GitHub Agent HQ: How to Run Claude, Codex, and Copilot Agents Side by Side

What Is Agent HQ?

GitHub Agent HQ is a new feature for Copilot Pro+ and Enterprise subscribers that lets you run multiple AI coding agents in parallel. Instead of being locked into one AI assistant, you choose between Claude Code, OpenAI Codex, and GitHub Copilot for each task.

Why This Matters

Different agents excel at different tasks:

Claude Code: Best for complex planning, large codebase navigation, code review, and debugging across many files
OpenAI Codex: Fastest for scaffolding, rapid prototyping, and generating boilerplate code
GitHub Copilot: Strong for inline completions, pull request reviews, and issue-to-code workflows

How to Set It Up

Ensure you have a Copilot Pro+ or Enterprise subscription
Open Agent HQ in GitHub, GitHub Mobile, or VS Code
Configure per-agent settings and preferences
Start assigning tasks to different agents

Multi-Agent Workflow Patterns

Pattern 1: Parallel Development

Run Claude on architecture and Codex on implementation simultaneously. Review both outputs and merge the best approaches.

Pattern 2: Specialized Review

Use Copilot for initial code generation, then have Claude review the output for quality, security, and architectural consistency.

Pattern 3: Custom Agents

Build team-specific agents for your workflows. Agent HQ supports custom agents that understand your codebase conventions and standards.

Pricing and Availability

Agent HQ is included with Copilot Pro+ ($39/mo) and Enterprise subscriptions. GitHub plans to expand access to other tiers soon. BYOK support means you can also bring your own API keys for AWS Bedrock and Google AI Studio.

The Bigger Picture

Agent HQ makes GitHub the central orchestration platform for AI-assisted development. Rather than switching between tools, developers stay in one environment while leveraging the strengths of each AI provider.