GitHub has announced Agent HQ, a vision to evolve its platform by integrating coding agents from Anthropic, OpenAI, Google, Cognition, and xAI within the developer workflow.
Agent HQ is designed to create an open ecosystem where developers can orchestrate multiple agents in parallel, without the need to move away from established developer tools and practices. Coding agents will be available directly within GitHub as part of the paid Copilot subscription.
According to GitHub, its user base has grown to 180 million developers, increasing at its fastest rate to date, with a new developer joining every second. The company states that 80% of new developers use Copilot in their first week, highlighting the growing reliance on AI-assisted development. GitHub aims to integrate AI as an essential aspect of the development process, rather than an add-on, ensuring that the technology works in tandem with users' established workflows.
Agent HQ introduces a suite of capabilities, beginning with "mission control"-a command centre designed to assign, monitor and approve the work of multiple agents from anywhere. This is available across GitHub, VS Code, mobile, and the command line interface (CLI). New features include deeper integrations within VS Code, such as Plan Mode, AGENTS.md files for custom agent behaviour, and access to the GitHub MCP Registry for discovering and enabling specialist MCP servers.
The platform's new mission control is aimed at providing oversight and control for developers as they work with specialist agents. Developers can view, assign, and track agent activities across devices, with an interface that spans existing development environments.
New controls have also been added, such as granular oversight for running continuous integration (CI) and checks for agent-generated code, identity features for managing agent permissions, and improved tools for handling merge conflicts and code navigation. Expanded integrations enable developers to connect mission control with Slack and Linear, in addition to previously supported services like Atlassian Jira, Microsoft Teams, Azure Boards, and Raycast.
In addition, Plan Mode in VS Code allows developers to build task-specific plans in collaboration with Copilot, documenting requirements and clarifying project context before coding begins. Custom agents can be configured using AGENTS.md files, allowing teams to tailor Copilot behaviour to their specific needs without repeatedly prompting the system.
Through the GitHub MCP Registry in VS Code, developers can discover, install and activate MCP servers from third parties such as Stripe, Figma, and Sentry. This feature supports creating custom agents with defined system prompts and tools, extending the flexibility of AI-assisted coding within GitHub's ecosystem.
There is also a focus on code quality and governance. The GitHub Code Quality feature, now in public preview, provides visibility and reporting on code maintainability, reliability, and testing across all repositories within an organisation. This feature extends Copilot's security checks beyond vulnerabilities to include maintainability and reliability considerations for new code. Copilot's AI agents will now perform an initial review of code before it is surfaced to developers.
To give organisations insights into Copilot utility, GitHub has released a public preview of the Copilot metrics dashboard, providing data on usage and impact across entire organisations. For enterprise administrators, GitHub has introduced a control plane-a governance layer for AI agents-enabling configuration of security policies, access management, and audit logging. Administrators can control which agents are allowed, model access, and obtain usage metrics for Copilot across their teams.
"We built Agent HQ because we're developers, too. We know what it's like when it feels like your tools are fighting you instead of helping you. When 'AI-powered' ends up meaning more context-switching, more babysitting, more subscriptions, and more time explaining what you need to get the value you were promised. That ends today," said Kyle Daigle, COO, GitHub.
Daigle added: "Agent HQ isn't about the hype of AI. It's about the reality of shipping code. It's about bringing order and governance to this new era without compromising choice. It's about giving you the power to build faster, with more confidence, and on your terms."
"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 millions more developers who use GitHub and VS Code, extending the power of Codex everywhere code gets written," said Alexander Embiricos, Codex Product Lead, OpenAI.
"We're partnering with GitHub to bring Claude even closer to how teams build software. With Agent HQ, Claude can pick up issues, create branches, commit code, and respond to pull requests, working alongside your team like any other collaborator. This is how we think the future of development works: agents and developers building together, on the infrastructure you already trust," said Mike Krieger, Chief Product Officer, Anthropic.
All features introduced as part of Agent HQ will become available to developers through their existing paid Copilot subscriptions. Integration with external agents and specialist servers is aimed at providing greater flexibility and capabilities as part of standard developer toolkits on GitHub and VS Code.
