> ## Documentation Index¶
> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt¶
> Use this file to discover all available pages before exploring further.¶
¶
# Common workflows¶
¶
> Step-by-step guides for exploring codebases, fixing bugs, refactoring, testing, and other everyday tasks with Claude Code.¶
¶
This page covers practical workflows for everyday development: exploring unfamiliar code, debugging, refactoring, writing tests, creating PRs, and managing sessions. Each section includes example prompts you can adapt to your own projects. For higher-level patterns and tips, see [Best practices](/en/best-practices).¶
¶
## Understand new codebases¶
¶
### Get a quick codebase overview¶
¶
Suppose you've just joined a new project and need to understand its structure quickly.¶
¶
<Steps>¶
<Step title="Navigate to the project root directory">¶
```bash theme={null}¶
cd /path/to/project ¶
```¶
</Step>¶
¶
<Step title="Start Claude Code">¶
```bash theme={null}¶
claude ¶
```¶
</Step>¶
¶
<Step title="Ask for a high-level overview">¶
```¶
> give me an overview of this codebase ¶
```¶
</Step>¶
¶
<Step title="Dive deeper into specific components">¶
```¶
> explain the main architecture patterns used here ¶
```¶
¶
```¶
> what are the key data models?¶
```¶
¶
```¶
> how is authentication handled?¶
```¶
</Step>¶
</Steps>¶
¶
<Tip>¶
Tips:¶
¶
* Start with broad questions, then narrow down to specific areas¶
* Ask about coding conventions and patterns used in the project¶
* Request a glossary of project-specific terms¶
</Tip>¶
¶
### Find relevant code¶
¶
Suppose you need to locate code related to a specific feature or functionality.¶
¶
<Steps>¶
<Step title="Ask Claude to find relevant files">¶
```¶
> find the files that handle user authentication ¶
```¶
</Step>¶
¶
<Step title="Get context on how components interact">¶
```¶
> how do these authentication files work together? ¶
```¶
</Step>¶
¶
<Step title="Understand the execution flow">¶
```¶
> trace the login process from front-end to database ¶
```¶
</Step>¶
</Steps>¶
¶
<Tip>¶
Tips:¶
¶
* Be specific about what you're looking for¶
* Use domain language from the project¶
* Install a [code intelligence plugin](/en/discover-plugins#code-intelligence) for your language to give Claude precise "go to definition" and "find references" navigation¶
</Tip>¶
¶
***¶
¶
## Fix bugs efficiently¶
¶
Suppose you've encountered an error message and need to find and fix its source.¶
¶
<Steps>¶
<Step title="Share the error with Claude">¶
```¶
> I'm seeing an error when I run npm test ¶
```¶
</Step>¶
¶
<Step title="Ask for fix recommendations">¶
```¶
> suggest a few ways to fix the @ts-ignore in user.ts ¶
```¶
</Step>¶
¶
<Step title="Apply the fix">¶
```¶
> update user.ts to add the null check you suggested ¶
```¶
</Step>¶
</Steps>¶
¶
<Tip>¶
Tips:¶
¶
* Tell Claude the command to reproduce the issue and get a stack trace¶
* Mention any steps to reproduce the error¶
* Let Claude know if the error is intermittent or consistent¶
</Tip>¶
¶
***¶
¶
## Refactor code¶
¶
Suppose you need to update old code to use modern patterns and practices.¶
¶
<Steps>¶
<Step title="Identify legacy code for refactoring">¶
```¶
> find deprecated API usage in our codebase ¶
```¶
</Step>¶
¶
<Step title="Get refactoring recommendations">¶
```¶
> suggest how to refactor utils.js to use modern JavaScript features ¶
```¶
</Step>¶
¶
<Step title="Apply the changes safely">¶
```¶
> refactor utils.js to use ES2024 features while maintaining the same behavior ¶
```¶
</Step>¶
¶
<Step title="Verify the refactoring">¶
```¶
> run tests for the refactored code ¶
```¶
</Step>¶
</Steps>¶
¶
<Tip>¶
Tips:¶
¶
* Ask Claude to explain the benefits of the modern approach¶
* Request that changes maintain backward compatibility when needed¶
* Do refactoring in small, testable increments¶
</Tip>¶
¶
***¶
¶
## Use specialized subagents¶
¶
