docs(claude-code): update outdated node descriptions (#10491)
Refresh 10 nodes against the current Claude Code docs: permission modes (auto, dontAsk), headless mode (Agent SDK), scheduling (cloud and desktop scheduled tasks, fix dead link), loops (/goal), scaling, channels, mobile (rename from mobile-channels), /agents and /rewind. Existing community resources are kept. Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>
This commit is contained in:
Javier CanalesandClaude Sonnet 5.5 authored and GitHub committed
1 parent
e52acc9a80
commit
bb1212d36c
11 files changed
+38
-28
No files matched your search
@@ -1,10 +1,12 @@
|
||||
# /agents
|
||||
|
||||
The `/agents` command is a specialized management interface used to create, configure, and orchestrate sub-agents within your Claude Code environment. This command transitions from a single-assistant model to a multi-agent workflow, allowing you to delegate specific types of work—like security auditing, unit testing, or architectural planning—to "specialists" that have their own isolated context windows and tailored system prompts.
|
||||
The `/agents` command points you to subagent management. On current versions it prints a reminder to ask Claude to create or manage subagents, or to edit the files in `.claude/agents/` or `~/.claude/agents/` yourself. On v2.1.197 and earlier it opened an interactive interface for creating and editing subagents. Subagents run in their own context window with their own system prompt and tools, which keeps specialized work like security audits or test writing out of the main conversation.
|
||||
|
||||
Visit the following resources to learn more:
|
||||
|
||||
- [@official@Create custom subagents](https://code.claude.com/docs/en/sub-agents)
|
||||
- [@official@Commands](https://code.claude.com/docs/en/commands)
|
||||
- [@official@Built-in commands](https://code.claude.com/docs/en/interactive-mode)
|
||||
- [@official@Create custom subagents](https://code.claude.com/docs/en/sub-agents#use-the-agents-command)
|
||||
- [@official@Agents flag format](https://code.claude.com/docs/en/cli-reference)
|
||||
- [@video@Claude Code NEW Sub Agents in 7 Minutes](https://www.youtube.com/watch?v=DNGxMX7ym44)
|
||||
- [@video@Claude Code NEW Sub Agents in 7 Minutes](https://www.youtube.com/watch?v=DNGxMX7ym44)
|
||||
@@ -6,4 +6,4 @@ Visit the following resources to learn more:
|
||||
|
||||
- [@official@Push events into a running session with channels](https://code.claude.com/docs/en/channels)
|
||||
- [@official@Channels reference](https://code.claude.com/docs/en/channels-reference)
|
||||
- [@video@Claude Code Channels in 8 Minutes](https://www.youtube.com/watch?v=QZXaAc80OL0)
|
||||
- [@video@Claude Code Channels in 8 Minutes](https://www.youtube.com/watch?v=QZXaAc80OL0)
|
||||
@@ -1,9 +1,10 @@
|
||||
# Headless mode
|
||||
|
||||
Headless mode runs Claude Code without the interactive interface by passing the -p (or --print) flag with a prompt. It is part of the Agent SDK and is meant for scripts and CI/CD pipelines. Output can be plain text, json or stream-json, and --json-schema forces the result into a defined structure. Flags like --allowedTools, --permission-mode and --bare control which tools can run and how fast the session starts. For full programmatic control with callbacks and message objects, the Python and TypeScript Agent SDK packages offer the same agent loop.
|
||||
Headless mode runs Claude Code without the interactive interface by passing the `-p` (or `--print`) flag with a prompt. It is the CLI form of the Agent SDK and is meant for scripts and CI/CD pipelines. Output can be plain text, `json` or `stream-json`, and `--json-schema` forces the result into a defined structure. Flags like `--allowedTools`, `--permission-mode` and `--bare` control which tools can run and how fast the session starts. For full programmatic control with callbacks and message objects, the Python and TypeScript Agent SDK packages offer the same agent loop.
