GitHub opens its platform to third-party AI agents with “Agent HQ”

GitHub has taken a significant step forward in integrating artificial intelligence into software development. The platform has announced the launch of ‘Agent HQ’, a centralized hub that allows developers to access multiple third-party coding AI agents — going beyond just GitHub Copilot.

A control center for multiple programming AIs

With Agent HQ, users will soon be able to invoke OpenAI Codex, Anthropic Claude, Google Jules, xAI, and Cognition Devin — all directly from GitHub, without leaving their working environment.

Subscribers to GitHub Copilot will have access to a “Mission Control” dashboard, designed as a command center to manage and monitor multiple AI agents simultaneously.

Parallel AI agents and an intelligent work plan

Another major innovation: Agent HQ will allow multiple AI agents to execute in parallel on the same task. Developers can then compare results and select the most relevant or elegant solution.

Before the official launch, OpenAI Codex is already available to Copilot Pro Plus subscribers enrolled in the VS Code Insiders program.

New tools integrated into VS Code

GitHub is also introducing a “Plan Mode” in VS Code, a feature that utilizes Copilot to create a step-by-step action plan before executing code with an AI agent. A new code review stage is also being introduced: Copilot will now use tools like CodeQL to analyze and validate the code before it is submitted to the developer.

With Agent HQ, GitHub is transforming its ecosystem into a true orchestration platform for AI agents, where multiple intelligences can collaborate or compete to produce the best possible code. This evolution clearly illustrates the current trend: moving from a single assistant to a team of specialized AIs working alongside developers.