Suppose you want to use specialized AI subagents to handle specific tasks more effectively.¶
¶
<Steps>¶
<Step title="View available subagents">¶
```¶
> /agents¶
```¶
¶
This shows all available subagents and lets you create new ones.¶
</Step>¶
¶
<Step title="Use subagents automatically">¶
Claude Code automatically delegates appropriate tasks to specialized subagents:¶
¶
```¶
> review my recent code changes for security issues¶
```¶
¶
```¶
> run all tests and fix any failures¶
```¶
</Step>¶
¶
<Step title="Explicitly request specific subagents">¶
```¶
> use the code-reviewer subagent to check the auth module¶
```¶
¶
```¶
> have the debugger subagent investigate why users can't log in¶
```¶
</Step>¶
¶
<Step title="Create custom subagents for your workflow">¶
```¶
> /agents¶
```¶
¶
Then select "Create New subagent" and follow the prompts to define:¶
¶
* A unique identifier that describes the subagent's purpose (for example, `code-reviewer`, `api-designer`).¶
* When Claude should use this agent¶
* Which tools it can access¶
* A system prompt describing the agent's role and behavior¶
</Step>¶
</Steps>¶
¶
<Tip>¶
Tips:¶
¶
* Create project-specific subagents in `.claude/agents/` for team sharing¶
* Use descriptive `description` fields to enable automatic delegation¶
* Limit tool access to what each subagent actually needs¶
* Check the [subagents documentation](/en/sub-agents) for detailed examples¶
</Tip>¶
¶
***¶
¶
## Use Plan Mode for safe code analysis¶
¶
Plan Mode instructs Claude to create a plan by analyzing the codebase with read-only operations, perfect for exploring codebases, planning complex changes, or reviewing code safely. In Plan Mode, Claude uses [`AskUserQuestion`](/en/settings#tools-available-to-claude) to gather requirements and clarify your goals before proposing a plan.¶
¶
### When to use Plan Mode¶
¶
* **Multi-step implementation**: When your feature requires making edits to many files¶
* **Code exploration**: When you want to research the codebase thoroughly before changing anything¶
* **Interactive development**: When you want to iterate on the direction with Claude¶
¶
### How to use Plan Mode¶
¶
**Turn on Plan Mode during a session**¶
¶
You can switch into Plan Mode during a session using **Shift+Tab** to cycle through permission modes.¶
¶
If you are in Normal Mode, **Shift+Tab** first switches into Auto-Accept Mode, indicated by `⏵⏵ accept edits on` at the bottom of the terminal. A subsequent **Shift+Tab** will switch into Plan Mode, indicated by `⏸ plan mode on`.¶
¶
**Start a new session in Plan Mode**¶
¶
To start a new session in Plan Mode, use the `--permission-mode plan` flag:¶
¶
```bash theme={null}¶
claude --permission-mode plan¶
```¶
¶
**Run "headless" queries in Plan Mode**¶
¶
You can also run a query in Plan Mode directly with `-p` (that is, in ["headless mode"](/en/headless)):¶
¶
```bash theme={null}¶
claude --permission-mode plan -p "Analyze the authentication system and suggest improvements"¶
```¶
¶
### Example: Planning a complex refactor¶
¶
```bash theme={null}¶
claude --permission-mode plan¶
```¶
¶
```¶
> I need to refactor our authentication system to use OAuth2. Create a detailed migration plan.¶
```¶
¶
Claude analyzes the current implementation and create a comprehensive plan. Refine with follow-ups:¶
¶
```¶
> What about backward compatibility?¶
> How should we handle database migration?¶
```¶
¶
<Tip>Press `Ctrl+G` to open the plan in your default text editor, where you can edit it directly before Claude proceeds.</Tip>¶
¶
### Configure Plan Mode as default¶
¶
```json theme={null}¶
// .claude/settings.json¶
{¶
"permissions": {¶
"defaultMode": "plan"¶
}¶
}¶
```¶
¶
See [settings documentation](/en/settings#available-settings) for more configuration options.¶
¶
***¶
¶
## Work with tests¶
¶
Suppose you need to add tests for uncovered code.¶
¶
<Steps>¶
<Step title="Identify untested code">¶
