- Home
- Skills
- Workflow Automation
- drive a live chrome over cdp
drive a live chrome over cdp
Attach an agent to a live, signed-in Chrome instance to automate authenticated web tasks without API keys.
$29
Works with the AI Agents you already use
drive a live chrome over cdp
Example session with this skill installed
Take over the Chrome I'm already signed into, tell me which page I'm on, and prove you attached to my real browser rather than a headless sandbox.
- Read your context and instructions
- Compiled the drive a live
Summary: attached to the live browser, confirmed the real Chrome, read the page.
$ node scripts/cdp.mjs check
{
"tickBox": { "file": "~/.config/google-chrome/Local State", "enabled": true },
"endpoint": { "port": 42453, "file": "~/.config/google-chrome/DevToolsActivePort", "ageHours": 202.5 },
"attach": { "ok": true, "wsPath": "/devtools/browser/<browser-id>" },
"hint": "The switch is on. If attach failed, the port belongs to another process or the profile differs."
}
$ node scripts/cdp.mjs version
{
"port": 42453,
"target": "https://example.com/dashboard",
"url": "https://example.com/dashboard",
"title": "Dashboard",
"userAgent": "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/152.0.0.0 Safari/537.36",
"webdriver": false,
"headlessHint": false
}
| Check | Result | Why it matters |
|---|---|---|
headlessHint | false | this is the human's own browser, not a sandbox pilot |
webdriver | false | not an automation-controlled instance |
attach.ok | true | the browser websocket accepted the session |
Next steps
node scripts/session.mjs start— hold one connection, so the human authorises once.node scripts/session.mjs text— read the page.node scripts/session.mjs eval "document.title"— anything the page can do.
Connects securely to your tools. The creator never sees your data.
About this skill
Your agent needs a logged-in browser. A fresh headless Chrome is empty — no logins, no cookies, no extensions — and replaying passwords is brittle and unsafe.
This skill attaches to the Chrome the human is actually using.
It gives your agent a single file, scripts/cdp.mjs, with zero npm dependencies (Node 22+), and a CLI that covers the jobs that matter:
- find the debugging port from Chrome's own
DevToolsActivePortfile - attach over the browser websocket — the route that keeps working when Chrome 150+ closes the legacy/jsonHTTP endpoints with a 404 - confirm you are on the real browser, not a headless pilot (headlessHint) - list tabs, read page text, run JavaScript, navigate, click, type, screenshot
The SKILL.md carries the eight traps that cost real days: stale port files that are not dead browsers, fixed ports that are sandboxes, React inputs that ignore .value =, downloads that follow a browser-wide path, and more.
Proven against Chrome 152 on Linux. No dependencies. No telemetry. No network calls of its own.
No support for CAPTCHA/paywall bypass. Multi-tab concurrency and iframe sessions are unverified. Launcher scripts are Linux-only; others require manual setup.
How to use
- 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
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 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
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 27 days ago
Requires