claude-code-20261007T173702Z / documentation capture

managed-mcp Changed · +17 / -5 lines

from line 301

301301

302302#### How `serverUrl` entries match

303303

304URLs support `*` wildcards anywhere in the pattern, including the scheme. Hostname matching is case-insensitive and ignores a trailing FQDN dot, so `https://Mcp.Example.com/*` matches `https://mcp.example.com/api`. Paths stay case-sensitive.

304URLs support `*` wildcards, including `*` as the whole scheme. Hostname matching is case-insensitive and ignores a trailing FQDN dot, so `https://Mcp.Example.com/*` matches `https://mcp.example.com/api`. Paths stay case-sensitive. If you give no port, how you write the hostname decides whether the pattern matches only the scheme's default port or every port:

305305

306* **Hostname written out in full**: the default port only, 443 for `https` and 80 for `http`

307* **Hostname with a `*` in it**: every port

308

306309The table shows what common patterns allow:

307310

308311| Pattern | Allows |

309312| :- | :- |

310| `https://mcp.example.com/*` | All paths on a specific domain |

311| `https://mcp.example.com` | Also all paths on that domain. A pattern with no path matches any path |

312| `https://*.example.com/*` | Any subdomain of `example.com` |

313| `https://mcp.example.com/*` | All paths on a specific domain, on port 443 only |

314| `https://mcp.example.com` | Also all paths on that domain, on port 443 only. A pattern with no path matches any path |

315| `https://mcp.example.com:8443/*` | All paths on that domain, on port 8443 only |

316| `https://mcp.example.com:*/*` | All paths on that domain, on any port, 443 included |

317| `https://*.example.com/*` | Any subdomain of `example.com`, on any port |

313318| `http://localhost:*/*` | Any port on localhost |

314| `*://mcp.example.com/*` | Any scheme to a specific domain |

319| `*://mcp.example.com/*` | Any scheme to a specific domain, each scheme on its default port only |

315320

321Entries in `deniedMcpServers` match ports the same way, so choose an entry for `staging.example.com` by the ports and schemes you need to block:

322

323* `https://staging.example.com/*`: blocks `https` servers on that host on port 443 only, so it doesn't block a server at `https://staging.example.com:8443/api`

324* `https://staging.example.com:*/*`: blocks `https` servers on that host on every port

325* `*://staging.example.com:*/*`: blocks that host over any scheme and on any port

326

316327<h4 id="how-policy-entries-expand">

317328 Environment variables in `serverCommand` and `serverUrl` entries

318329</h4>

from line 488

477488 | :- | :- |

478489 | HTTP server at `https://mcp.example.com/api` | Allowed: matches allowlist URL pattern, no denylist match |

479490 | HTTP server at `https://staging.example.com/api` | Blocked: matches both, but the denylist takes precedence |

491 | HTTP server at `https://staging.example.com:8443/api` | Allowed: matches allowlist URL pattern, [no denylist match on this port](#how-serverurl-entries-match) |

480492 | HTTP server at `https://other.com/mcp` | Blocked: doesn't match the allowlist |

481493</Accordion>

482494

overview Changed · +6 / -4 lines

from line 22

2222 curl -fsSL https://claude.ai/install.sh | bash

2323 ```

2424

25 On Windows, your prompt shows `PS C:\` when you're in PowerShell and `C:\` without the `PS` when you're in CMD.

26

2527 **Windows PowerShell:**

2628

2729 ```powershell theme={null}

from line 38

3638

3739 When the installer finishes, open a new terminal window and run `claude --version`. A working installation prints a version number. If your shell says `claude` isn't found or isn't recognized, the install directory isn't on your PATH yet: see [Fix your PATH](/docs/en/troubleshoot-install#command-not-found-claude-after-installation).

3840

39 If you see `The token '&&' is not a valid statement separator`, you're in PowerShell, not CMD. If you see `'irm' is not recognized as an internal or external command`, you're in CMD, not PowerShell. Your prompt shows `PS C:\` when you're in PowerShell and `C:\` without the `PS` when you're in CMD.

