lackeyjb/playwright-skill

What it solves

It allows coding agents to perform general-purpose browser automation by writing and executing Playwright code on the fly. Unlike simple interactive browsing tools, this project enables agents to create complex, reusable automation scripts involving loops, assertions, network interception, and multi-step flows.

How it works

When a user describes a browser task, the agent uses this skill to write custom Playwright code. This code is then processed by a universal executor (run.js) that handles module resolution and executes the script. By default, the browser runs in non-headless mode so the user can see the automation in real-time, and the agent returns the results via console output and screenshots.

Who it’s for

Developers using AI coding agents (such as Claude Code, Cursor, or GitHub Copilot) who need their agents to perform automated testing, visual verification, or complex web interactions.

Highlights

• Dynamic Code Generation: Agents write custom scripts for specific requests rather than relying on pre-built templates.
• Visible Execution: Defaults to headless: false for real-time monitoring of the automation process.
• étude Portable Executor: Includes a dedicated executor to ensure stable module resolution for both inline and file-based scripts.
• Progressive Disclosure: Uses a concise instruction file (SKILL.md) and only loads the full API reference when the agent requires detailed technical documentation.