AI-Powered Development: GitHub Copilot in Visual Studio Code

Course details

What is this course about?
Explore how AI is revolutionizing software development within Visual Studio Code (VS Code) and learn how to use GitHub Copilot to generate code, guide your coding with chat prompts, and perform agent-based tasks for streamlined development. Discover how to customize VS Code with AI extensions, keybindings, and snippets to optimize your workflow. Find out how to optimize writing unit tests, managing code completions, and running security scanning to ensure code reliability and safety. Navigate through setting agent permissions for chat tools and understand how to apply AI responsibly in your projects.
Upon completion of the course, you will have developed a robust framework to integrate AI seamlessly into your daily coding activities. This intermediate-level course is suited for developers looking to enhance workflows by leveraging AI-driven solutions to improve efficiency and provide creative approaches to coding. With guidance on practical applications, you can become equipped to critically evaluate AI outputs and maintain high quality in your projects.

Note: This course was created by LearnIt Labs. We are pleased to host this training in our library.

Objectives
What will I be able to do by the end of this course?
Identify the fundamental principles of AI-based code generation in Visual Studio Code.
Utilize GitHub Copilot to automate routine coding tasks, improving coding efficiency.
Apply AI-driven techniques for writing unit tests and conducting code security scans.
Evaluate AI-generated outputs for accuracy, licensing compliance, and style.
Customize AI tools and settings in Visual Studio Code to optimize workflow and productivity.

Audience
Who is this course for?
Intermediate developers
Software engineers
Development team leads
Technical project managers

Concepts

Introduction - Start Smart with AI in Visual Studio Code

AI-powered development in Visual Studio Code
How to access the sample code on GitHub

Writing Code with AI

How AI tools support developers
AI, LLMs, and Copilot explained
Non generative AI - The other option
Generate a working project with Copilot agents

Get Ready to Use Copilot

GitHub Copilot - The backbone of AI in the IDE
Subscription options
Activate Copilot in Visual Studio Code
Verify Copilot is activated
Choose an Copilot AI model

Completions - Smarter Code, Faster Flow

Completions, chat, and agent modes
Code competion - JavaScript
Making sense of nondeterministic predictions
Code completion - Python
HTML and CSS completion
More inline suggestions features
Use comments to guide completion

Interactive Chat - The Ask Agent

Use chat to ask code specific questions
Use chat to explore a project
Explore the Word Frequency Analyzer page
Explain a section of code
Guidance on potential issues
Get refactoring suggestions
Have Copilot draft an editing prompt
Review the hidden reasoning workflow
Inline chat - Exploring the code editor

Chat Window Features

Create new chats
Chat panel vs. agent sessions
Manage chat threads and history
Rename a chat thread
Save a chat thread
Delete a chat thread
Customize chat settings
Use voice and audio

Interactive Chat - The Edit Agent

Edit agent explained
The edit agent
Find the proposed changes
The three Copilot diff tools
Track and manage pending changes
Keep or undo the changes
Revert workspace state with checkpoints

Interactive Chat - The Multistep Agent

Quick look - Edit vs. multistep agent
Evaluate and improve Python code
Verify config file reading
Implement a print feature
Add a security feature to the app

Use Agent Mode to Refactor a Web Project

Prompt engineering for the agent
Start an agent-assisted refactor
Add blocked words to the UI
Add additional files
Clean up invalid charactors
Modernize the UI layer
Refactor review - Outcomes and impact

Set Chat Context

Implicit context - What Copilot infers
Default context - What Copilot sees
Explicit context - Narrow or expand scope

Broaden Scope with Participants

Explore the participants
Use the @vscode participant
Use the @github participant

Chat Tools and Actions

Actions overview
Preview the chat tools
Set agent permissions for chat tools
Reference web content with #fetch

Greenfield AI

Prototype a Greenfield project
Tour the graphic generation site
Import a saved chat conversation
How to prompt a graphic generation site
Examine the generated code and UI
Tour the Pedal and Pines site
Review refactored site sections
Review updated code and UI
Prototype setbacks and frustrations
Postmortems and prompt refinement

Unit Tests

The unit test project
Add test framework with SetupTests
Write unit tests with test
Inspect generated code
Test for raised errors
Data-driven tests

Use Images in Chat Prompts

Use images in chat
Generate CSS based on mock-up
Generate components based on a sketch
Create a class library from a diagram

Copilot in the Terminal

Terminal helpers overview
The three helpers
Use the three terminal helpers
Generate terminal commands
Fix terminal errors
Get tooling help
Use inline chat

Interactive Chat - The Plan Agent

Understand cloud and background agents
Plan mode in Copilot - Quick review
Session panel
Create a plan
Save plan as markdown
Run the plan on background agent
Use terminal with background agent
The Git worktree
Review the changes
Merge the changes

Git Panel Integration

Commit messages
Commit message customization
Write pull request descriptions
Pull request code review
Approve code review steps

Work with MCP Servers

MCP overview
Query a database with an MCP server
Generate TypeScript DB access code
Understand mcp.json setting
Enable catalog browsing in the Extensions panel
Install other MCP servers with the Extension panel
Find MCP servers on GitHub

Conclusion - Continue Your Journey

Custom instructions and skills
Next steps and learning resources