winget install GitHub.Copilotbrew install copilot-clinpm install -g @github/copilotgh copilotcopilotcopilot --continue to pick up exactly where you left off in your most recent Copilot CLI session.copilot --continuecopilot --resume, you can view a list of past sessions and select which one to load, maintaining context across multiple projects.copilot --resume/model

copilot help [topic]/help command provides a list of interactive shortcuts and in-app commands. /help--yolo: Extreme AutonomyWhen using —yolo, Copilot CLI operates without any safety nets. It grants all necessary permissions to execute commands, modify files, and access URLs without requiring explicit user approval. This hands complete control to the AI.
copilot --yolo/plan [prompt]Press Shift + Tab to cycle in and out of plan mode.
/review [prompt]/diff/login/logout

Can I use this with any code editor?Copilot CLI operates independently in your terminal and can modify files that any editor can then display, making it useful regardless of your editor choice, especially for editors that don't have official Copilot extensions.
Yes! Copilot CLI supports Model Context Protocol (MCP) server integrations, allowing you to add custom capabilities and contextual richness tailored to your unique development environment.This extensibility means you can integrate your own tools, APIs, and workflows directly into the Copilot experience.
Direct integration into VSCode is not available in the current public preview release, but it is being actively considered in the roadmap for future releases.
Every file change and command execution requires your explicit approval before being applied. You maintain complete visibility and control over all autonomous actions that Copilot CLI suggests.