is now available! Read about the new features and fixes from November.

GitHub Copilot in VS Code cheat sheet

GitHub Copilot in Visual Studio Code provides AI-powered features to help you write code faster and with less effort. This cheat sheet provides a quick overview of the features for GitHub Copilot in Visual Studio Code.

You can access GitHub Copilot in VS Code through the Chat view, directly in the editor, from the integrated terminal, and via AI-powered enhancements in the VS Code user interface.

The team is continuously working on improving Copilot in VS Code and adding new features. Items in the cheat sheet marked with are experimental features. Try them out and share your feedback in our issues.

Chat with GitHub Copilot

Use natural language to chat with GitHub Copilot and get help with coding tasks. For example, ask Copilot to explain a block of code or a programming concept. Get more information about using Copilot Chat.

Action Description
⌃⌘I (Windows, Linux Ctrl+Alt+I) Open the Chat view and start a chat conversation with Copilot by using natural language.
⇧⌘I (Windows Ctrl+Shift+I, Linux Ctrl+Shift+Alt+I) Open the Copilot Edits view and start a code editing session across multiple files.
⇧⌥⌘L (Windows, Linux Ctrl+Shift+Alt+L) Open Quick Chat and ask a quick question to Copilot.
⌘I (Windows, Linux Ctrl+I) Start Inline Chat to send a chat request to Copilot directly from the editor. Use natural language or use / commands to give instructions to Copilot.
Type @ in chat or select to view the list of chat participants, which are domain experts that can help you in a specific area. Extensions can also contribute additional participants.
Example: @workspace how is auth implemented?
Participant detection Copilot Chat can also automatically route your question to the appropriate participant. Get more info.
/ Invoke a slash command to prompt for commonly used actions, such as explaining a block of code, generating tests or documentation.
/explain Ask Copilot to explain a block of code or a programming concept.
Select this icon in the Chat view to access your history of chat sessions.
Enter a chat prompt by using speech (voice chat). The chat response is read out aloud.

Tips

  • Use / commands and @ participants to get more precise and relevant answers.
  • Be specific, keep it simple, and ask follow-up questions to get the best results.
  • Provide context by attaching files, symbols, or selections to your chat prompt.

Code editing session (Preview)

Use Copilot Edits to start a code editing session where you can iterate quickly on AI-generated code edits that are applied directly across multiple files in your workspace.

Action Description
⇧⌘I (Windows Ctrl+Shift+I, Linux Ctrl+Shift+Alt+I) Open the Copilot Edits view and start a code editing session across multiple files.
Start a new edit session.
Accept Accept all current edits.
Discard Discard all current edits.
View all edits in a multi-file diff editor.
Undo the last edit.
Redo the last edit.
Add Files... Attach files to working set.

Tips

  • Add all files for which you want to get edits to the working set.
  • Be specific and precise about the changes you want Copilot Edits to make.
  • If you have a larger task, decompose it in smaller tasks and iterate often.

Generate code from chat

Copilot can generate code blocks in response to your chat prompts. Quickly apply the generated code in your project or insert it in a new file. For example, ask Copilot to optimize an algorithm in your code.

Action Description
Smart-apply the generated code block in the active editor.
Insert the generated code block at the cursor.
Copy the generated code block to the clipboard.
Insert the generated code block in the terminal as a shell command.
Insert into New File Insert the generated code block in a new file.

Tips

Attach context to your prompt

When you send a chat prompt to Copilot, you can attach context to help Copilot understand your question better. For example, add the current editor selection, a file, or a symbol to your chat prompt. Get more information about best practices for using Copilot.

