VS Code has evolved into a rich "AI workbench," where GitHub Copilot can do far more than autocomplete. In this course, LinkedIn Learning Principal Instructor Morten Rand-Hendriksen guides you through the most useful (and often overlooked) Copilot integrations inside VS Code—how to shape Copilot behavior with agent instructions and custom instructions, how to use slash commands and promptable actions for common dev tasks, and how to extend Copilot with tools via Model Context Protocol (MCP) servers. You'll also explore building and using custom agents and skills/scripts/templates to create repeatable, team-friendly workflows—so Copilot becomes a dependable collaborator, not just a chat box.

Learning objectives

• Identify the major Copilot "integration surfaces" in VS Code (agents, tools, workspace context, and command-style interactions) and when to use each.
  
• Apply agent instructions and repository/custom instructions to make Copilot more consistent, project-aware, and aligned with team conventions.
  
• Use slash commands and VS Code Copilot features to accelerate common tasks like explaining, fixing, documenting, optimizing, and generating tests.
  
• Understand MCP at a practical level—what it is, what it unlocks, and how MCP servers provide tools to connect Copilot to external systems and services through a standard interface.
  
• Create or adopt custom agents in VS Code to support specialized workflows (planner/reviewer/implementer-style roles and handoffs).
  
• Organize reusable workflows using skills/scripts/templates so Copilot can load the right instructions and resources at the right time.
  
• Navigate practical constraints (security prompts, trust boundaries, and org policy limitations) when enabling tool integrations.
  

Exercise Files on GitHub

Homepage