41 If you see `The token '&&' is not a valid statement separator`, you're in PowerShell, not CMD. If you see `'irm' is not recognized as an internal or external command`, you're in CMD, not PowerShell.

4042

41 If the install command fails with `syntax error near unexpected token '<'`, a `403`, or another curl error, see [Troubleshoot installation](/docs/en/troubleshoot-install#find-your-error) to match the error to a fix and for alternative install methods.

43 If the install command fails with `syntax error near unexpected token '<'`, a `403`, or any other error, see [Troubleshoot installation](/docs/en/troubleshoot-install#find-your-error) to match the error to a fix and for alternative install methods.

4244

4345 [Git for Windows](https://git-scm.com/downloads/win) is recommended on native Windows so Claude Code can use the Bash tool. If Git for Windows is not installed, Claude Code uses PowerShell as the shell tool instead. WSL setups do not need Git for Windows.

4446

from line 81

7981 claude

8082 ```

8183

82 You'll be prompted to log in on first use. If you've set the `ANTHROPIC_API_KEY` environment variable, Claude Code skips the login prompt and asks you to approve the key instead. That's it! [Continue with the Quickstart →](/docs/en/quickstart)

84 Claude Code prompts you to log in on first use. If you've set the `ANTHROPIC_API_KEY` environment variable and you approve the key when Claude Code asks whether to use it, Claude Code skips the login prompt. [Continue with the Quickstart →](/docs/en/quickstart)

8385

8486 <Tip>

8587 See [advanced setup](/docs/en/setup) for installation options, manual updates, or uninstallation instructions. Visit [installation troubleshooting](/docs/en/troubleshoot-install) if you hit issues.

quickstart Changed · +42 / -74 lines

## Step 2: Start your first session ## Step 3: Ask your first question ## Step 4: Make your first code change ## Step 5: Use Git with Claude Code ## Step 6: Fix a bug or add a feature ## Step 7: Test out other common workflows ## Step 2: Log in to your account ## Step 3: Start your first session ## Step 4: Ask your first question ## Step 5: Make your first code change ## Step 6: Use Git with Claude Code ## Step 7: Fix a bug or add a feature ## Step 8: Test out other common workflows

from line 1

11# Quickstart

22

3> Welcome to Claude Code!

3> Install Claude Code in your terminal, sign in, and use the CLI to explore your codebase and make your first code change.

44

5This quickstart guide will have you using AI-powered coding assistance in a few minutes. By the end, you'll understand how to use Claude Code for common development tasks.

5This quickstart covers Claude Code in your terminal: installing the CLI, signing in from your first session, and using it for common development tasks in your own project.

66

77## Before you begin

88

from line 9

99Make sure you have:

1010

1111* A terminal or command prompt open

12 * If you've never used the terminal before, check out the [terminal guide](/docs/en/terminal-guide)

1312* A code project to work with

1413* A [Claude subscription](https://claude.com/pricing?utm_source=claude_code\&utm_medium=docs\&utm_content=quickstart_prereq) (Pro, Max, Team, or Enterprise), [Claude Console](https://platform.claude.com/) account, or access through a [supported cloud provider](/docs/en/third-party-integrations)

1514

1615<Note>

17 This guide covers the terminal CLI. Claude Code is also available on the [web](https://claude.ai/code), as a [desktop app](/docs/en/desktop), in [VS Code](/docs/en/vs-code) and [JetBrains IDEs](/docs/en/jetbrains), in [Slack](/docs/en/slack), and in CI/CD with [GitHub Actions](/docs/en/github-actions) and [GitLab](/docs/en/gitlab-ci-cd). See [all interfaces](/docs/en/overview#use-claude-code-everywhere).

16 These cases are covered on other pages:

17

18 * **Never used a terminal before**: start with the [terminal guide](/docs/en/terminal-guide)

19 * **Want Claude Code somewhere other than the terminal**: Claude Code is also available on the [web](https://claude.ai/code), as a [desktop app](/docs/en/desktop), in [VS Code](/docs/en/vs-code) and [JetBrains IDEs](/docs/en/jetbrains), in [Slack](/docs/en/slack), and in CI/CD with [GitHub Actions](/docs/en/github-actions) and [GitLab](/docs/en/gitlab-ci-cd). See [all interfaces](/docs/en/overview#use-claude-code-everywhere).

