Playwrightによるブラウザ自動化のClaude Codeスキル。テストと検証のためのカスタム自動化を自律実行。
General-purpose Playwright automation for coding agents
An Agent Skill that enables coding agents to write and execute Playwright automation on the fly, from simple page tests to complex multi-step flows. It is also packaged as a Claude Code Plugin for convenient installation.
Claude autonomously decides when to use this skill based on your browser automation needs, loading only the minimal information required for your specific task.
Made using Claude Code.
headless: falseThis repository contains a standard Agent Skill and a Claude Code plugin wrapper. The recommended installation method is Vercel's skills CLI, which installs skills into the native locations for supported agents.
Use this skill when the agent needs to write a real Playwright program: loops, assertions, multiple contexts, network interception, screenshots, video, or a script you want to keep and rerun. It also provides dev-server detection and a small set of focused helpers.
For straightforward interactive browsing, start with Microsoft's official
@playwright/cli and install its
agent skills with playwright-cli install --skills. For tool-based browser
control with accessibility snapshots, use
playwright-mcp. This project is
the code-first option when the automation itself is the useful artifact.
This repository uses the plugin format with a nested structure:
playwright-skill/ # Plugin root
├── .claude-plugin/ # Plugin metadata
└── skills/
└── playwright-skill/ # The actual skill
└── SKILL.md
The repository keeps the skill inside the plugin's skills/ directory. Installers handle that layout automatically; manual copying is only a fallback for clients without an installer.
skills (Recommended)Install globally for your user:
npx skills add lackeyjb/playwright-skill --skill playwright-skill --global --yes
Install only for the current project by omitting --global:
npx skills add lackeyjb/playwright-skill --skill playwright-skill --yes
To target specific agents, add --agent followed by one or more agent IDs:
npx skills add lackeyjb/playwright-skill --skill playwright-skill --agent claude-code cursor --global --yes
After installation, run setup from the installed skill directory:
npm run setup
See the skills CLI documentation for supported agents and options.
Install via Claude Code's plugin system for automatic updates and team distribution:
# Add this repository as a marketplace
/plugin marketplace add lackeyjb/playwright-skill
# Install the plugin
/plugin install playwright-skill@playwright-skill
# Navigate to the skill directory and run setup
cd ~/.claude/plugins/marketplaces/playwright-skill/skills/playwright-skill
npm run setup
Verify installation by running /help to confirm the skill is available.
Agent Skills are supported by Claude Code, Cursor, GitHub Copilot, Codex,
Gemini CLI, OpenCode, and other clients. Install the directory containing
SKILL.md using the client’s documented skill path. If a client has no
installer, copy skills/playwright-skill/ into its documented skill directory
and run npm run setup there.
skills/playwright-skill/ folder to:
~/.claude/skills/playwright-skill/path/to/your/project/.claude/skills/playwright-skillcd ~/.claude/skills/playwright-skill # or your project path
npm run setup
Run /help to confirm the skill is loaded, then ask Claude to perform a simple browser task like "Test if google.com loads".
After installation, ask your agent to test or automate a browser task. It will write custom Playwright code, execute it, and return results with screenshots and console output.
"Test the homepage"
"Check if the contact form works"
"Verify the signup flow"
"Take screenshots of the dashboard in mobile and desktop"
"Test responsive design across different viewports"
"Fill out the registration form and submit it"
"Click through the main navigation"
"Test the search functionality"
"Check for broken links"
"Verify all images load"
"Test form validation"
Default settings:
false (browser visible unless explicitly requested otherwise)0ms by default; set SLOW_MO when usefulPW_ARTIFACT_DIR to choose another locationplaywright-skill/
├── .claude-plugin/
│ ├── plugin.json # Plugin metadata for distribution
│ └── marketplace.json # Marketplace configuration
├── skills/
│ └── playwright-skill/ # The actual skill (Claude discovers this)
│ ├── SKILL.md # What Claude reads
│ ├── run.js # Universal executor (proper module resolution)
│ ├── package.json # Dependencies & setup scripts
│ └── lib/
│ └── helpers.js # Optional utility functions
│ └── API_REFERENCE.md # Full Playwright API reference
├── README.md # This file - user documentation
├── CONTRIBUTING.md # Contribution guidelines
└── LICENSE # MIT License
Claude will automatically load API_REFERENCE.md when needed for comprehensive documentation on selectors, network interception, authentication, visual regression testing, mobile emulation, performance testing, and debugging.
npm run setup)npm run setup)Playwright not installed?
Navigate to the skill directory and run npm run setup.
Module not found errors?
Ensure automation runs via run.js, which handles module resolution.
Browser doesn't open?
Verify headless: false is set. The skill defaults to visible browser unless headless mode is requested.
Install all browsers?
Run npm run install-all-browsers from the skill directory.
Agent Skills are folders of instructions, scripts, and resources that agents can discover and use to do things more accurately and efficiently. When you ask Claude to test a webpage or automate browser interactions, Claude discovers this skill, loads the necessary instructions, executes custom Playwright code, and returns results with screenshots and console output.
This Playwright skill implements the open Agent Skills specification, making it compatible across agent platforms.
Contributions are welcome. Fork the repository, create a feature branch, make your changes, and submit a pull request. See CONTRIBUTING.md for details.
MIT License - see LICENSE file for details.
互換性
トピック