Pick your agent: Use Claude and Codex on Agent HQ 

Context switching equals friction in software development. Today, we’re removing some of that friction with the latest updates to Agent HQ which lets you run coding agents from multiple providers directly inside GitHub and your editor, keeping context, history, and review attached to your work.

Copilot Pro+ and Copilot Enterprise users can now run multiple coding agents directly inside GitHub, GitHub Mobile, and Visual Studio Code (with Copilot CLI support coming soon). That means you can use agents like GitHub Copilot, Claude by Anthropic, and OpenAI Codex (both in public preview) today.

With Codex, Claude, and Copilot in Agent HQ, you can move from idea to implementation using different agents for different steps without switching tools or losing context. 

We’re bringing Claude into GitHub to meet developers where they are. With Agent HQ, Claude can commit code and comment on pull requests, enabling teams to iterate and ship faster and with more confidence. Our goal is to give developers the reasoning power they need, right where they need it.

Katelyn Lesse, Head of Platform, Anthropic

How to get started 

When you have a task, you can choose Copilot, Claude, Codex, or a custom agent. You can make them work asynchronously and then review the results on your own time. 

Each coding agent session consumes one premium request. You can currently start Claude and Codex sessions from github.com, GitHub Mobile, and in VS Code. Claude and Codex must be explicitly enabled in your settings before use. See details on how. 

Create and view agent sessions on GitHub 

You can start and manage agent sessions directly from GitHub: 

• Open the Agents tab in any repository where agents are enabled. 
• Enter your request and select an agent using the Copilot icon. 
• Submit the request to start the session. 

Agents run asynchronously by default. You can follow their progress in real time or review completed sessions later, with detailed logs showing what the agent did and why. 

Each session produces artifacts like comments, drafts, and proposed changes that can be reviewed like any other contribution. 

Assign agents to issues and pull requests 

You can also use agents directly in your existing collaboration workflows: 

• Assign an issue to Copilot, Claude, Codex, or all three to compare results. 
• Agents can submit draft pull requests for review. 
• Assign agents to existing pull requests to request changes or further analysis. 
• Mention @Copilot, @Claude, or @Codex in pull request comments to prompt follow-up work. 

Agent activity is logged and reviewable, so their output fits naturally into the same workflows you already use to evaluate developer contributions. 

Agents can still make mistakes. That’s exactly why their output is designed to be reviewed, compared, and challenged, not blindly accepted. 

Start agent sessions in Visual Studio Code 

You can also work with agents directly from VS Code (version 1.109 or later): 

1. Open the Agent sessions view from the title bar or by pressing Ctrl+Shift+P /CMD+Shift+P and searching for “agent sessions”  
2. Choose a session type and your agent: 
   • Local (Copilot), Claude, or Codex for fast, interactive assistance 
   • Cloud for autonomous tasks that run on GitHub, then select an agent 
   • Background for asynchronous local work (Copilot only) 

This lets you explore ideas in the editor and then delegate longer-running work to GitHub without losing context or history.

From faster code to better decisions 

Agent HQ also lets you compare how different agents approach the same problem, too. You can assign multiple agents to a task, and see how Copilot, Claude, and Codex reason about tradeoffs and arrive at different solutions.  

In practice, this helps you surface issues earlier by using agents for different kinds of review:  

• Architectural guardrails: Ask one or more agents to evaluate modularity and coupling, helping identify changes that could introduce unintended side effects. 
• Logical pressure testing: Use another agent to hunt for edge cases, async pitfalls, or scale assumptions that could cause problems in production. 
• Pragmatic implementation: Have a separate agent propose the smallest, backward-compatible change to keep the blast radius of a refactor low.

This method of working moves your reviews and thinking to strategy over syntax. 

Our collaboration with GitHub has always pushed the frontier of how developers build software. The first Codex model helped power Copilot and inspired a new generation of AI-assisted coding. We share GitHub’s vision of meeting developers wherever they work, and we’re excited to bring Codex to GitHub and VS Code. Codex helps engineers work faster and with greater confidence—and with this integration, millions more developers can now use it directly in their primary workspace, extending the power of Codex everywhere code gets written.

Alexander Embiricos, OpenAI 

Why running agents on GitHub matters 

GitHub is already where code lives, collaboration happens, and decisions are reviewed, governed, and shipped. 

Making coding agents native to that workflow, rather than external tools, makes them even more useful at scale. Instead of copying and pasting context between tools, documents, and threads, all discussion and proposed changes stay attached to the repository itself. 

With Copilot, Claude, and Codex working directly in GitHub and VS Code, you can: 

• Explore tradeoffs early: Run agents in parallel to surface competing approaches and edge cases before code hardens. 
• Keep context attached to the work: Agents operate inside your repository, issues, and pull requests instead of starting from stateless prompts. 
• Avoid new review processes: Agent-generated changes show up as draft pull requests and comments, reviewed the same way you’d review a teammate’s work. 

There are no new dashboards to learn, and no separate AI workflows to manage. Everything runs inside the environments you already use. 

Built for teams, not just individuals 

These workflows don’t just benefit individual developers. Agent HQ gives you org-wide visibility and systematic control over how AI interacts with your codebase: 

• Agent controls: Manage access and security policies in one place, allowing enterprise admins to define which agents and models are permitted across the organization. 
• Code quality checks: GitHub Code Quality (in public preview) extends Copilot’s security checks to evaluate the maintainability and reliability impact of changed code, helping ensure “LGTM” reflects long-term code health. 
• Automated first-pass review: We have integrated a code review step directly into the Copilot’s workflow, allowing Copilot to address initial problems before a developer ever sees the code. 
• Impact metrics: Use the Copilot metrics dashboard (in public preview) to track usage and impact across your entire organization, providing clear traceability for agent-generated work. 
• Security and auditability: Maintain full control with audit logging and enterprise-grade access management, ensuring agents work with your security posture instead of against it. 

This allows teams to adopt agent-based workflows without sacrificing code quality, accountability, or trust. 

More agents coming soon 

Access to Claude and Codex will soon expand to more Copilot subscription types. In the meantime, we’re actively working with partners, including Google, Cognition, and xAI to bring more specialized agents into GitHub, VS Code, and Copilot CLI workflows. 

Read the docs to get started >