Action Description
(⌘/ (Windows, Linux Ctrl+/)) Open a Quick Pick to select relevant context for your chat prompt. Choose from workspace files, symbols, current editor selection or visible contents, terminal selection or last run command, or the VS Code API.
Enable/disable automatically attaching open editors as context.
#codebase Context variable: add relevant workspace content as context to your prompt.
#editor Context variable: add the visible contents of the active editor as context for your prompt.
#selection Context variable: add the current editor selection as context to your prompt.
#terminalSelection Context variable: add the current terminal selection as context to your chat prompt.
#terminalLastCommand Context variable: add the last run terminal command as context to your chat prompt.
#VSCodeAPI Context variable: add the VS Code API as context to your prompt to ask questions related to VS Code extension development.
#file Open a Quick Pick to select a file from your workspace and add it as context for your prompt.
#<filename> Type #, followed by a filename, to get filename suggestions for workspace files and attach as context.
#sym Open a Quick Pick to select a symbol from your workspace and add it as context for your prompt.
#<symbol> Type #, followed by a symbol name, to get symbol suggestions for workspace files and attach as context.
Drag & drop file Drag & drop a file or editor onto the chat to attach the file as context.
Recent files Automatically include recently opened and edited files in your chat prompt. Get more info.

Tips

  • Quickly add multiple files as context by using the right arrow key in the Quick Pick
  • Explicitly include relevant context by selecting code or mentioning chat variables
  • Review the used references in the chat response to ensure that the context is relevant

Copilot in the editor

As you're coding in the editor, you can use Copilot to generate code completions as you're typing. Invoke Inline Chat to ask questions and get help from Copilot, while staying in the flow of coding. For example, ask Copilot to generate unit tests for a function or method. Get more information about code completions and Inline Chat.

Action Description
Code completions Start typing in the editor and Copilot provides code suggestions that match your coding style and take your existing code into account.
Code comments Provide a code completion prompt to Copilot by writing instructions in a code comment.
Example: # write a calculator class with methods for add, subtract, and multiply. Use static methods.
⌘I (Windows, Linux Ctrl+I) Start Inline Chat to send a chat request to Copilot directly from the editor. Use natural language or use / commands to give instructions to Copilot.
Prompt from the editor Start typing natural language directly in code and Copilot detects that you're not writing code but prompting, and will automatically start Inline Chat for your prompt.
F2 Get AI-powered suggestions when renaming symbols in your code.

Tips

  • Use meaningful method or function names to get better code completions quicker.
  • Select a code block to scope your Inline Chat prompt or attach relevant context by attaching files or symbols.
  • Use the editor context menu options to access common Copilot actions directly from the editor.

Customize AI code generation

Copilot can generate responses that match the coding style, tools, and developer workflow of your team or project, provided it has the right context. You can provide custom instructions to help Copilot adjust to your preferences, so that you don't have to provide these details every time you make a Copilot request. These custom instructions are automatically added to your requests. Get more information about customizing Copilot in VS Code.

Action Description
File-based instructions (Preview) Define shared instructions for code generation in a .gitHub/copilot-instructions.md file in your workspace. These common instructions supplement your own personal code-generation instructions.
Code-review instructions (Preview) Define instructions for using Copilot to review an editor selection in settings or import from a file. You can define language-specific instructions.
Code-generation instructions Define instructions for code generation with GitHub Copilot in settings or import from a file. You can define language-specific instructions.
Test-generation instructions Define instructions for test generation with GitHub Copilot in settings or import from a file. You can define language-specific instructions.
Commit-message generation instructions Define instructions for commit message generation with GitHub Copilot in settings or import from a file. You can define language-specific instructions.

Tips

  • Define language-specific instructions to get more accurate generated code for each language.
  • Store your instructions in a file to easily share them with your team and across projects.

Review code (Preview)

Copilot can do a quick review pass of a code block or perform a review of uncommitted changes in your workspace. Review feedback shows up as comments in the editor, where you can apply the suggestions.

Action Description
Review and Comment (Preview) Select a block of code, and select Copilot > Review and Comment from the editor context menu for quick review pass.
Copilot Code Review Select the Copilot Code Review button in the Source Control view for a deeper review of all uncommitted changes. Join the waitlist.

Generate tests

Copilot can generate tests for functions and methods in your codebase. Get more information about slash commands in Chat.

Action Description
/tests Generate tests for all or only the selected methods and functions in the editor. The generated tests are appended in an existing tests file or a new tests file is created.
/setupTests Get help setting up a testing framework for your code. Get recommendation for a relevant testing framework, steps to set up and configure it, and suggestions for VS Code testing extensions.
/fixTestFailure Ask Copilot for suggestions on how to fix failing tests.
Test coverage Generate tests for functions and methods that are not yet covered by tests. Get more information.