|
||||
|
||||
Visit the following resources to learn more:
|
||||
|
||||
- [@official@Run Claude Code programmatically](https://code.claude.com/docs/en/headless)
|
||||
- [@official@Agent SDK overview](https://code.claude.com/docs/en/agent-sdk/overview)
|
||||
- [@article@Headless Mode: Unleash AI in Your CI/CD Pipeline](https://dev.to/rajeshroyal/headless-mode-unleash-ai-in-your-cicd-pipeline-1imm)
|
||||
- [@video@Building headless automation with Claude Code | Code w/ Claude](https://www.youtube.com/watch?v=dRsjO-88nBs)
|
||||
- [@video@Building headless automation with Claude Code | Code w/ Claude](https://www.youtube.com/watch?v=dRsjO-88nBs)
|
||||
@@ -1,10 +1,11 @@
|
||||
# Loops
|
||||
|
||||
`/loop` is a Claude Code command that reruns a prompt on a repeating interval within an open session, useful for polling a deployment or babysitting a pull request. Giving it an interval and a prompt runs on a fixed schedule, giving just a prompt lets Claude pick the interval dynamically each iteration, and giving neither runs a built-in maintenance prompt that checks unfinished work and pending PR comments. Loops are session-scoped and expire after seven days, and can be stopped early by pressing Es
|
||||
`/loop` reruns a prompt on a repeating interval inside an open session, which is useful for polling a deployment or watching a pull request. With an interval and a prompt it runs on a fixed schedule, with only a prompt Claude chooses the interval each time, and with neither it runs a built-in maintenance prompt or your own `loop.md` file. Loops are session-scoped, stop when the session closes, and recurring ones expire after seven days. To keep Claude working until a condition is met instead of on a timer, use `/goal`.
|
||||
|
||||
Visit the following resources to learn more:
|
||||
|
||||
- [@article@Run a prompt repeatedly with /loop](https://code.claude.com/docs/en/scheduled-tasks#run-a-prompt-repeatedly-with-/loop)
|
||||
- [@official@Run a prompt repeatedly with /loop](https://code.claude.com/docs/en/scheduled-tasks#run-a-prompt-repeatedly-with-/loop)
|
||||
- [@official@Keep Claude working toward a goal](https://code.claude.com/docs/en/goal)
|
||||
- [@article@What Is the Claude Code /loop Command?](https://www.mindstudio.ai/blog/what-is-claude-code-loop-command-recurring-tasks)
|
||||
- [@video@Stop Prompting Claude. Start Loop Engineering.](https://www.youtube.com/watch?v=YAS4ojuhbW4)
|
||||
- [@video@Claude Code Loops in 7 Minutes](https://www.youtube.com/watch?v=pWZh37iRnDA)
|
||||
- [@video@Claude Code Loops in 7 Minutes](https://www.youtube.com/watch?v=pWZh37iRnDA)
|
||||
@@ -1,8 +0,0 @@
|
||||
# Mobile
|
||||
|
||||
Claude Code on mobile lets you start and follow coding sessions from your phone. Cloud sessions run on Anthropic infrastructure, and Remote Control lets you continue a session running on your own computer from the phone or tablet. This makes it possible to check progress, answer a permission prompt, or send a new task while you are away from your desk.
|
||||
|
||||
Visit the following resources to learn more:
|
||||
|
||||
- [@official@Claude Code on mobile](https://code.claude.com/docs/en/mobile)
|
||||
- [@official@Continue local sessions from any device with Remote Control](https://code.claude.com/docs/en/remote-control)
|
||||
@@ -0,0 +1,8 @@
|
||||
# Mobile
|
||||
|
||||
Claude Code does not have a separate mobile app. The Claude app for iOS and Android is a client that reaches Claude Code sessions in the Code tab. From your phone you can start and follow cloud sessions, drive a session running on your own computer through Remote Control, or message Dispatch a task. The `/mobile` command shows a QR code that opens the right app store for your phone.
|
||||
|
||||
Visit the following resources to learn more:
|
||||
|
||||
- [@official@Claude Code on mobile](https://code.claude.com/docs/en/mobile)
|
||||
- [@official@Remote Control](https://code.claude.com/docs/en/remote-control)
|
||||
@@ -1,9 +1,11 @@
|
||||
# Modes
|
||||
|
||||
Modes set the level of autonomy Claude Code has in a session. Manual mode asks before edits and shell commands, Accept Edits mode approves file changes and common filesystem commands automatically, and Plan mode keeps Claude read-only while it outlines a strategy. Auto mode has a classifier model check actions in the background instead of prompting you, dontAsk denies every call that would prompt, and Bypass Permissions removes all checks. You switch between modes during a session with the mode shortcut, or you start in one with the --permission-mode flag.
