gantt charts from structured input
• Run linting/auto-formatting across many markdown files or fix common issues
• Perform bulk search-and-replace or preview changes with dry-run
• Collect quick file statistics (lines, words, headings, tables, links) without reading full content

Best practices

• Auto-activate the skill for structural queries (headers, tables, stats) on files >500 characters for best token savings
• Use --json output when you need machine-readable results for downstream automation
• Run dry-run before mass replacements and enable backups for safety
• Limit colored/terminal formatting calls to 2–3 outputs to avoid CLI task noise
• Prefer structure extraction with this skill; open full file in context when you need deep semantic understanding

Example use cases

• Extract all H2/H3 headings from README.md to build a project outline
• Parse tables from CHANGELOG.md and export them as CSV for reporting
• Generate a Mermaid flowchart for the checkout process and output as SVG
• Lint and auto-fix formatting in project-tasks/**/*.md across a repo
• Find and replace 'http://' with 'https://' across all markdown files with a dry-run preview