```¶
> find functions in NotificationsService.swift that are not covered by tests ¶
```¶
</Step>¶
¶
<Step title="Generate test scaffolding">¶
```¶
> add tests for the notification service ¶
```¶
</Step>¶
¶
<Step title="Add meaningful test cases">¶
```¶
> add test cases for edge conditions in the notification service ¶
```¶
</Step>¶
¶
<Step title="Run and verify tests">¶
```¶
> run the new tests and fix any failures ¶
```¶
</Step>¶
</Steps>¶
¶
Claude can generate tests that follow your project's existing patterns and conventions. When asking for tests, be specific about what behavior you want to verify. Claude examines your existing test files to match the style, frameworks, and assertion patterns already in use.¶
¶
For comprehensive coverage, ask Claude to identify edge cases you might have missed. Claude can analyze your code paths and suggest tests for error conditions, boundary values, and unexpected inputs that are easy to overlook.¶
¶
***¶
¶
## Create pull requests¶
¶
You can create pull requests by asking Claude directly ("create a pr for my changes"), or guide Claude through it step-by-step:¶
¶
<Steps>¶
<Step title="Summarize your changes">¶
```¶
> summarize the changes I've made to the authentication module¶
```¶
</Step>¶
¶
<Step title="Generate a pull request">¶
```¶
> create a pr¶
```¶
</Step>¶
¶
<Step title="Review and refine">¶
```¶
> enhance the PR description with more context about the security improvements¶
```¶
</Step>¶
</Steps>¶
¶
When you create a PR using `gh pr create`, the session is automatically linked to that PR. You can resume it later with `claude --from-pr <number>`.¶
¶
<Tip>¶
Review Claude's generated PR before submitting and ask Claude to highlight potential risks or considerations.¶
</Tip>¶
¶
## Handle documentation¶
¶
Suppose you need to add or update documentation for your code.¶
¶
<Steps>¶
<Step title="Identify undocumented code">¶
```¶
> find functions without proper JSDoc comments in the auth module ¶
```¶
</Step>¶
¶
<Step title="Generate documentation">¶
```¶
> add JSDoc comments to the undocumented functions in auth.js ¶
```¶
</Step>¶
¶
<Step title="Review and enhance">¶
```¶
> improve the generated documentation with more context and examples ¶
```¶
</Step>¶
¶
<Step title="Verify documentation">¶
```¶
> check if the documentation follows our project standards ¶
```¶
</Step>¶
</Steps>¶
¶
<Tip>¶
Tips:¶
¶
* Specify the documentation style you want (JSDoc, docstrings, etc.)¶
* Ask for examples in the documentation¶
* Request documentation for public APIs, interfaces, and complex logic¶
</Tip>¶
¶
***¶
¶
## Work with images¶
¶
Suppose you need to work with images in your codebase, and you want Claude's help analyzing image content.¶
¶
<Steps>¶
<Step title="Add an image to the conversation">¶
You can use any of these methods:¶
¶
1. Drag and drop an image into the Claude Code window¶
2. Copy an image and paste it into the CLI with ctrl+v (Do not use cmd+v)¶
3. Provide an image path to Claude. E.g., "Analyze this image: /path/to/your/image.png"¶
</Step>¶
¶
<Step title="Ask Claude to analyze the image">¶
```¶
> What does this image show?¶
```¶
¶
```¶
> Describe the UI elements in this screenshot¶
```¶
¶
```¶
> Are there any problematic elements in this diagram?¶
```¶
</Step>¶
¶
<Step title="Use images for context">¶
```¶
> Here's a screenshot of the error. What's causing it?¶
```¶
¶
```¶
> This is our current database schema. How should we modify it for the new feature?¶
```¶
</Step>¶
¶
<Step title="Get code suggestions from visual content">¶
```¶
> Generate CSS to match this design mockup¶
```¶
¶
```¶
> What HTML structure would recreate this component?¶
```¶
</Step>¶
</Steps>¶
¶
<Tip>¶
Tips:¶
¶
* Use images when text descriptions would be unclear or cumbersome¶
* Include screenshots of errors, UI designs, or diagrams for better context¶
* You can work with multiple images in a conversation¶
