URL to Markdown
Fetches any URL via Chrome CDP and converts HTML to clean markdown.
Script Directory
Important: All scripts are located in
subdirectory of this skill.
Agent Execution Instructions:
- Determine this SKILL.md file's directory path as
- Script path =
${SKILL_DIR}/scripts/<script-name>.ts
- Replace all in this document with actual path
Script Reference:
| Script | Purpose |
|---|
| CLI entry point for URL fetching |
Features
- Chrome CDP for full JavaScript rendering
- Two capture modes: auto or wait-for-user
- Clean markdown output with metadata
- Handles login-required pages via wait mode
Usage
bash
# Auto mode (default) - capture when page loads
npx -y bun ${SKILL_DIR}/scripts/main.ts <url>
# Wait mode - wait for user signal before capture
npx -y bun ${SKILL_DIR}/scripts/main.ts <url> --wait
# Save to specific file
npx -y bun ${SKILL_DIR}/scripts/main.ts <url> -o output.md
Options
| Option | Description |
|---|
| URL to fetch |
| Output file path (default: auto-generated) |
| Wait for user signal before capturing |
| Page load timeout (default: 30000) |
Capture Modes
Auto Mode (default)
Page loads → waits for network idle → captures immediately.
Best for:
- Public pages
- Static content
- No login required
Wait Mode ()
Page opens → user can interact (login, scroll, etc.) → user signals ready → captures.
Best for:
- Login-required pages
- Dynamic content needing interaction
- Pages with lazy loading
Agent workflow for wait mode:
- Run script with flag
- Script outputs:
Page opened. Press Enter when ready to capture...
- Use to ask user if page is ready
- When user confirms, send newline to stdin to trigger capture
Output Format
markdown
---
url: https://example.com/page
title: "Page Title"
description: "Meta description if available"
author: "Author if available"
published: "2024-01-01"
captured_at: "2024-01-15T10:30:00Z"
---
# Page Title
Converted markdown content...
Mode Selection Guide
When user requests URL capture, help select appropriate mode:
Suggest Auto Mode when:
- URL is public (no login wall visible)
- Content appears static
- User doesn't mention login requirements
Suggest Wait Mode when:
- User mentions needing to log in
- Site known to require authentication
- User wants to scroll/interact before capture
- Content is behind paywall
Ask user when unclear:
The page may require login or interaction before capturing.
Which mode should I use?
1. Auto - Capture immediately when loaded
2. Wait - Wait for you to interact first
Output Directory
Each capture creates a file organized by domain:
url-to-markdown/
└── <domain>/
└── <slug>.md
Path Components:
- : Site domain (e.g., , )
- : Generated from page title or URL path (kebab-case)
Slug Generation:
- Extract from page title (preferred) or URL path
- Convert to kebab-case, 2-6 words
- Example: "Getting Started with React" →
getting-started-with-react
Conflict Resolution:
If
url-to-markdown/<domain>/<slug>.md
already exists:
- Append timestamp:
<slug>-YYYYMMDD-HHMMSS.md
- Example: exists →
getting-started-20260118-143052.md
Error Handling
| Error | Resolution |
|---|
| Chrome not found | Install Chrome or set env |
| Page timeout | Increase value |
| Capture failed | Try wait mode for complex pages |
| Empty content | Page may need JS rendering time |
Environment Variables
| Variable | Description |
|---|
| Custom Chrome executable path |
| Custom data directory |
| Custom Chrome profile directory |
Extension Support
Custom configurations via EXTEND.md.
Check paths (priority order):
.content-gen-skills/url-to-markdown/EXTEND.md
~/.content-gen-skills/url-to-markdown/EXTEND.md
If found, load before workflow. Extension content overrides defaults.