Skip to content
Get packs

Browser Use (Live Chrome Control)

Lets an agent drive your signed-in Chrome profile through a relay for login-dependent work and UI checks

Source Download skill
Repository
steipete/agent-scripts
Stars
7,433
Licence
MIT
Files
3 files

What it does

Once installed, the agent is told to control your real, already-signed-in Chrome browser rather than a fresh test browser, so it can work on pages that depend on your cookies, single sign-on and extensions. It prefers a Chrome plugin if available, then an OpenClaw extension relay through the mcporter tool, and only falls back to direct DevTools attachment as a last resort. It also includes a helper script that watches for Chrome's 'Allow remote debugging?' prompt and presses Allow automatically.

Great for

  • Developers verifying a rendered web UI bug in their own logged-in Chrome
  • People automating work on sites that need their existing login, SSO or device trust
  • Users already running the OpenClaw and mcporter tooling who want a reliable local Chrome relay

Skip it if

  • Anyone who does not want an agent operating their personal signed-in browser: it can act with your accounts and sessions
  • Sensitive sessions (banking, email, admin consoles) where an auto-pressed Allow prompt could expose data; the auto-allow script removes a security confirmation step
  • Users not running OpenClaw or mcporter, since most of the routing depends on that setup

Try asking

  • “Check whether the checkout page in my logged-in Chrome shows the correct discount after my last change”
  • “Open my admin dashboard in Chrome and verify the new settings panel renders without errors”
  • “Use my existing Chrome session to export the list of open tabs in the OpenClaw group”
Step-by-step guide: how to use Browser Use (Live Chrome Control)Install, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 3 original files, the MIT licence and an attribution note.

    browser-use.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip browser-use.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip browser-use.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Read the auto-allow.py script before running it, since it presses the Allow button on a remote-debugging prompt; keep it off for sensitive browsing sessions.
  • Set the relay access mode to Selected tabs and use the 'OpenClaw' tab group if you want to limit what the agent can see.
  • Never paste relay keys or credentials into chat; the skill says they should stay in the protected OpenClaw credentials directory.

What's inside

  • auto-allow.py
  • mcporter-config.md
  • SKILL.md

Redistributed unmodified from steipete/agent-scripts under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.