* Image analysis works with diagrams, screenshots, mockups, and more¶
* When Claude references images (for example, `[Image #1]`), `Cmd+Click` (Mac) or `Ctrl+Click` (Windows/Linux) the link to open the image in your default viewer¶
</Tip>¶
¶
***¶
¶
## Reference files and directories¶
¶
Use @ to quickly include files or directories without waiting for Claude to read them.¶
¶
<Steps>¶
<Step title="Reference a single file">¶
```¶
> Explain the logic in @src/utils/auth.js¶
```¶
¶
This includes the full content of the file in the conversation.¶
</Step>¶
¶
<Step title="Reference a directory">¶
```¶
> What's the structure of @src/components?¶
```¶
¶
This provides a directory listing with file information.¶
</Step>¶
¶
<Step title="Reference MCP resources">¶
```¶
> Show me the data from @github:repos/owner/repo/issues¶
```¶
¶
This fetches data from connected MCP servers using the format @server:resource. See [MCP resources](/en/mcp#use-mcp-resources) for details.¶
</Step>¶
</Steps>¶
¶
<Tip>¶
Tips:¶
¶
* File paths can be relative or absolute¶
* @ file references add `CLAUDE.md` in the file's directory and parent directories to context¶
* Directory references show file listings, not contents¶
* You can reference multiple files in a single message (for example, "@file1.js and @file2.js")¶
</Tip>¶
¶
***¶
¶
## Use extended thinking (thinking mode)¶
¶
[Extended thinking](https://platform.claude.com/docs/en/build-with-claude/extended-thinking) is enabled by default, giving Claude space to reason through complex problems step-by-step before responding. This reasoning is visible in verbose mode, which you can toggle on with `Ctrl+O`.¶
¶
Additionally, Opus 4.6 introduces adaptive reasoning: instead of a fixed thinking token budget, the model dynamically allocates thinking based on your [effort level](/en/model-config#adjust-effort-level) setting. Extended thinking and adaptive reasoning work together to give you control over how deeply Claude reasons before responding.¶
¶
Extended thinking is particularly valuable for complex architectural decisions, challenging bugs, multi-step implementation planning, and evaluating tradeoffs between different approaches.¶
¶
<Note>¶
Phrases like "think", "think hard", "ultrathink", and "think more" are interpreted as regular prompt instructions and don't allocate thinking tokens.¶
</Note>¶
¶
### Configure thinking mode¶
¶
Thinking is enabled by default, but you can adjust or disable it.¶
¶
| Scope | How to configure | Details |¶
| ------------------------ | ------------------------------------------------------------------------------------------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------- |¶
| **Effort level** | Adjust in `/model` or set [`CLAUDE_CODE_EFFORT_LEVEL`](/en/settings#environment-variables) | Control thinking depth for Opus 4.6 and Sonnet 4.6: low, medium, high. See [Adjust effort level](/en/model-config#adjust-effort-level) |¶
| **`ultrathink` keyword** | Include "ultrathink" anywhere in your prompt | Sets effort to high for that turn on Opus 4.6 and Sonnet 4.6. Useful for one-off tasks requiring deep reasoning without permanently changing your effort setting |¶
| **Toggle shortcut** | Press `Option+T` (macOS) or `Alt+T` (Windows/Linux) | Toggle thinking on/off for the current session (all models). May require [terminal configuration](/en/terminal-config) to enable Option key shortcuts |¶
| **Global default** | Use `/config` to toggle thinking mode | Sets your default across all projects (all models).<br />Saved as `alwaysThinkingEnabled` in `~/.claude/settings.json` |¶
| **Limit token budget** | Set [`MAX_THINKING_TOKENS`](/en/settings#environment-variables) environment variable | Limit the thinking budget to a specific number of tokens (ignored on Opus 4.6 unless set to 0). Example: `export MAX_THINKING_TOKENS=10000` |¶
¶