1820</Note>

1921

2022## Step 1: Install Claude Code

from line 31

2931 curl -fsSL https://claude.ai/install.sh | bash

3032 ```

3133

34 On Windows, your prompt shows `PS C:\` when you're in PowerShell and `C:\` without the `PS` when you're in CMD.

35

3236 **Windows PowerShell:**

3337

3438 ```powershell theme={null}

from line 47

4347

4448 When the installer finishes, open a new terminal window and run `claude --version`. A working installation prints a version number. If your shell says `claude` isn't found or isn't recognized, the install directory isn't on your PATH yet: see [Fix your PATH](/docs/en/troubleshoot-install#command-not-found-claude-after-installation).

4549

46 If you see `The token '&&' is not a valid statement separator`, you're in PowerShell, not CMD. If you see `'irm' is not recognized as an internal or external command`, you're in CMD, not PowerShell. Your prompt shows `PS C:\` when you're in PowerShell and `C:\` without the `PS` when you're in CMD.

50 If you see `The token '&&' is not a valid statement separator`, you're in PowerShell, not CMD. If you see `'irm' is not recognized as an internal or external command`, you're in CMD, not PowerShell.

4751

48 If the install command fails with `syntax error near unexpected token '<'`, a `403`, or another curl error, see [Troubleshoot installation](/docs/en/troubleshoot-install#find-your-error) to match the error to a fix and for alternative install methods.

52 If the install command fails with `syntax error near unexpected token '<'`, a `403`, or any other error, see [Troubleshoot installation](/docs/en/troubleshoot-install#find-your-error) to match the error to a fix and for alternative install methods.

4953

5054 [Git for Windows](https://git-scm.com/downloads/win) is recommended on native Windows so Claude Code can use the Bash tool. If Git for Windows is not installed, Claude Code uses PowerShell as the shell tool instead. WSL setups do not need Git for Windows.

5155

from line 91

8791

8892The command prints a version number followed by `(Claude Code)`.

8993

90## Step 2: Log in to your account

94## Step 2: Start your first session

9195

92Claude Code requires an account to use. Start an interactive session with the `claude` command and you'll be prompted to log in on first use:

96Open your terminal in any project directory and start Claude Code:

9397

9498```bash theme={null}

99cd /path/to/your/project

95100claude

96101```

97102

98For Claude subscription or Console accounts, follow the prompts to complete authentication in your browser. If you've set the `ANTHROPIC_API_KEY` environment variable, Claude Code skips the login prompt and asks you to approve the key instead. To switch accounts later or re-authenticate, type `/login` inside the running session:

103Replace `/path/to/your/project` with the path to the project you want to work on.

99104

100```text wrap theme={null}

101/login

102```

105Claude Code prompts you to log in on first use. For Claude subscription or Console accounts, follow the prompts to complete authentication in your browser. If you've set the `ANTHROPIC_API_KEY` environment variable and you approve the key when Claude Code asks whether to use it, Claude Code skips the login prompt.

103106

104107You can log in using any of these account types:

105108

from line 113

110113

111114Once logged in, your credentials are stored and you won't need to log in again. Learn more in [Credential Management](/docs/en/authentication#credential-management).

112115

113## Step 3: Start your first session

116The Claude Code prompt appears with the version, current model, and working directory shown above it. Type `/help` for available commands or `/resume` to continue a previous conversation. To switch accounts later or re-authenticate, type `/login` inside the running session.

114117

115Open your terminal in any project directory and start Claude Code:

118## Step 3: Ask your first question

116119

117```bash theme={null}

118cd /path/to/your/project

119claude

120```

120Try one of these commands:

121121

122Replace `/path/to/your/project` with the path to the project you want to work on.