Tips

  • Provide details about the testing frameworks or libraries to use.

Generate documentation

Generate code documentation for functions and methods in your codebase. Get more information about slash commands in Chat.

Action Description
/docs Generate documentation comments for all or only the selected methods and functions in the editor.

Debug and fix problems

Use Copilot to help fix coding problems and to get help with configuring and starting debugging sessions in VS Code.

Action Description
/fix Ask Copilot for suggestions on how to fix a block of code or how to resolve any compiler or linting errors in your code. For example, to help fix unresolved Node.js package names.
/fixTestFailure Ask Copilot for suggestions on how to fix failing tests.
/startDebugging Generate a launch.json debug configuration file and start a debugging session from the Chat view. Get more information.
copilot-debug command Terminal command to help you debug your programs. Prefix a run command to start a debugging session for it (for example, copilot-debug python foo.py). Get more information.

Tips

  • Provide additional information about the type of fix you need, such as optimizing the memory consumption or performance.
  • Watch for Copilot Code Actions in the editor that indicate suggestions for fixing problems in your code.

Scaffold a new project

Copilot can help you create a new project by generating a scaffold of the project structure, or generate a notebook based on your requirements.

Action Description
/new Use the /new command in the Chat view to scaffold a new project or a new file. Use natural language to describe the type of project/file you need, and preview the scaffolded content before creating it.
Example: /new Express app using typescript and svelte
/newNotebook Use the /newNotebook command in the Chat view to generate a new Jupyter notebook based on your requirements. Use natural language to describe what the notebook should contain.
Example: /newNotebook get census data and preview key insights with Seaborn.

Source control and issues

Copilot can analyze the changes in your commits and pull requests and provide suggestions for commit messages and pull request descriptions.

Action Description
Commit Generate a commit message for the current changes in a source control commit.
Pull request Generate a pull request title and description that correspond with the changes in your pull request.
@github Use the @github participant in chat to ask about issues, pull requests, and more across your repositories. Get more information about the available GitHub skills.
Example: @github What are all of the open PRs assigned to me?, @github Show me the recent merged pr's from @dancing-mona

Search

Use Copilot to get more relevant search results in the Search view.

Action Description
Semantic search Include search results from Copilot in the Search view that are semantically relevant.

Terminal

Get help about shell commands and how to resolve errors when running commands in the terminal.

Action Description
⌘I (Windows, Linux Ctrl+I) Start Inline Chat within the terminal to use natural language to quickly get and run a shell command.
Example: how many cores on this machine?

Fix using Copilot		 Select the icon on a failed shell command and get suggestions on how to resolve the error.

Explain using Copilot		 Select the icon on a failed shell command and get an explanation about why the command failed.
@terminal Use the @terminal participant in the Chat view to ask questions about the integrated terminal or shell commands.
Example: @terminal list the 5 largest files in this workspace
@terminal /explain Use the /explain command in the Chat view to explain something from the terminal.
Example: @terminal /explain top shell command

Python and Notebook support

You can use Copilot Chat to help you with Python programming tasks in the Native Python REPL and in Jupyter notebooks.

Action Description
Generate
⌘I (Windows, Linux Ctrl+I)		 Start Inline Chat in a notebook to generate a codeblock or Markdown block.
# Attach variables from the Jupyter kernel in your chat prompt to get more relevant responses.
Native REPL + ⌘I (Windows, Linux Ctrl+I) Start Inline Chat in the Native Python REPL and run the generated commands.

VS Code commands and APIs

You can use Copilot to get help about VS Code features, settings, and the VS Code extension APIs. Get more information about chat participants.

Action Description
@vscode Use the @vscode chat participant to ask questions about VS Code by using natural language.
Example: @vscode how to enable word wrapping?
@vscode /runCommand Use /runCommand with the @vscode chat participant to run a VS Code command.
@vscode /runCommand enable developer mode
@vscode /search Use /search with the @vscode chat participant to generate a VS Code search.
Example: @vscode /search python files without imports

Tips

  • Use the #vscodeAPI chat variable if you're asking about the VS Code extension API.

Next steps

12/11/2024