To view Claude's thinking process, press `Ctrl+O` to toggle verbose mode and see the internal reasoning displayed as gray italic text.¶
¶
### How extended thinking works¶
¶
Extended thinking controls how much internal reasoning Claude performs before responding. More thinking provides more space to explore solutions, analyze edge cases, and self-correct mistakes.¶
¶
**With Opus 4.6**, thinking uses adaptive reasoning: the model dynamically allocates thinking tokens based on the [effort level](/en/model-config#adjust-effort-level) you select (low, medium, high). This is the recommended way to tune the tradeoff between speed and reasoning depth.¶
¶
**With other models**, thinking uses a fixed budget of up to 31,999 tokens from your output budget. You can limit this with the [`MAX_THINKING_TOKENS`](/en/settings#environment-variables) environment variable, or disable thinking entirely via `/config` or the `Option+T`/`Alt+T` toggle.¶
¶
`MAX_THINKING_TOKENS` is ignored on Opus 4.6 and Sonnet 4.6, since adaptive reasoning controls thinking depth instead. The one exception: setting `MAX_THINKING_TOKENS=0` still disables thinking entirely on any model. To disable adaptive thinking and revert to the fixed thinking budget, set `CLAUDE_CODE_DISABLE_ADAPTIVE_THINKING=1`. See [environment variables](/en/settings#environment-variables).¶
¶
<Warning>¶
You're charged for all thinking tokens used, even though Claude 4 models show summarized thinking¶
</Warning>¶
¶
***¶
¶
## Resume previous conversations¶
¶
When starting Claude Code, you can resume a previous session:¶
¶
* `claude --continue` continues the most recent conversation in the current directory¶
* `claude --resume` opens a conversation picker or resumes by name¶
* `claude --from-pr 123` resumes sessions linked to a specific pull request¶
¶
From inside an active session, use `/resume` to switch to a different conversation.¶
¶
Sessions are stored per project directory. The `/resume` picker shows sessions from the same git repository, including worktrees.¶
¶
### Name your sessions¶
¶
Give sessions descriptive names to find them later. This is a best practice when working on multiple tasks or features.¶
¶
<Steps>¶
<Step title="Name the current session">¶
Use `/rename` during a session to give it a memorable name:¶
¶
```¶
> /rename auth-refactor¶
```¶
¶
You can also rename any session from the picker: run `/resume`, navigate to a session, and press `R`.¶
</Step>¶
¶
<Step title="Resume by name later">¶
From the command line:¶
¶
```bash theme={null}¶
claude --resume auth-refactor¶
```¶
¶
Or from inside an active session:¶
¶
```¶
> /resume auth-refactor¶
```¶
</Step>¶
</Steps>¶
¶
### Use the session picker¶
¶
The `/resume` command (or `claude --resume` without arguments) opens an interactive session picker with these features:¶
¶
**Keyboard shortcuts in the picker:**¶
¶
| Shortcut | Action |¶
| :-------- | :------------------------------------------------ |¶
| `↑` / `↓` | Navigate between sessions |¶
| `→` / `←` | Expand or collapse grouped sessions |¶
| `Enter` | Select and resume the highlighted session |¶
| `P` | Preview the session content |¶
| `R` | Rename the highlighted session |¶
| `/` | Search to filter sessions |¶
| `A` | Toggle between current directory and all projects |¶
| `B` | Filter to sessions from your current git branch |¶
| `Esc` | Exit the picker or search mode |¶
¶
**Session organization:**¶
¶
The picker displays sessions with helpful metadata:¶
¶
* Session name or initial prompt¶
* Time elapsed since last activity¶
* Message count¶
* Git branch (if applicable)¶
¶
Forked sessions (created with `/rewind` or `--fork-session`) are grouped together under their root session, making it easier to find related conversations.¶
¶
<Tip>¶
Tips:¶
¶
* **Name sessions early**: Use `/rename` when starting work on a distinct task—it's much easier to find "payment-integration" than "explain this function" later¶