|
||||
Modes set the level of autonomy Claude Code has in a session. Manual mode asks before edits and shell commands, Accept Edits mode approves file changes and common filesystem commands automatically, and Plan mode keeps Claude read-only while it outlines a strategy. Auto mode has a classifier model check actions in the background instead of prompting you, `dontAsk` denies every call that would prompt, and Bypass Permissions removes all checks. You switch modes during a session with `Shift+Tab`, or you start in one with the `--permission-mode` flag.
|
||||
|
||||
Visit the following resources to learn more:
|
||||
|
||||
- [@official@Choose a permission mode](https://code.claude.com/docs/en/permission-modes)
|
||||
- [@official@Interactive mode](https://code.claude.com/docs/en/interactive-mode)
|
||||
- [@official@Configure permissions](https://code.claude.com/docs/en/permissions)
|
||||
- [@article@A complete guide to Claude Code permissions](https://www.eesel.ai/blog/claude-code-permissions)
|
||||
- [@video@Claude Code Tutorial #4 - Tools & Permissions](https://www.youtube.com/watch?v=TU0ZcDFq0e0)
|
||||
- [@video@Claude Code Tutorial #4 - Tools & Permissions](https://www.youtube.com/watch?v=TU0ZcDFq0e0)
|
||||
@@ -1,11 +1,13 @@
|
||||
# Permission Modes
|
||||
|
||||
Permission modes control how much Claude Code can do without asking you first. Manual (default) only runs reads without approval, acceptEdits also allows file edits and common filesystem commands, and plan restricts Claude to research and planning. auto lets a separate classifier model review each action and block risky ones, and it is the built-in starting mode in interactive sessions on recent versions. dontAsk denies anything that would prompt, and bypassPermissions runs everything and should only be used in isolated containers or VMs.
|
||||
Permission modes control how much Claude Code can do without asking you first. Manual (`default`) only runs reads without approval, `acceptEdits` also allows file edits and common filesystem commands, and `plan` restricts Claude to research and planning. `auto` lets a separate classifier model review each action and block risky ones, and it is the built-in starting mode in interactive sessions on recent versions. `dontAsk` denies anything that would prompt, and `bypassPermissions` runs everything and should only be used in isolated containers or VMs.
|
||||
|
||||
Visit the following resources to learn more:
|
||||
|
||||
- [@official@Choose a permission mode](https://code.claude.com/docs/en/permission-modes)
|
||||
- [@official@Configure permissions](https://code.claude.com/docs/en/permissions)
|
||||
- [@official@Permission modes](https://code.claude.com/docs/en/permissions#permission-modes)
|
||||
- [@article@A complete guide to Claude Code permissions](https://www.eesel.ai/blog/claude-code-permissions)
|
||||
- [@article@3 Things You Must Know About /permissions in Claude Code](https://wmedia.es/en/tips/claude-code-permissions-3-key-concepts)
|
||||
- [@article@Claude Code on Loop: The Ultimate YOLO Mode](https://mfyz.com/claude-code-on-loop-autonomous-ai-coding/)
|
||||
- [@video@Claude Code Tutorial #4 - Tools & Permissions](https://www.youtube.com/watch?v=TU0ZcDFq0e0)
|
||||
- [@video@Claude Code Tutorial #4 - Tools & Permissions](https://www.youtube.com/watch?v=TU0ZcDFq0e0)
|
||||
@@ -1,9 +1,10 @@
|
||||
# /rewind
|
||||
|
||||
The `/rewind` command is a history management tool that allows you to undo recent turns in your conversation. When executed, it removes the most recent prompt and response from the active context, effectively "winding back" the session to a previous state. This is particularly useful if the AI misunderstood a complex instruction or if a code generation task went in the wrong direction, as it allows you to re-issue the command with better clarity without the baggage of the failed attempt cluttering the model's memory or inflating your token usage.
|
||||
The `/rewind` command restores a session to an earlier point. You can rewind the conversation, the code changes, or both, and you can also summarize the conversation from a selected message instead of discarding it. It relies on checkpoints that Claude Code saves automatically before each file edit. The aliases are `/checkpoint` and `/undo`, and pressing `Esc` twice on an empty prompt opens the same menu.
