新闻分析：非洲加深一体化推动区域经济发展
百度 在一个乡土诗国度创建一种基于城市生活和城市意识的城市诗，无疑是今天最大的先锋之举。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.
If you don't yet have a Copilot subscription, you can use Copilot for free by signing up for the Copilot Free plan and get a monthly limit of completions and chat interactions.
Essential keyboard shortcuts
- ??I (Windows, Linux Ctrl+Alt+I) - Open the Chat view
- ?I (Windows, Linux Ctrl+I) - Enter voice chat prompt in Chat view
- ?N (Windows, Linux Ctrl+N) - Start a new chat session in Chat view
- ??I (Windows Ctrl+Shift+I, Linux Ctrl+Shift+Alt+I) - Switch to agent mode
- ?I (Windows, Linux Ctrl+I) - Start inline chat in the editor or terminal
- ?I (Windows, Linux Ctrl+I) (hold) - Start inline voice chat
- Tab - Accept inline suggestion or navigate to the next edit suggestion
- Escape - Dismiss inline suggestion
Access AI in VS Code
-
Start a chat conversation using natural language
- Chat view (??I (Windows, Linux Ctrl+Alt+I)): keep an ongoing chat conversation in the Secondary Side Bar
- Inline chat in the editor or terminal (?I (Windows, Linux Ctrl+I)): ask questions while you're in the flow
- Quick Chat (???L (Windows, Linux Ctrl+Shift+Alt+L)): ask quick questions without leaving your current task
-
AI in the editor
- Code completions: get suggestions as you type, press Tab to accept a suggestion
- Edit context menu actions: access common AI actions like explaining or fixing code, generating tests, or reviewing a text selection
- Code actions: get editor code actions (lightbulb) to fix linting and compiler errors
-
Task-specific smart actions across VS Code
- Generate commit messages and pull request titles and descriptions
- Fix testing errors
- Semantic file search suggestions
Chat experience in VS Code
Start a natural language chat conversation to get help with coding tasks. For example, ask to explain a block of code or a programming concept, refactor a piece of code, or implement a new feature. Get more information about using Copilot Chat.
| Action | Description |
|---|---|
| ??I (Windows, Linux Ctrl+Alt+I) | Open the Chat view in the Secondary Side Bar. |
| ?I (Windows, Linux Ctrl+I) | Start inline chat to open chat in the editor or terminal. |
| ???L (Windows, Linux Ctrl+Shift+Alt+L) | Open Quick Chat without interrupting your workflow. |
| ?N (Windows, Linux Ctrl+N) | Start a new chat session in the Chat view. |
| ?. (Windows, Linux Ctrl+.) | Toggle between different chat modes in the Chat view. |
| ??. (Windows, Linux Ctrl+Alt+.) | Show the model picker to select a different AI model for chat. |
Add Context... |
Attach different types of context to your chat prompt. |
/-command |
Use slash commands for common tasks or invoke a reusable chat prompt. |
#-mention |
Reference common tools or chat variables to provide context within in your prompt. |
@-mention |
Reference chat participants to handle domain-specific requests. |
| Edit () | Edit a previous chat prompt and revert changes. |
| History () | Access your history of chat sessions. |
| Voice () | Enter a chat prompt by using speech (voice chat). The chat response is read out aloud. |
Tips
- Use
#-mentions to add more context to your chat prompt.- 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.
- Choose a chat mode that fits your needs: ask, edit, agent, or create a custom mode.
Add context to your prompt
Get more relevant responses by providing context to your chat prompt. Choose from different context types, such as files, symbols, editor selections, source control commits, test failures, and more.
| Action | Description |
|---|---|
| Add Context | Open a Quick Pick to select relevant context for your chat prompt. Choose from different context types, such as workspace files, symbols, current editor selection, terminal selection, and more. |
| Drag & drop files | Drag & drop a file from the Explorer or Search view, or drag an editor tab onto the Chat view. |
| Drag & drop folders | Drag & drop a folder onto the Chat view to attach the files within it. |
| Drag & drop problem | Drag & drop an item from the Problems panel. |
#<file|folder|symbol> |
Type #, followed by a file, folder, or symbol name, to add it as chat context. |
| `#-mention | Type #, followed by a chat variable to add a specific context type or tool. |
Chat variables
Use chat variables in your chat prompt to reference context that is relevant to your question or to reference specific tools. Reference chat variables or tools in your chat prompt by using the #-mention syntax.