123

124You'll see the Claude Code prompt with the version, current model, and working directory shown above it. Type `/help` for available commands or `/resume` to continue a previous conversation.

125

126## Step 4: Ask your first question

127

128Start by understanding your codebase. Try one of these commands:

129

130122```text wrap theme={null}

131123what does this project do?

132124```

from line 155

163155 Claude Code reads your project files as needed. You don't have to manually add context.

164156</Note>

165157

166## Step 5: Make your first code change

158## Step 4: Make your first code change

167159

168Now have Claude Code do some actual coding. Try a simple task:

160Try a small task:

169161

170162```text wrap theme={null}

171163add a hello world function to the main file

from line 165

173165

174166Claude Code finds the appropriate file and shows you the change. If it asks before making the change, select **Yes** to approve.

175167

176With Claude Code v2.1.283 or later, auto mode is the [built-in starting permission mode](/docs/en/permission-modes#eliminate-prompts-with-auto-mode) for interactive terminal sessions: a classifier reviews actions instead of you, and Claude edits most files and runs most commands without asking you. On earlier versions, auto mode is the built-in starting permission mode only on Pro, Max, and Team plans. For the session you start right after installing, see [First session after an install or upgrade](/docs/en/env-vars#first-session-after-an-install-or-upgrade).

168The session's [permission mode](/docs/en/permission-modes) sets which actions Claude can take without asking you first. Press `Shift+Tab` at any time to switch the permission mode of the session you're in.

177169

178<Note>

179 Your settings or your organization can set a different starting permission mode. [Which permission mode a session starts in](/docs/en/permission-modes#which-mode-a-session-starts-in) lists what does. Press `Shift+Tab` at any time to switch the permission mode of the session you're in.

180</Note>

170## Step 5: Use Git with Claude Code

181171

182## Step 6: Use Git with Claude Code

183

184172Claude Code makes Git operations conversational:

185173

186174```text wrap theme={null}

from line 193

205193help me resolve merge conflicts

206194```

207195

208## Step 7: Fix a bug or add a feature

196## Step 6: Fix a bug or add a feature

209197

210Claude is proficient at debugging and feature implementation.

211

212198Describe what you want in natural language:

213199

214200```text wrap theme={null}

from line 207

221207there's a bug where users can submit empty forms - fix it

222208```

223209

224Claude Code will:

210## Step 7: Test out other common workflows

225211

226* Locate the relevant code

227* Understand the context

228* Implement a solution

229* Run tests if available

230

231## Step 8: Test out other common workflows

232

233212There are a number of ways to work with Claude:

234213

235214**Refactor code**

from line 308

329308

330309Now that you've learned the basics, explore more advanced features:

331310

332<CardGroup cols={2}>

333 <Card title="How Claude Code works" icon="microchip" href="/docs/en/how-claude-code-works">

334 Understand the agentic loop, built-in tools, and how Claude Code interacts with your project

335 </Card>

311* [How Claude Code works](/docs/en/how-claude-code-works): understand the agentic loop, built-in tools, and how Claude Code interacts with your project

312* [Best practices](/docs/en/best-practices): get better results with effective prompting and project setup

313* [Common workflows](/docs/en/common-workflows): step-by-step guides for common tasks

314* [Extend Claude Code](/docs/en/features-overview): customize with CLAUDE.md, skills, hooks, MCP, and more

336315

337 <Card title="Best practices" icon="star" href="/docs/en/best-practices">

338 Get better results with effective prompting and project setup

339 </Card>

316See [advanced setup](/docs/en/setup) for installation options, manual updates, or uninstallation instructions.