* Use `--continue` for quick access to your most recent conversation in the current directory¶
* Use `--resume session-name` when you know which session you need¶
* Use `--resume` (without a name) when you need to browse and select¶
* For scripts, use `claude --continue --print "prompt"` to resume in non-interactive mode¶
* Press `P` in the picker to preview a session before resuming it¶
* The resumed conversation starts with the same model and configuration as the original¶
¶
How it works:¶
¶
1. **Conversation Storage**: All conversations are automatically saved locally with their full message history¶
2. **Message Deserialization**: When resuming, the entire message history is restored to maintain context¶
3. **Tool State**: Tool usage and results from the previous conversation are preserved¶
4. **Context Restoration**: The conversation resumes with all previous context intact¶
</Tip>¶
¶
***¶
¶
## Run parallel Claude Code sessions with Git worktrees¶
¶
When working on multiple tasks at once, you need each Claude session to have its own copy of the codebase so changes don't collide. Git worktrees solve this by creating separate working directories that each have their own files and branch, while sharing the same repository history and remote connections. This means you can have Claude working on a feature in one worktree while fixing a bug in another, without either session interfering with the other.¶
¶
Use the `--worktree` (`-w`) flag to create an isolated worktree and start Claude in it. The value you pass becomes the worktree directory name and branch name:¶
¶
```bash theme={null}¶
# Start Claude in a worktree named "feature-auth"¶
# Creates .claude/worktrees/feature-auth/ with a new branch¶
claude --worktree feature-auth¶
¶
# Start another session in a separate worktree¶
claude --worktree bugfix-123¶
```¶
¶
If you omit the name, Claude generates a random one automatically:¶
¶
```bash theme={null}¶
# Auto-generates a name like "bright-running-fox"¶
claude --worktree¶
```¶
¶
Worktrees are created at `<repo>/.claude/worktrees/<name>` and branch from the default remote branch. The worktree branch is named `worktree-<name>`.¶
¶
You can also ask Claude to "work in a worktree" or "start a worktree" during a session, and it will create one automatically.¶
¶
### Subagent worktrees¶
¶
Subagents can also use worktree isolation to work in parallel without conflicts. Ask Claude to "use worktrees for your agents" or configure it in a [custom subagent](/en/sub-agents#supported-frontmatter-fields) by adding `isolation: worktree` to the agent's frontmatter. Each subagent gets its own worktree that is automatically cleaned up when the subagent finishes without changes.¶
¶
### Worktree cleanup¶
¶
When you exit a worktree session, Claude handles cleanup based on whether you made changes:¶
¶
* **No changes**: the worktree and its branch are removed automatically¶
* **Changes or commits exist**: Claude prompts you to keep or remove the worktree. Keeping preserves the directory and branch so you can return later. Removing deletes the worktree directory and its branch, discarding all uncommitted changes and commits¶
¶
To clean up worktrees outside of a Claude session, use [manual worktree management](#manage-worktrees-manually).¶
¶
<Tip>¶
Add `.claude/worktrees/` to your `.gitignore` to prevent worktree contents from appearing as untracked files in your main repository.¶
</Tip>¶
¶
### Manage worktrees manually¶
¶
For more control over worktree location and branch configuration, create worktrees with Git directly. This is useful when you need to check out a specific existing branch or place the worktree outside the repository.¶
¶
```bash theme={null}¶
# Create a worktree with a new branch¶
git worktree add ../project-feature-a -b feature-a¶
¶