Tools can be invoked as part of an autonomous coding workflow with agent mode. Extensions and MCP servers can provide additional tools that you can use in your chat prompts.
| Chat variable/Tool | Description |
|---|---|
#changes |
List of source control changes. |
#codebase |
Perform a code search in the current workspace to automatically find relevant context for the chat prompt. |
#editFiles |
Tool set to enable creating and editing files in the workspace. |
#extensions |
Tool to find and ask questions about VS Code extensions. For example, "how to get started with Python #extensions?" |
#fetch |
Fetch the content from a web page - provide the URL. |
#findTestFiles |
Tool for finding test files in the current workspace. |
#<file|folder|symbol> |
Add a file, folder, or code symbol as context. |
#githubRepo |
Tool to perform a code search in a GitHub repo. For example, "what is a global snippet #githubRepo microsoft/vscode." |
#new |
Tool to scaffold a new VS Code workspace. |
#openSimpleBrowser |
Tool to open the built-in Simple Browser and preview a locally-deployed web app. |
#problems |
Add workspace issues and problems from the Problems panel as context. Useful while fixing code or debugging. |
#readCellOutput |
Tool for reading the output from a notebook cell. |
#runCommands |
Tool for running commands in the terminal and reading the output. |
#runNotebooks |
Tool for running cells in a notebook and reading the output. |
#runTasks |
Tool for running tasks in the workspace and reading the output. |
#runTests |
Tool for running tests in the workspace and reading the output. |
#search |
Tool set for searching for files in the current workspace. |
#searchResults |
Add the results from the Search view as context to your prompt. |
#selection |
Add the current editor selection as context to your prompt. |
#terminalSelection |
Add the current terminal selection as context to your chat prompt. |
#terminalLastCommand |
Add the last run terminal command as context to your chat prompt. |
#testFailure |
Add test failure information as context. Useful when running and diagnosing tests. |
#usages |
Combination of "Find All References", "Find Implementation", and "Go to Definition". |
#VSCodeAPI |
Tool to reference the VS Code extension APIs and ask questions related to VS Code extension development. |
Slash commands
Slash commands are shortcuts to specific functionality within the chat. You can use them to quickly perform actions, like fixing issues, generating tests, or explaining code.
| Slash command | Description |
|---|---|
/docs |
Generate code documentation comments from editor inline chat. |
/explain |
Explain a code block, file, or programming concept. |
/fix |
Ask Copilot to fix a code block or resolve compiler or linting errors. |
/help |
Get help about using chat in VS Code. |
/tests |
Generate tests for all or only the selected methods and functions in the editor. |
/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. |
/clear |
Start a new chat session in the Chat view. |
/new |
Scaffold a new VS Code workspace or file. Use natural language to describe the type of project/file you need, and preview the scaffolded content before creating it. |
/newNotebook |
Scaffold a new Jupyter notebook based on your requirements. Use natural language to describe what the notebook should contain. |
/search |
Generate a search query for the Search view. Use natural language to describe what you want to search for. |
/startDebugging |
Generate a launch.json debug configuration file and start a debugging session from the Chat view. |
/<prompt file name> |
Run a reusable prompt file in chat. |
Chat participants
Use chat participants to handle domain-specific requests in chat. Chat participants are prefixed with @ and can be used to ask questions about specific topics. VS Code provides built-in chat participants, such as @github, @terminal, and @vscode, and extensions can provide additional participants.
| Chat participant | Description |
|---|---|
@github |
Use the @github participant to ask questions about GitHub repositories, issues, pull requests, and more. 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 PRs from @dancing-mona |
@terminal |
Use the @terminal participant to ask questions about the integrated terminal or shell commands.Example: @terminal list the 5 largest files in this workspace |
@vscode |
Use the @vscode participant to ask questions about VS Code features, settings, and the VS Code extension APIs.Example: @vscode how to enable word wrapping? |
@workspace |
Use the @workspace participant to ask questions about the current workspace.Example: @workspace how is authentication implemented? |
Agent mode
With chat agent mode, you can use natural language to specify a high-level task, and let AI autonomously reason about the request, plan the work needed, and apply the changes to your codebase. Agent mode uses a combination of code editing and tool invocation to accomplish the task you specified. As it processes your request, it monitors the outcome of edits and tools, and iterates to resolve any issues that arise.
