GitHub Copilot推出全新编码助手，旨在提升开发者的工作效率。该助手直接嵌入GitHub，通过GitHub Actions驱动安全且可定制的开发环境，在开发者将GitHub issue分配给Copilot或在VS Code中提示它时启动。助手将提交推送到草稿pull request，并通过会话日志跟踪每一步。它擅长处理低到中等复杂度的任务，例如添加功能、修复错误、扩展测试、重构代码和改进文档。通过Model Context Protocol (MCP)，编码助手可以访问GitHub之外的数据和功能，利用视觉模型处理issue中的图像。Copilot完成后，会通知开发者进行审查，并根据反馈自动修改代码。

👀GitHub Copilot编码助手通过将GitHub issue分配给Copilot来运行，它会自动添加一个“眼睛”的emoji表情符号，并在后台启动工作，包括启动虚拟机、克隆存储库、配置环境，并使用GitHub代码搜索支持的增强生成（RAG）分析代码库。

🚀该助手使用最先进的模型，擅长处理经过充分测试的代码库中的低到中等复杂度的任务，例如添加功能、修复错误、扩展测试、重构代码和改进文档，从而解放开发者，专注于更有价值的工作。

🛡️GitHub Copilot编码助手的设计从一开始就考虑到项目的安全性，并确保其工作在投入生产之前得到审查。默认情况下，助手只能推送到它创建的分支，并且要求开发者不能批准自己要求助手打开的pull request，从而保障代码安全。

💡借助模型上下文协议（MCP），开发者可以授予编码助手访问GitHub外部数据和功能权限。MCP服务器可以在存储库的设置中进行配置，所有GitHub数据都可以从官方GitHub MCP服务器中提取。此外，助手还可以查看GitHub issue中包含的图像，从而更好地理解bug截图或新功能模型。

✅Copilot编码助手现已向所有Copilot企业版和Copilot Pro+客户开放。只需在希望使用它的存储库中启用该代理，如果是Copilot企业版用户，管理员还需要开启策略。

We are excited to introduce a new coding agent for GitHub Copilot. Embedded directly into GitHub, the agent starts its work when you assign a GitHub issue to Copilot or prompt it in VS Code. The agent spins up a secure and fully customizable development environment powered by GitHub Actions. 

As the agent works, it pushes commits to a draft pull request, and you can track it every step of the way through the agent session logs. Having Copilot on your team doesn’t mean weakening your security posture – existing policies like branch protections still apply in exactly the way you’d expect. Plus, the agent’s pull requests require human approval before any CI/CD workflows are run, creating an extra protection control for the build and deployment environment. 

Using state-of-the-art models, the agent excels at low-to-medium complexity tasks in well-tested codebases, from adding features and fixing bugs to extending tests, refactoring code, and improving documentation. You can hand off the time-consuming, but boring tasks to Copilot that will use pull requests, CI/CD, and all of your existing tooling while you focus on the interesting work.

A detailed look

To run the new coding agent, assign one or more GitHub issues to Copilot. It’s as easy as that. You can do this on github.com, in GitHub Mobile, or through the GitHub CLI, just as you would assign the same issue to one of your team members or yourself. You can also ask Copilot to open a pull request from Copilot Chat on GitHub or right in VS Code like this:

> @github Open a pull request to refactor this query generator into its own class

Once an issue is assigned to it, the agent adds an 👀 emoji reaction and starts its work in the background. It boots a virtual machine, clones the repository, configures the environment, and analyzes the codebase with advanced retrieval augmented generation (RAG) powered by GitHub code search. As the agent works, it regularly pushes its changes to a draft pull request as git commits and updates the pull request’s description. Along the way, you’ll see the agent’s reasoning and validation steps in the session logs, making it easy to trace decisions and spot issues. 

With the power of Model Context Protocol (MCP), you can give the coding agent access to data and capabilities from outside of GitHub. MCP servers can be configured in the repository’s settings. And, of course, all your GitHub data can be pulled in from the official GitHub MCP Server. And the agent isn’t limited to just text – thanks to the power of vision models, it can see images included in GitHub issues you assign to it, so you can share screenshots of a bug or mockups of what your new feature should look like.

Once Copilot is done, it’ll tag you for review and you can leave comments asking for it to make changes. It will pick those comments up automatically and propose code changes. The agent also incorporates context from related issue or PR discussions and follows any custom repository instructions, allowing it to understand both the intent behind the task and the coding standards of the project. 

The Copilot coding agent is opening up doors for human developers to have their own agent-driven team, all working in parallel to amplify their work. We’re now able to assign tasks that would typically detract from deeper, more complex work—allowing developers to focus on high-value coding tasks.

James Zabinski, DevEx Lead at EY

Integrated, configurable, and secure

All SWE agents need a compute environment to do their work. For agent mode in VS Code, it’s the CPU of your PC or Mac, or a remote development container like GitHub Codespaces. When we started the work on Project Padawan, there was only one answer to which cloud compute we should use: GitHub Actions. Introduced in 2018, Actions is the largest CI/CD ecosystem in the world with more than 25,000 actions in the GitHub Marketplace. Every weekday, GitHub-hosted and self-hosted runners execute more than 40 million daily jobs. Open source projects, startups, and large enterprises all rely on GitHub Actions to run their build, test, and deploy workloads.

Our coding agent is designed from the ground up to keep your project secure and ensures that its work gets a review before it goes to production. The following policies are applied by default:

The agent can only push to branches it created, keeping your default branch and the ones your team created safe and secureThe developer who asks the agent to open a pull request cannot be the one to approve it – so any “required reviews” rule you have set up in your repository will be honored The agent’s internet access is tightly limited to a trusted list of destinations that you can customizeGitHub Actions workflows won’t run without your approval, giving you the chance to spot-check the agent’s code

Existing repository rulesets and organization policies are considered as well.

The GitHub Copilot coding agent fits into our existing workflow and converts specifications to production code in minutes. This increases our velocity and enables our team to channel their energy toward higher-level creative work.

Alex Devkar, Senior Vice President, Engineering and Analytics, Carvana

Get started today

The new coding agent is available to all Copilot Enterprise and Copilot Pro+ customers. All you need to get started is enabling the agent in the repositories where you want to use it, and if you’re a Copilot Enterprise user, an administrator will need to turn on the policy too. Additionally, you can now activate agent mode in more IDEs, including Xcode, Eclipse, Jetbrains, and Visual Studio.

Beginning June 4, 2025, Copilot coding agent will use one premium request per model request the agent makes. 

Whether it’s code completions, next edit suggestions, chat, agent mode, or now coding agent, GitHub Copilot has always had one mission: To keep you in the magical flow state. To do the things you don’t want to do, so you have more time for the things you do. To find the creative spark that brought you to this industry in the first place. Happy coding!

The post GitHub Copilot: Meet the new coding agent appeared first on The GitHub Blog.