Playwright Skill 🧩 Skill Free

What it does

Playwright Skill is a community Agent Skill that lets a coding agent write and run its own Playwright programs for browser automation and testing. You describe a check in plain words, such as testing a signup flow, screenshotting a dashboard at mobile and desktop sizes, or looking for broken links. The agent writes a custom script, runs it, and reports back with console output and screenshots. By default the browser is visible, so you can watch the run.

What is inside

The skill lives in skills/playwright-skill/ and includes the following files:

• A short SKILL.md
• run.js, an executor that runs both file and inline scripts with reliable module resolution
• lib/helpers.js, with optional helper functions
• API_REFERENCE.md, which the agent loads only when it needs detail on selectors, network interception, authentication, visual regression or mobile emulation

It also detects running dev servers and cleans up temporary files. You can set SLOW_MO to slow down a run and PW_ARTIFACT_DIR to choose where screenshots are saved.

Works with

It follows the open Agent Skills format and is also packaged as a Claude Code plugin. The README names Claude Code, Cursor, GitHub Copilot, Codex, Gemini CLI and OpenCode as clients that can use it.

How to install

With Vercel's skills CLI:

npx skills add lackeyjb/playwright-skill --skill playwright-skill --global --yes

Or as a Claude Code plugin:

/plugin marketplace add lackeyjb/playwright-skill
/plugin install playwright-skill@playwright-skill

Then run npm run setup in the installed skill folder to install Playwright and Chromium.

Maintenance and safety

The skill is MIT licensed and maintained by one developer. It needs Node.js, and it runs code the agent writes against whatever sites you point it at, so review scripts that submit forms or touch production. The README itself recommends Microsoft's Playwright CLI or Playwright MCP for simple interactive browsing. This skill is the better fit when the script is the thing you want to keep.

Who should use it

Developers who want repeatable, code-first browser tests or automations written by their agent, with loops, assertions and several browser contexts, and who prefer scripts to an MCP server.