| Action | Description |
|---|---|
| ??I (Windows Ctrl+Shift+I, Linux Ctrl+Shift+Alt+I) | Switch to agent mode in the Chat view |
| Tools () | Configure which tools are available in agent mode. Select from built-in tools, MCP servers, and extension-provided tools. |
| Auto-approve tools (Experimental) | Enable auto-approval of all tools in agent mode (chat.tools.autoApprove). |
| Auto-approve terminal commands (Experimental) | Enable auto-approval of terminal commands in agent mode with an allow and deny list. |
| MCP | Configure MCP servers to extend agent mode with extra capabilities and tools. |
Tips
- Add extra tools to agent mode to extend its capabilities.
- Configure custom chat modes to define how agent mode operates, for example to implement a read-only planning mode.
- Define custom instructions to guide agent mode on how to generate and structure code.
Customize your chat experience
Customize your chat experience to generate responses that match your coding style, tools, and developer workflow. There are several ways to customize your chat experience in VS Code:
-
Custom instructions: Define common guidelines or rules for tasks like generating code, performing code reviews, or generating commit messages. Custom instructions describe the conditions in which the AI should perform operate (how a task should be done).
-
Reusable prompt files: Define reusable prompts for common tasks like generating code or performing a code review. Prompt files are standalone prompts that you can run directly in chat. They describe the task to be performed (what should be done).
-
Chat modes: Define how chat operates, which tools it can use, and how it interacts with the codebase. Each chat prompt is run within the boundaries of the chat mode, without having to configure tools and instructions for every request.
Tips
- Define language-specific instructions to get more accurate generated code for each language.
- Store your instructions in your workspace to easily share them with your team.
- Define reusable prompt files for common tasks to save time and help team members get started quickly.
Editor AI features
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 get code suggestions that match your coding style and take your existing code into account. |
| Code comments | Provide a code completion prompt 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 editor inline chat to send a chat request directly from the editor. Use natural language and reference chat variables and slash commands to provide context. |
| F2 | Get AI-powered suggestions when renaming symbols in your code. |
| Context menu actions | Use the editor context menu to access common AI actions, such as explaining code, generating tests, reviewing code, and more. Right-click in the editor to open the context menu and select Copilot. |
| Code Actions (lightbulb) | Select the Code Action (lightbulb) in the editor for fixing linting or compiler errors 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.
Source control and issues
Use AI to analyze the changes in your commits and pull requests and provide suggestions for commit messages and pull request descriptions.
| Action | Description |
|---|---|
#changes |
Add the current source control changes as context in your chat prompt. |
| Commit as context | Add a commit from the source control history as context in your chat prompt. |
| Summarize commit | Right-click a commit in the source control graph and select Copilot > Summarize to generate a summary of the commit changes. |
| Commit message | Generate a commit message for the current changes in a source control commit. |
| Pull request description | 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 |
Review code (experimental)
Use AI to 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. |
Search and settings
Get semantically relevant search results in the Search view or help with searching for settings in the Settings editor.
| Action | Description |
|---|---|
| Semantic search (Preview) | Include semantic search results in the Search view (search.searchView.semanticSearchBehavior). |
| Settings search (Preview) | Include semantic search results in the Settings editor (workbench.settings.showAISearchToggle). |
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 (Experimental) | 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.
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 (Experimental) |
Generate a launch.json debug configuration file and start a debugging session from the Chat view. |
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). |
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 |
|---|---|
| Agent mode | Use agent mode and use a natural language prompt to create a new project or file. For example, Create a svelte web application to track my tasks. |
/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. |
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 terminal inline chat to use natural language for asking about shell commands and the terminal. Example: how many cores on this machine? |
@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 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. |
| ??I (Windows, Linux Ctrl+Alt+I) | Open the Chat view and use edit or agent mode to make notebook edits. |
/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. |