|
||||
|
||||
Visit the following resources to learn more:
|
||||
|
||||
- [@official@Built-in commands](https://code.claude.com/docs/en/interactive-mode)
|
||||
- [@official@Checkpointing](https://code.claude.com/docs/en/checkpointing)
|
||||
- [@article@Rewind Changes Instantly with Checkpoints](https://wmedia.es/en/tips/rewind-changes-instantly-with-checkpoints)
|
||||
- [@official@Commands](https://code.claude.com/docs/en/commands)
|
||||
- [@official@Built-in commands](https://code.claude.com/docs/en/interactive-mode)
|
||||
- [@article@Rewind Changes Instantly with Checkpoints](https://wmedia.es/en/tips/rewind-changes-instantly-with-checkpoints)
|
||||
@@ -1,9 +1,9 @@
|
||||
# Scaling Claude Code
|
||||
|
||||
Claude Code can split work across several agents or sessions. Subagents handle a side task in their own context and return a summary. Agent view lets you send independent sessions to the background and watch them from one screen. Agent teams coordinate multiple sessions through a shared task list and direct messages, led by one lead agent. Worktrees give each parallel session its own git checkout so they never edit the same files. For larger jobs, dynamic workflows orchestrate many subagents from a rerunnable script, and sessions can message each other directly.
|
||||
Claude Code offers several ways to parallelize work across multiple agents or sessions. Subagents handle a side task in their own context and return a summary. Agent view lets you send independent sessions to the background and watch them from one screen, and agent teams coordinate multiple sessions through a shared task list and direct messages. Worktrees give each parallel session its own git checkout so they never edit the same files. For larger jobs, dynamic workflows run many subagents from a script, and sessions can message each other directly.
|
||||
|
||||
Visit the following resources to learn more:
|
||||
|
||||
- [@official@Run agents in parallel](https://code.claude.com/docs/en/agents)
|
||||
- [@official@Agent view](https://code.claude.com/docs/en/agent-view)
|
||||
- [@official@Orchestrate teams of Claude Code sessions](https://code.claude.com/docs/en/agent-teams)
|
||||
- [@official@Orchestrate teams of Claude Code sessions](https://code.claude.com/docs/en/agent-teams)
|
||||
@@ -1,10 +1,11 @@
|
||||
# Scheduling Jobs
|
||||
|
||||
Scheduling jobs means running Claude Code prompts automatically, without you starting them each time. There are three options. /loop repeats a prompt inside an open session, desktop scheduled tasks run on your machine, and cloud scheduled tasks (formerly called routines) run on Anthropic infrastructure even when your laptop is closed. A cloud task can start from a schedule, an API call, or a GitHub event such as a new pull request. You create them with the /schedule command or on the web.
|
||||
Scheduling jobs means running Claude Code prompts automatically, without you starting them each time. There are three options. `/loop` repeats a prompt inside an open session, desktop scheduled tasks run on your machine, and cloud scheduled tasks (formerly called routines) run on Anthropic infrastructure even when your laptop is closed. A cloud task can start from a schedule, an API call, or a GitHub event such as a new pull request. To react to events as they happen instead of polling on a timer, use Channels.
|
||||
|
||||
Visit the following resources to learn more:
|
||||
|
||||
- [@official@Run prompts on a schedule](https://code.claude.com/docs/en/scheduled-tasks)
|
||||
- [@official@Schedule tasks on the web](https://code.claude.com/docs/en/web-scheduled-tasks)
|
||||
- [@official@Automate work with scheduled tasks](https://code.claude.com/docs/en/routines)
|
||||
- [@official@Desktop scheduled tasks](https://code.claude.com/docs/en/desktop-scheduled-tasks)
|
||||
- [@article@Claude Code Loop vs Scheduled Tasks: Which Should You Use?](https://www.mindstudio.ai/blog/claude-code-loop-vs-scheduled-tasks)
|
||||
- [@video@Claude Code Scheduled Tasks Are Insane](https://www.youtube.com/watch?v=U_cDKkDvPAQ)
|
||||
- [@video@Claude Code Scheduled Tasks Are Insane](https://www.youtube.com/watch?v=U_cDKkDvPAQ)
|
||||
Reference in new issue
Block a user