Browser Access

    1

    Routes browser tasks to the right MCP (Claude in Chrome or Chrome DevTools) and handles logins without touching passwords

    Free

    1 useSecurity scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Browser Access

    Example session with this skill installed

    Open my local dev server at localhost:3000, check if the login button is visible, and if so, record a GIF of the loading state when I click it.

    • Read your context and instructions
    • Compiled the browser access

    I will now inspect your local development server at localhost:3000 using the Claude in Chrome MCP to verify the UI and record the interaction.

    1. Tab Context: Checked existing tabs and created a new tab for localhost:3000.
    2. UI Verification: Located the login button in the DOM and confirmed visibility via screenshot.
    3. Interaction: Triggered the gif_creator tool while initiating the click event.

    Browser Inspection Results

    ComponentStatusObservation
    URLhttp://localhost:3000Page loaded successfully.
    Login ButtonVisibleSelector button[type="submit"] found in header.
    Console LogsCleanNo errors detected during initial load.
    GIF RecordingCompleteRecorded 3.5 seconds of the transition to the dashboard.

    The login button is present. I have recorded the loading state GIF as requested. The recording shows the spinner appearing for 400ms before the navigation occurs.

    Next steps

    • Review the captured GIF in the output directory to verify the spinner animation.
    • Let me know if you want to run a Lighthouse audit on this page to check performance scores.
    • I can also inspect the network requests if the loading state takes longer than expected.

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

    About this skill

    What it does

    Browser Access tells your agent which browser MCP to use for a task, and what to do when it fails.

    • Default to Claude in Chrome for opening pages, clicking, filling forms, screenshots, reading DOM/console/network, and anything behind a login. It attaches to your real Chrome profile, so existing sessions just work.
    • Fall back to Chrome DevTools MCP only for the instrumentation it uniquely provides (e.g. performance traces), or when Claude in Chrome is unavailable.
    • Login stays with you. When a page shows a login screen or an MCP returns 401/403, the agent asks you to sign in in the browser and only verifies the result. It never types your password.
    • Troubleshooting playbook for disconnected extensions, missing site permissions, and DevTools ports that are not running.

    Who it's for

    Developers who use Claude Code (or another SKILL.md-compatible agent) for UI verification, visual checks, network debugging, or dev-server inspection, and are tired of the agent picking the wrong browser tool or getting stuck on auth.

    Requirements

    • Claude in Chrome extension and/or Chrome DevTools MCP configured in your agent.
    • No API keys or credentials needed.

    Cannot handle raw credentials in chat. Browser dialogs (alert/confirm) block execution. Performance tools require remote debugging port 9222. Claude in Chrome requires extension installation.

    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

    1 use

    Times used to date

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 15 days ago

    • Free with an Agensi account

    Requires

    Chrome Devtools MCP
    Node.js

    Permissions

    Terminal / Shell
    Listed1 month ago
    Updated15 days ago

    What's inside

    Frequently Asked Questions

    This skill uses a dual-path workflow that prioritizes your existing Chrome profile. Unlike standard agents that start fresh, this skill attaches to your active session, allowing it to navigate dashboards and internal tools that are already logged in.

    Yes, it is specifically optimized for use with Claude in Chrome and the Chrome DevTools MCP. It requires the Google Chrome browser and the remote debugging protocol to be enabled for deep instrumentation tasks like memory profiling.

    The skill includes a pre-configured workflow for both visual interaction and technical auditing. You get the logic for automated UI testing, screenshot/GIF capture, and deep-dive performance tools like Lighthouse and heap snapshots.

    While the agent can automate clicks and data entry, it cannot see or interact with native browser pop-ups like 'Alert' or 'Confirm' boxes. You will need to manually click 'OK' or 'Cancel' on those specific modals before the agent can continue its task.

    Absolutely. This skill is designed for technical QA, allowing you to trigger Lighthouse audits, track network requests, and perform heap snapshots to identify memory leaks or performance bottlenecks in real-time.

    New in Agents & Orchestration