pair-agent

Share a local browser session with a remote AI agent via setup key exchange.

Updated Feb 24, 2026
One-click install
npx skills add https://github.com/Jia-Hong-Peng/jolike.com --skill pair-agent-jia-hong-peng
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: pair-agent
Source: https://github.com/Jia-Hong-Peng/jolike.com/tree/main/.claude/skills/gstack/pair-agent
Command: npx skills add https://github.com/Jia-Hong-Peng/jolike.com --skill pair-agent-jia-hong-peng

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill eliminates the manual hassle of configuring browser sharing between AI agents. It provides a secure, one-command way to generate a setup key and connect a remote agent to your local browser session with scoped permissions.

Core Features & Use Cases

  • One-Command Setup: Generates an expiring setup key and prints paste-ready instructions for the remote agent.
  • Scoped Access Control: Remote agents get isolated tabs with read+write access by default, and admin access on request.
  • Multi-Agent Compatibility: Works with OpenClaw, Hermes, Codex, Cursor, or any agent capable of making HTTP requests.
  • Use Case: You are debugging a UI issue in Claude Code and want a Cursor agent to inspect the same live browser session to compare DOM state without switching contexts.

Quick Start

Use the pair-agent skill to generate a setup key and share your browser with another AI agent.

Frequently Asked Questions about pair-agent

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I share my browser session with another AI agent?

You can share your browser session with a remote AI agent by starting a local HTTP server, generating a one-time setup key, and passing the paste-ready connection instructions to the remote agent.

Can I use this to let a Cursor agent inspect a live browser session from Claude Code?

Yes, this works with multi-agent setups like Claude Code and Cursor, allowing a remote agent to inspect the same live browser session and compare DOM state without switching contexts.

What access controls apply when a remote agent connects to my browser?

Remote agents get isolated tabs with read+write access by default, and you can grant admin access on request, ensuring scoped permissions for inspection or collaborative browsing.

How does the setup key exchange mechanism work for AI pairing?

The setup key exchange generates an expiring key via a one-command setup, which authenticates the remote AI agent's session token to securely connect to your local browser instance.

Do I need a local HTTP server to enable browser sharing between agents?

Yes, a local HTTP server is required to establish the connection, enforce session token authentication, and maintain tab isolation for secure browser sharing between agents.

What is the best way to debug a UI issue collaboratively using two different AI agents?

The best way is to use a one-command setup to generate a setup key, granting the second agent scoped access to inspect the live browser session and compare DOM state for debugging.