340317

341 <Card title="Common workflows" icon="graduation-cap" href="/docs/en/common-workflows">

342 Step-by-step guides for common tasks

343 </Card>

344

345 <Card title="Extend Claude Code" icon="puzzle-piece" href="/docs/en/features-overview">

346 Customize with CLAUDE.md, skills, hooks, MCP, and more

347 </Card>

348</CardGroup>

349

350318## Getting help

351319

352* **In Claude Code**: Type `/help` or ask a "how do I" question

353* **Documentation**: You're here! Browse other guides

354* **Courses**: Take [Claude Code 101](https://academy.claude.com/courses/claude-code-101) and other free self-paced courses on [Claude Academy](https://academy.claude.com/)

355* **Community**: Join the [Discord server](https://www.anthropic.com/discord) for tips and support

320* **In Claude Code**: type `/help` or ask a "how do I" question

321* **Documentation**: browse the other guides on this site

322* **Courses**: take [Claude Code 101](https://academy.claude.com/courses/claude-code-101) and other free self-paced courses on [Claude Academy](https://academy.claude.com/)

323* **Community**: join the [Discord server](https://www.anthropic.com/discord) for tips and support

356324

setup Changed · +6 / -4 lines

from line 39

3939 curl -fsSL https://claude.ai/install.sh | bash

4040 ```

4141

42 On Windows, your prompt shows `PS C:\` when you're in PowerShell and `C:\` without the `PS` when you're in CMD.

43

4244 **Windows PowerShell:**

4345

4446 ```powershell theme={null}

from line 55

5355

5456 When the installer finishes, open a new terminal window and run `claude --version`. A working installation prints a version number. If your shell says `claude` isn't found or isn't recognized, the install directory isn't on your PATH yet: see [Fix your PATH](/docs/en/troubleshoot-install#command-not-found-claude-after-installation).

5557

56 If you see `The token '&&' is not a valid statement separator`, you're in PowerShell, not CMD. If you see `'irm' is not recognized as an internal or external command`, you're in CMD, not PowerShell. Your prompt shows `PS C:\` when you're in PowerShell and `C:\` without the `PS` when you're in CMD.

58 If you see `The token '&&' is not a valid statement separator`, you're in PowerShell, not CMD. If you see `'irm' is not recognized as an internal or external command`, you're in CMD, not PowerShell.

5759

58 If the install command fails with `syntax error near unexpected token '<'`, a `403`, or another curl error, see [Troubleshoot installation](/docs/en/troubleshoot-install#find-your-error) to match the error to a fix and for alternative install methods.

60 If the install command fails with `syntax error near unexpected token '<'`, a `403`, or any other error, see [Troubleshoot installation](/docs/en/troubleshoot-install#find-your-error) to match the error to a fix and for alternative install methods.

5961

6062 [Git for Windows](https://git-scm.com/downloads/win) is recommended on native Windows so Claude Code can use the Bash tool. If Git for Windows is not installed, Claude Code uses PowerShell as the shell tool instead. WSL setups do not need Git for Windows.

6163

from line 188

186188

187189Claude Code requires a Pro, Max, Team, Enterprise, or Console account. The free claude.ai plan does not include Claude Code access. You can also use Claude Code with a third-party API provider like [Amazon Bedrock](/docs/en/amazon-bedrock), [Google Cloud's Agent Platform](/docs/en/google-vertex-ai), or [Microsoft Foundry](/docs/en/microsoft-foundry).

188190

189After installing, log in by running `claude` and following the browser prompts. If the `ANTHROPIC_API_KEY` environment variable is set, Claude Code prompts you once to approve the key instead of opening a browser. See [Authentication](/docs/en/authentication) for all account types and team setup options.

191After installing, log in by running `claude` and following the browser prompts. If you've set the `ANTHROPIC_API_KEY` environment variable and you approve the key when Claude Code asks whether to use it, Claude Code skips the login prompt. See [Authentication](/docs/en/authentication) for all account types and team setup options.

190192

191193## Update Claude Code

192194

authentication Changed · +1 / -1 lines

from line 6

66

77## Log in to Claude Code

88

9After [installing Claude Code](/docs/en/setup#install-claude-code), run `claude` in your terminal. On first launch, Claude Code opens a browser window for you to log in. If you've set the `ANTHROPIC_API_KEY` environment variable, Claude Code skips the login prompt and asks you to approve the key instead.

9After [installing Claude Code](/docs/en/setup#install-claude-code), run `claude` in your terminal. On first launch, Claude Code opens a browser window for you to log in. If you've set the `ANTHROPIC_API_KEY` environment variable and you approve the key when Claude Code asks whether to use it, Claude Code skips the login prompt.