# Create a worktree with an existing branch¶
git worktree add ../project-bugfix bugfix-123¶
¶
# Start Claude in the worktree¶
cd ../project-feature-a && claude¶
¶
# Clean up when done¶
git worktree list¶
git worktree remove ../project-feature-a¶
```¶
¶
Learn more in the [official Git worktree documentation](https://git-scm.com/docs/git-worktree).¶
¶
<Tip>¶
Remember to initialize your development environment in each new worktree according to your project's setup. Depending on your stack, this might include running dependency installation (`npm install`, `yarn`), setting up virtual environments, or following your project's standard setup process.¶
</Tip>¶
¶
### Non-git version control¶
¶
Worktree isolation works with git by default. For other version control systems like SVN, Perforce, or Mercurial, configure [WorktreeCreate and WorktreeRemove hooks](/en/hooks#worktreecreate) to provide custom worktree creation and cleanup logic. When configured, these hooks replace the default git behavior when you use `--worktree`.¶
¶
For automated coordination of parallel sessions with shared tasks and messaging, see [agent teams](/en/agent-teams).¶
¶
***¶
¶
## Get notified when Claude needs your attention¶
¶
When you kick off a long-running task and switch to another window, you can set up desktop notifications so you know when Claude finishes or needs your input. This uses the `Notification` [hook event](/en/hooks-guide#get-notified-when-claude-needs-input), which fires whenever Claude is waiting for permission, idle and ready for a new prompt, or completing authentication.¶
¶
<Steps>¶
<Step title="Open the hooks menu">¶
Type `/hooks` and select `Notification` from the list of events.¶
</Step>¶
¶
<Step title="Configure the matcher">¶
Select `+ Match all (no filter)` to fire on all notification types. To notify only for specific events, select `+ Add new matcher…` and enter one of these values:¶
¶
| Matcher | Fires when |¶
| :------------------- | :---------------------------------------------- |¶
| `permission_prompt` | Claude needs you to approve a tool use |¶
| `idle_prompt` | Claude is done and waiting for your next prompt |¶
| `auth_success` | Authentication completes |¶
| `elicitation_dialog` | Claude is asking you a question |¶
</Step>¶
¶
<Step title="Add your notification command">¶
Select `+ Add new hook…` and enter the command for your OS:¶
¶
<Tabs>¶
<Tab title="macOS">¶
Uses [`osascript`](https://ss64.com/mac/osascript.html) to trigger a native macOS notification through AppleScript:¶
¶
```¶
osascript -e 'display notification "Claude Code needs your attention" with title "Claude Code"'¶
```¶
</Tab>¶
¶
<Tab title="Linux">¶
Uses `notify-send`, which is pre-installed on most Linux desktops with a notification daemon:¶
¶
```¶
notify-send 'Claude Code' 'Claude Code needs your attention'¶
```¶
</Tab>¶
¶
<Tab title="Windows (PowerShell)">¶
Uses PowerShell to show a native message box through .NET's Windows Forms:¶
¶
```¶
powershell.exe -Command "[System.Reflection.Assembly]::LoadWithPartialName('System.Windows.Forms'); [System.Windows.Forms.MessageBox]::Show('Claude Code needs your attention', 'Claude Code')"¶
```¶
</Tab>¶
</Tabs>¶
</Step>¶
¶
<Step title="Save to user settings">¶
Select `User settings` to apply the notification across all your projects.¶
</Step>¶
</Steps>¶
¶
For the full walkthrough with JSON configuration examples, see [Automate workflows with hooks](/en/hooks-guide#get-notified-when-claude-needs-input). For the complete event schema and notification types, see the [Notification reference](/en/hooks#notification).¶
¶
***¶
¶
## Use Claude as a unix-style utility¶
¶
### Add Claude to your verification process¶
¶
Suppose you want to use Claude Code as a linter or code reviewer.¶
¶
**Add Claude to your build script:**¶
¶
```json theme={null}¶
// package.json¶
{¶
...¶
"scripts": {¶
...¶
