Paul Krill
Editor at Large

GitHub Copilot previews agent mode

news
Feb 6, 20253 mins

New agent mode can iterate on its own code, infer tasks not specifed, recognize errors, and fix them automatically, GitHub said.

Credit: Shutterstock / kung_tom

GitHub has added new capabilities to its GitHub Copilot AI-powered coding assistant, including a preview of a new agent mode capable of iterating on its own code. The company also announced the general availability of Copilot Edits in Visual Studio Code.

The updates, announced February 6, serve as an evolution of GitHub Copilot from an AI pair programmer to an agentic peer programmer, as agents become increasingly integral to software development, according to GitHub.

Copilot Edits allows developers to specify a set of files to be edited and prompt Copilot in natural language to immediately make inline changes across multiple files. Agent mode enables Copilot to iterate on its own output as well as the results of that output to complete a user request. Agent mode can recognize and fix its own errors, suggest terminal commands, and analyze runtime errors with self-healing capabilities, GitHub said.

To access agent mode, developers need to download VS Code Insiders and then enable the agent mode setting for GitHub Copilot Chat.

Other features in preview include:

  • New models from industry leaders including Google Gemini 2.0 Flash and OpenAI o3-mini available in Copilot Chat. Administrators have organization-wide access control.
  • Next edit suggestions to accelerate code changes by identifying and proposing the next edit based on the context of previous changes. By pressing tab, users can implement suggestions throughout an open file with insertions, deletions, and replacements.
  • Prompt files that allow users to store and share reusable prompt instructions in their VS Code workspace. These โ€œblueprintsโ€ include self-contained markdown files that blend natural language guidance, file references, and linked snippets to โ€œsuperchargeโ€ coding tasks.
  • Vision for Copilot, a feature that allows users to generate a UI, alt text, and code by feeding Copilot a screen, image, or snip.

The company also unveiled โ€œProject Padawan,โ€ which involves plans for autonomous SWE agents on GitHub that will independently handle entire tasks at the developerโ€™s direction. This effort represents a future where developers can assign GitHub issues to GitHub Copilot, let AI complete the task autonomously, and then review the work.

GitHub also announced provisioning and authentication support for GitHub Copilot Workspace for Enterprise Managed Users. This allows organizations to configure and control Workspace access securely, GitHub said.

Paul Krill

Paul Krill is editor at large at InfoWorld. Paul has been covering computer technology as a news and feature reporter for more than 35 years, including 30 years at InfoWorld. He has specialized in coverage of software development tools and technologies since the 1990s, and he continues to lead InfoWorldโ€™s news coverage of software development platforms including Java and .NET and programming languages including JavaScript, TypeScript, PHP, Python, Ruby, Rust, and Go. Long trusted as a reporter who prioritizes accuracy, integrity, and the best interests of readers, Paul is sought out by technology companies and industry organizations who want to reach InfoWorldโ€™s audience of software developers and other information technology professionals. Paul has won a โ€œBest Technology News Coverageโ€ award from IDG.

More from this author