1010

1111If the browser doesn't open automatically, press `c` to copy the login URL to your clipboard, then paste it into your browser.

1212

claude-tag Changed · +1 / -1 lines

This page is larger than the 256 KiB this site keeps, so one side of the diff below stops where the stored text does.

Nothing in the body moved in this read. What changed is above.

costs Changed · +1 / -1 lines

from line 334

334334* **Use plan mode for complex tasks**: Press Shift+Tab to cycle to [plan mode](/docs/en/permission-modes#analyze-before-you-edit-with-plan-mode) before implementation. Claude explores the codebase and proposes an approach for your approval, preventing expensive re-work when the initial direction is wrong.

335335* **Course-correct early**: If Claude starts heading the wrong direction, press Escape to stop immediately. Use `/rewind` or double-tap Escape to restore conversation and code to a previous checkpoint.

336336* **Give verification targets**: Include test cases, paste screenshots, or define expected output in your prompt. When Claude can verify its own work, it catches issues before you need to request fixes.

337* **Test incrementally**: Write one file, test it, then continue. This catches issues early when they're cheap to fix.

337* **Test incrementally**: Write one file, test it, then continue. This catches issues early.

338338

339339## Background token usage

340340

fast-mode Changed · +1 / -1 lines

from line 76

7676

7777Fast mode pricing is flat across the full 1M token context window. For the standard Opus rate to compare against, see the [Claude pricing reference](https://platform.claude.com/docs/en/about-claude/pricing).

7878

79The first time you enable fast mode in a conversation, you pay the full fast mode uncached input token price for the entire conversation context. The deeper into a conversation you are, the more this costs, so enabling fast mode from the start is cheaper. The cost applies once per conversation, so toggling fast mode off and on again later does not repeat it. For the mechanism, see [how fast mode interacts with the prompt cache](/docs/en/prompt-caching#turning-on-fast-mode).

79The first time you enable fast mode in a conversation, you pay the full fast mode uncached input token price for the entire conversation context. The deeper into a conversation you are, the more this costs, so the charge is smallest when you enable fast mode at the start. The cost applies once per conversation, so toggling fast mode off and on again later doesn't repeat it. For the mechanism, see [how fast mode interacts with the prompt cache](/docs/en/prompt-caching#turning-on-fast-mode).

8080

8181### See where fast mode spend appears

8282

plugin-evals Changed · +1 / -1 lines

from line 106

106106

107107 The most common first finding is a `Δ` near zero with the case's `tool_used: Skill` grader failing, which means Claude isn't choosing your skill on natural phrasing. Adjust the skill's [`description`](/docs/en/skills#frontmatter-reference), run `claude plugin eval .` again, and compare.

108108

109 To iterate on one case cheaply, run a single arm once. A single run is noisy, so confirm any change at the default three runs before you trust it. With one arm the table shows `SCORE` and `PASS%` columns instead of `WITH`, `W/OUT`, and `Δ`:

109 To iterate on one case with fewer runs, run a single arm once. A single run is noisy, so confirm any change at the default three runs before you trust it. With one arm the table shows `SCORE` and `PASS%` columns instead of `WITH`, `W/OUT`, and `Δ`:

110110

111111 ```bash theme={null}

112112 claude plugin eval . --case <case-name> --runs 1 --ablation none

prompt-caching Changed · +2 / -2 lines

from line 252

252252

253253On a Pro or Max plan, when you resume a large session after a long break, Claude Code [offers to resume from a summary](/docs/en/sessions#resume-from-a-summary) so later requests don't carry the full history.

254254

255The time to live (TTL) controls how long a gap the cache survives. The API offers two: a five-minute TTL, and a [one-hour TTL](https://platform.claude.com/docs/en/build-with-claude/prompt-caching#1-hour-cache-duration) that keeps the cache warm through longer breaks but [bills cache writes at a higher rate](https://platform.claude.com/docs/en/build-with-claude/prompt-caching#pricing). The longer TTL helps when you leave a session idle and come back to it, because you skip the reprocessing an expired prefix costs. It costs more on short bursts of work that never idle past five minutes, where the higher write rate applies and the longer cache lifetime goes unused.

