Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    BrowserLearn Agent Skills

    2

    Record Chrome tasks and export them to Playwright/Stagehand scripts or run them via MCP tools.

    Free

    Security scanned
    BrowserLearn Agent Skills

    BrowserLearn Agent Skills

    Example session with this skill installed

    List my current BrowserLearn workflows and export the one named "Github Star" to a Playwright script in the current directory.

    • Read your context and instructions
    • Compiled the browserlearn agent

    I have listed your local BrowserLearn workflows and identified the "Github Star" workflow. I am now exporting it to a Playwright-compatible JavaScript file.

    Exported Playwright Script

    import { test, expect } from '@playwright/test';
    
    /**
     * BrowserLearn Export: Github Star
     * Format: playwright
     */
    test('Github Star workflow', async ({ page }) => {
      // Navigate to initial URL
      await page.goto('https://github.com/droidmaximus/browserlearn');
    
      // Click the Star button
      // Note: Selectors captured from recording
      await page.click('button[aria-label="Star this repository"]');
    
      // Parameterized inputs would be injected here if defined
      console.log('Workflow execution complete.');
    });
    

    Next steps

    1. Install dependencies using npm install @playwright/test.
    2. Run the script with npx playwright test GithubStar.mjs.
    3. Check for any required environment variables if the workflow uses parameters.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Stop manually scripting browser automations from scratch. This skill provides the interface to BrowserLearn, allowing you to record complex browser tasks in Chrome and immediately convert them into structured workflows or production-ready code.

    What it does

    • Record workflows via a Chrome extension to capture clicks, typing, and navigation.
    • Export scripts to standalone Playwright or Stagehand .mjs files for CI/CD or agent stacks.
    • Execute automations directly through MCP tools or the CLI with full parameter support.
    • Inspect steps and input schemas to debug locators and required variables before running.
    • Manage recordings through a local-first dashboard and searchable database.

    How it works

    1. Setup: Install the local server and pair the Chrome extension using a unique code.
    2. Capture: Record a browser task; BrowserLearn redacts sensitive values and saves the flow locally.
    3. Connect: Link the MCP server to your agent host to enable list_workflows and run_workflow tools.
    4. Automate: Run workflows via the agent with custom parameters or export them to code for your repo.

    Frameworks & tools

    Built for Node.js 22.13+, Playwright, Stagehand, and the Model Context Protocol (MCP). It integrates with Chrome via a manifest v3 extension.

    Why this beats prompting it yourself

    Prompting an LLM to write a browser script often results in broken selectors and hallucinated DOM structures. This tool captures real-world locators and interactions during a live session, ensuring the agent executes against verified selectors rather than guesses.

    Use cases

    • Generate Playwright scripts for regression testing from a manual walkthrough.
    • Build parameterized browser agents that perform repetitive SaaS tasks like invoicing.
    • Discover and trigger local automation workflows through natural language.
    • Scaffold Stagehand-based agent flows without writing boilerplate interaction code.

    Known limitations

    Requires Node 22.13 or newer. Does not support training browser models or modifying MCP tool names. Exporting as an Agensi skill is not currently supported by the CLI.

    Requires Node 22.13+. Limited to Playwright and Stagehand exports. fresh automation contexts may require passing credentials as parameters rather than relying on session cookies.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 13 days ago

    • Free with an Agensi account

    Requires

    browserlearn
    Node.js
    playwright
    Git

    Permissions

    Terminal / Shell
    Browser
    Write Files
    Network Access
    Listed13 days ago

    What's inside

    Frequently Asked Questions

    By recording live interactions in Chrome, the skill captures real-world CSS selectors and DOM structures, preventing the common hallucinations and broken locators that occur when asking an LLM to write scripts from scratch.

    The skill is compatible with any agent host that supports the Model Context Protocol (MCP) and requires Node.js 22.13 or newer for the local server environment.

    You receive the BrowserLearn local server, the Chrome manifest v3 extension for workflow capture, and the MCP integration tools to connect the workflows to your AI agent.

    Yes, you can export your captured workflows as standalone Playwright or Stagehand .mjs files, making it easy to integrate them into existing CI/CD pipelines or larger agentic codebases.

    Setup involves installing the local server via Node, pairing the included Chrome extension using a unique connection code, and then linking the MCP server to your preferred agent host.

    While it automates browser tasks, it is not a training tool for new LLM browser models; it focuses on executing and exporting reliable, human-verified automation flows.

    New in Testing & Debugging