Playwright Browser Automation

0•

Automates browser interactions and web testing using Playwright CLI and specialized interaction templates.

This skill provides a powerful hybrid approach to browser automation within Claude Code, leveraging the Playwright CLI for quick tasks and JavaScript templates for complex interactions. It enables developers to capture high-quality screenshots and PDFs, automate form submissions, and monitor console or network activity without the need for a dedicated MCP server. Ideal for developers needing to perform end-to-end testing, visual regression checks, or automated web data collection directly from their terminal-based AI environment, this tool streamlines web workflows with minimal configuration.

主な機能

01Perform complex user interactions including clicks, hovers, and drag-and-drop
02Capture full-page screenshots and generate PDFs via simple CLI commands
03Automate multi-field form filling with built-in validation
04Monitor real-time console messages and network requests
05Conduct accessibility audits and snapshot captures for UI/UX testing
060 GitHub stars

ユースケース

01Scraping dynamic content and monitoring network logs during page interactions
02Automating end-to-end user journey testing on web applications
03Generating visual documentation and performance snapshots for web pages
name description
browser-automation
Automates browser interactions using Playwright CLI and templates - takes screenshots, generates PDFs, clicks elements, fills forms, monitors console/network. Use when testing web pages, automating browser tasks, or when user mentions screenshots, web testing, form automation, or Playwright

Browser Automation

Overview

Hybrid browser automation using Playwright CLI for simple tasks and templates for complex interactions. Provides efficient automation without MCP server overhead.

When to Use

Use this skill when you need to:

  • Take screenshots or generate PDFs
  • Test web pages or forms
  • Click elements or fill inputs
  • Monitor console logs or network requests
  • Automate browser interactions

Don't use when:

  • Simple HTTP requests suffice (use fetch/curl)
  • Need persistent browser sessions across conversations

Hybrid Approach

CLI-Native (Simple Tasks)

Use Playwright CLI directly for:

  • Screenshot: npx playwright screenshot <url> <filename>
  • PDF: npx playwright pdf <url> <filename>
  • Open browser: npx playwright open <url> (for manual inspection)

Templates (Complex Interactions)

Use templates (16 total) for tasks requiring state management, event handling, or multi-step logic:

User Interactions (6)

  • Click element with wait
  • Fill form field with validation
  • Hover over element
  • Drag and drop
  • Press keyboard key
  • Select dropdown option

Form Handling (2)

  • Fill multiple fields (bulk)
  • Upload file

Page Monitoring (3)

  • Capture accessibility snapshot
  • Get console messages (event listening)
  • Get network requests (event listening)

Advanced (5)

  • Evaluate JavaScript
  • Navigate back with wait
  • Resize browser window
  • Handle dialogs (alert/confirm/prompt)
  • Wait for element with timeout

Quick Start

First time setup:

npm install -g playwright
npx playwright install chromium

CLI Examples (Simple Tasks)

# Screenshot
npx playwright screenshot https://example.com ./tmp/playwright/screenshot.png

# PDF
npx playwright pdf https://example.com ./tmp/playwright/page.pdf

# Open browser for manual inspection
npx playwright open https://example.com

Template Examples (Complex Tasks)

# Click element (requires wait logic)
node $SKILL_DIR/templates/click.js https://example.com "button:has-text('Submit')"

# Fill form (multiple fields)
node $SKILL_DIR/templates/fill-form.js https://example.com '{"input[name=email]":"test@example.com"}'

# Monitor console (event listening)
node $SKILL_DIR/templates/console-messages.js https://example.com

Important Guidelines:

  • Prefer CLI first: Use npx playwright for screenshots, PDFs, and simple evaluation
  • Use templates when needed: Complex interactions, event listening, multi-step logic
  • $SKILL_DIR is automatically set to the skill's absolute path by Claude
  • Never use cd before running commands
  • Never write inline Playwright code with node -e or heredocs

Note: Output files are saved to ./tmp/playwright/ in the current project directory by default.

Documentation

  • Complete usage guide: See reference/TEMPLATES.md
  • Template reference: Each template includes usage instructions in file header