255The time to live (TTL) controls how long a gap the cache survives. The API offers two: a five-minute TTL, and a [one-hour TTL](https://platform.claude.com/docs/en/build-with-claude/prompt-caching#1-hour-cache-duration) that keeps the cache warm through longer breaks but [charges a higher rate for cache writes](https://platform.claude.com/docs/en/build-with-claude/prompt-caching#pricing). The longer TTL helps when you leave a session idle and come back to it, because you skip the reprocessing an expired prefix costs. It costs more on short bursts of work that never idle past five minutes, where the higher write rate applies and the longer cache lifetime goes unused.

256256

257257### Which TTL each request gets

258258

from line 268

268268| Main conversation | One hour | Five minutes |

269269| Everything else | Five minutes, except the server-controlled helper requests, which get one hour | Five minutes |

270270

271Once you go over your plan's usage limit and Claude Code draws on [usage credits](https://support.claude.com/en/articles/12429409-extra-usage-for-paid-claude-plans), you are billed for that usage, so Claude Code drops the main conversation to the cheaper five-minute TTL. To keep the one-hour TTL there, [choose the TTL yourself](#choose-the-ttl-yourself).

271Once you go over your plan's usage limit and Claude Code draws on [usage credits](https://support.claude.com/en/articles/12429409-extra-usage-for-paid-claude-plans), you are billed for that usage, so Claude Code drops the main conversation to the five-minute TTL, which charges a lower rate for cache writes. To keep the one-hour TTL there, [choose the TTL yourself](#choose-the-ttl-yourself).

272272

273273### Choose the TTL yourself

274274

prompt-library Changed · +1 / -1 lines

from line 1338

13381338 },

13391339 "review-your-changes-before": {

13401340 title: "Review your changes before you commit",

1341 teaches: "Catch problems while they're still cheap to fix. Claude reads the changed files in full, not just the diff lines, so it spots issues a quick self-review misses.",

1341 teaches: "Catch problems while they take less work to fix. Claude reads the changed files in full, not just the diff lines, so it spots issues a quick self-review misses.",

13421342 next: "Run `/code-review` for the same check in one command",

13431343 prompt: "review my uncommitted changes and flag anything that looks risky before I commit"

13441344 },

sub-agents Changed · +2 / -2 lines

from line 16

1616* **Enforce constraints** by limiting which tools a subagent can use

1717* **Reuse configurations** across projects with user-level subagents

1818* **Specialize behavior** with focused system prompts for specific domains

19* **Control costs** by routing tasks to faster, cheaper models like Haiku

19* **Control costs** by routing tasks to faster, lower-cost models like Haiku

2020

2121Claude uses each subagent's description to decide when to delegate tasks. When you create a subagent, write a clear description so Claude knows when to use it.

2222

from line 1189

11891189| Permissions | Prompts surface in your terminal | [Prompts surface in your main session](#run-subagents-in-foreground-or-background) when running in the background |

11901190| Prompt cache | Shared with main session | Separate cache |

11911191

1192Because a fork's system prompt and tool definitions are identical to the parent, its first request reuses the parent's [prompt cache](/docs/en/prompt-caching#subagents-and-the-cache). This makes forking cheaper than spawning a fresh subagent for tasks that need the same context.

1192Because a fork's system prompt and tool definitions are identical to the parent, its first request reuses the parent's [prompt cache](/docs/en/prompt-caching#subagents-and-the-cache). Because of that reuse, a fork costs less than a fresh subagent for tasks that need the same context.

11931193

11941194When Claude spawns a fork through the Agent tool, it can pass `isolation: "worktree"` so the fork's file edits are written to a separate git worktree instead of your checkout. A fork can't spawn further forks.

11951195

One read of Claude Code CLIclaude-code-20261007T173702Z

What this read moved