"lint:claude": "claude -p 'you are a linter. please look at the changes vs. main and report any issues related to typos. report the filename and line number on one line, and a description of the issue on the second line. do not return any other text.'"¶
}¶
}¶
```¶
¶
<Tip>¶
Tips:¶
¶
* Use Claude for automated code review in your CI/CD pipeline¶
* Customize the prompt to check for specific issues relevant to your project¶
* Consider creating multiple scripts for different types of verification¶
</Tip>¶
¶
### Pipe in, pipe out¶
¶
Suppose you want to pipe data into Claude, and get back data in a structured format.¶
¶
**Pipe data through Claude:**¶
¶
```bash theme={null}¶
cat build-error.txt | claude -p 'concisely explain the root cause of this build error' > output.txt¶
```¶
¶
<Tip>¶
Tips:¶
¶
* Use pipes to integrate Claude into existing shell scripts¶
* Combine with other Unix tools for powerful workflows¶
* Consider using --output-format for structured output¶
</Tip>¶
¶
### Control output format¶
¶
Suppose you need Claude's output in a specific format, especially when integrating Claude Code into scripts or other tools.¶
¶
<Steps>¶
<Step title="Use text format (default)">¶
```bash theme={null}¶
cat data.txt | claude -p 'summarize this data' --output-format text > summary.txt¶
```¶
¶
This outputs just Claude's plain text response (default behavior).¶
</Step>¶
¶
<Step title="Use JSON format">¶
```bash theme={null}¶
cat code.py | claude -p 'analyze this code for bugs' --output-format json > analysis.json¶
```¶
¶
This outputs a JSON array of messages with metadata including cost and duration.¶
</Step>¶
¶
<Step title="Use streaming JSON format">¶
```bash theme={null}¶
cat log.txt | claude -p 'parse this log file for errors' --output-format stream-json¶
```¶
¶
This outputs a series of JSON objects in real-time as Claude processes the request. Each message is a valid JSON object, but the entire output is not valid JSON if concatenated.¶
</Step>¶
</Steps>¶
¶
<Tip>¶
Tips:¶
¶
* Use `--output-format text` for simple integrations where you just need Claude's response¶
* Use `--output-format json` when you need the full conversation log¶
* Use `--output-format stream-json` for real-time output of each conversation turn¶
</Tip>¶
¶
***¶
¶
## Ask Claude about its capabilities¶
¶
Claude has built-in access to its documentation and can answer questions about its own features and limitations.¶
¶
### Example questions¶
¶
```¶
> can Claude Code create pull requests?¶
```¶
¶
```¶
> how does Claude Code handle permissions?¶
```¶
¶
```¶
> what skills are available?¶
```¶
¶
```¶
> how do I use MCP with Claude Code?¶
```¶
¶
```¶
> how do I configure Claude Code for Amazon Bedrock?¶
```¶
¶
```¶
> what are the limitations of Claude Code?¶
```¶
¶
<Note>¶
Claude provides documentation-based answers to these questions. For executable examples and hands-on demonstrations, refer to the specific workflow sections above.¶
</Note>¶
¶
<Tip>¶
Tips:¶
¶
* Claude always has access to the latest Claude Code documentation, regardless of the version you're using¶
* Ask specific questions to get detailed answers¶
* Claude can explain complex features like MCP integration, enterprise configurations, and advanced workflows¶
</Tip>¶
¶
***¶
¶
## Next steps¶
¶
<CardGroup cols={2}>¶
<Card title="Best practices" icon="lightbulb" href="/en/best-practices">¶
Patterns for getting the most out of Claude Code¶
</Card>¶
¶
<Card title="How Claude Code works" icon="gear" href="/en/how-claude-code-works">¶
Understand the agentic loop and context management¶
</Card>¶
¶
<Card title="Extend Claude Code" icon="puzzle-piece" href="/en/features-overview">¶
Add skills, hooks, MCP, subagents, and plugins¶
</Card>¶
¶
<Card title="Reference implementation" icon="code" href="https://github.com/anthropics/claude-code/tree/main/.devcontainer">¶
Clone our development container reference implementation¶
</Card>¶
</CardGroup>¶