This update brings support for enterprise-managed sandbox policies, cross-file cursor jumps for next edit suggestions, global project context in chat, enterprise policy diagnostics, and a new connection between terminal Copilot CLI sessions and JetBrains IDEs. It also improves model selection, the chat experience, and reliability across MCP servers and agent sessions.
## [What’s new](#whats-new)
### [Enterprise-managed sandbox policies in public preview](#enterprise-managed-sandbox-policies-in-public-preview)
Enterprise administrators can now centrally configure sandbox behavior for GitHub Copilot in JetBrains IDEs. Managed policies can control sandbox enablement, filesystem and network access, proxy settings, developer-tool access, macOS Keychain access, and more.
Managed restrictions take precedence over user settings. Copilot locks affected controls in the IDE and identifies settings managed by your organization, helping administrators enforce consistent development environment boundaries.
The sandbox settings under GitHub Copilot > Sandbox are visible when your organization enables the Editor Preview feature flag or configures a managed setting to enable or disable the sandbox. If neither condition applies, the sandbox settings do not appear.
For more information, see [Configuring local sandbox settings for GitHub Copilot](https://docs.github.com/copilot/how-tos/cloud-and-local-sandboxes/configuring-local-sandbox-settings).
### [Cross-file cursor jumps in next edit suggestions](#cross-file-cursor-jumps-in-next-edit-suggestions)
Next edit suggestions can now move your cursor across files, helping you navigate and apply coordinated code suggestions throughout a project. When a suggested change continues in another file, you can jump directly to the relevant location instead of finding it manually.
Global project context in chat
You can now add global files and folders to chat context, making it faster to include information that applies across your project. This reduces repetitive context setup and helps Copilot provide more relevant responses for changes that span multiple areas.
Enterprise policy diagnostics
You can now use enterprise policy diagnostics to verify that policies are correctly detected and enforced on your device. This makes it easier to confirm that your Copilot configuration follows your organization’s requirements.
Connect terminal Copilot CLI sessions to your IDE
You can now use /ide in GitHub Copilot CLI to connect a terminal session to your JetBrains IDE context, including selections, diagnostics, and file references. This integration is in public preview and helps terminal-based workflows stay grounded in what you are viewing and editing in the IDE.
Copilot shell commands can also use environment variables from the IDE terminal and the Python interpreter configured for your project. They can activate the project’s standard local Python virtual environment, making connected command execution more consistent with your development environment.
User experience enhancements
This release includes several interaction and navigation improvements for common workflows:
- Background agent prompts : Improved the ask-user card to better accommodate long questions.
- Model selection : Refined model picker behavior and BYOK grouping to make model choices easier to scan and select.
- Subagent models : Added support for selecting the session model used by built-in subagents in the Copilot agent harness.
- Agent debug logs : Added section copy support so you can share diagnostics and troubleshooting context faster.
- MCP configuration : Opened MCP configuration in the originating project window for better continuity.
- Plugin updates : Improved update reminders to make new plugin versions easier to discover.
Quality improvements
This release improves reliability across MCP servers and agent sessions, including BYOK provider persistence, OAuth challenge continuity, restored server state, and GitHub Enterprise authentication. It also resolves blank local chat sessions, incorrect working-set counts, working-set zoom and model picker interaction issues, and read-only file editing issues. Additional fixes address Claude plan stop behavior, misleading diagnostics for unopened files, stale global instructions in “Agent Customizations,” and chats becoming unresponsive after automatic compaction.
Generally Available #
OpenTelemetry settings in “GitHub Copilot – Chat” are now generally available to all users.
Try it out #
We encourage you to try out the latest version of the GitHub Copilot plugin and share your feedback. Your input is invaluable in helping us refine and improve the product.
Share your feedback #
We’d love to hear about your experience in the following channels:
- In-product feedback: Use the feedback options within your IDE.
- Feedback repository: Share your thoughts in the GitHub Copilot for JetBrains IDEs issues .