What this read moved
1–7 of 7context-window Changed · +6 / -6 lines
from line 108
108108 tokens: 380,
109109 color: '#4A9B8E',
110110 vis: 'brief',
111 restoredAfterCompact: true,
112111 desc: 'This rule in `.claude/rules/` has a `paths:` pattern matching `src/api/**`. It loaded automatically when Claude read a file in that directory. You see "Loaded .claude/rules/api-conventions.md" in your terminal, but not the rule content.',
113112 link: '/en/memory#path-specific-rules'
114113 }, {
from line 137
138137 tokens: 290,
139138 color: '#4A9B8E',
140139 vis: 'brief',
141 restoredAfterCompact: true,
142140 desc: 'Another path-scoped rule, this one matching `*.test.ts` files. Triggered when Claude read auth.test.ts. Shown as a one-line "Loaded" notice.',
143141 link: '/en/memory#path-specific-rules'
144142 }, {
from line 619
621619 if (detailRef.current) detailRef.current.scrollTop = 0;
622620 }, [hovEvent]);
623621 const focusT = hovEvent ? hovEvent.t : time;
624 const takeaway = isCompacted ? 'Compaction replaces the conversation with a structured summary. System prompt, CLAUDE.md, memory, and MCP tools reload automatically. Claude Code also re-reads up to five of the files modified most recently, reloads the rules that match them, and re-injects the skills you invoked. The skill listing does not reload.' : focusT < STARTUP_END ? 'A lot loads before you type anything. CLAUDE.md, memory, skills, and MCP tools are all in context before your first prompt.' : focusT < 0.28 ? "Your prompt is tiny compared to what's already loaded. Most of Claude's context is project knowledge, not your words." : focusT < 0.50 ? 'Each file Claude reads grows the context. Path-scoped rules load automatically alongside matching files.' : focusT < 0.71 ? 'Hooks fire automatically on tool events. Output reaches Claude via additionalContext JSON. Exit code 2 surfaces stderr to Claude. Plain stdout on exit 0 goes to the debug log, not the transcript.' : focusT < 0.79 ? 'Follow-up questions keep building on the same context. Everything from earlier is still there.' : focusT < 0.87 ? "The subagent works in its own separate context window. None of its file reads touch yours. Only the final summary comes back." : focusT < 0.88 ? 'Bang commands run in your shell and prefix the output to your next message. Useful for grounding Claude in command results without it running them.' : focusT < 0.90 ? 'User-only skills stay out of context entirely until you invoke them. The skill index at startup only lists skills Claude can call on its own.' : '/compact summarizes the conversation to free space while keeping key information. In a real session, run it when context starts affecting performance or before a long new task.';
625 const terminalView = isCompacted ? 'A "Conversation compacted" message, then a one-line "Read auth.ts" for each re-read file and "Skills restored (commit-push)". The rules show as "Loaded" lines on Claude\'s next turn. None of the content itself appears.' : focusT < STARTUP_END ? 'The input box, waiting for your first message. Everything above loads silently before you type anything.' : focusT < 0.28 ? 'Your prompt. Claude hasn\'t started working yet.' : focusT < 0.52 ? 'Your prompt and "Reading files...". Rules show as one-line "Loaded" notices, not their content.' : focusT < 0.72 ? "Claude's response and file diffs. Hooks fire silently. Tool output like npm test shows as a brief summary, not the full content." : focusT < 0.79 ? 'Your follow-up prompt.' : focusT < 0.86 ? "A brief notice that a subagent is working, then its result. You don't see the subagent's individual file reads." : focusT < 0.90 ? "Claude's response, your git status output, and the commit-push skill running." : 'Your full conversation. /compact is available to run.';
622 const takeaway = isCompacted ? 'Compaction replaces the conversation with a structured summary. System prompt, CLAUDE.md, memory, and MCP tools reload automatically. Claude Code also re-reads up to five of the files modified most recently and re-injects the skills you invoked. The skill listing does not reload.' : focusT < STARTUP_END ? 'A lot loads before you type anything. CLAUDE.md, memory, skills, and MCP tools are all in context before your first prompt.' : focusT < 0.28 ? "Your prompt is tiny compared to what's already loaded. Most of Claude's context is project knowledge, not your words." : focusT < 0.50 ? 'Each file Claude reads grows the context. Path-scoped rules load automatically alongside matching files.' : focusT < 0.71 ? 'Hooks fire automatically on tool events. Output reaches Claude via additionalContext JSON. Exit code 2 surfaces stderr to Claude. Plain stdout on exit 0 goes to the debug log, not the transcript.' : focusT < 0.79 ? 'Follow-up questions keep building on the same context. Everything from earlier is still there.' : focusT < 0.87 ? "The subagent works in its own separate context window. None of its file reads touch yours. Only the final summary comes back." : focusT < 0.88 ? 'Bang commands run in your shell and prefix the output to your next message. Useful for grounding Claude in command results without it running them.' : focusT < 0.90 ? 'User-only skills stay out of context entirely until you invoke them. The skill index at startup only lists skills Claude can call on its own.' : '/compact summarizes the conversation to free space while keeping key information. In a real session, run it when context starts affecting performance or before a long new task.';
623 const terminalView = isCompacted ? 'A "Conversation compacted" message, then a one-line "Read auth.ts" for each re-read file and "Skills restored (commit-push)". None of the content itself appears.' : focusT < STARTUP_END ? 'The input box, waiting for your first message. Everything above loads silently before you type anything.' : focusT < 0.28 ? 'Your prompt. Claude hasn\'t started working yet.' : focusT < 0.52 ? 'Your prompt and "Reading files...". Rules show as one-line "Loaded" notices, not their content.' : focusT < 0.72 ? "Claude's response and file diffs. Hooks fire silently. Tool output like npm test shows as a brief summary, not the full content." : focusT < 0.79 ? 'Your follow-up prompt.' : focusT < 0.86 ? "A brief notice that a subagent is working, then its result. You don't see the subagent's individual file reads." : focusT < 0.90 ? "Claude's response, your git status output, and the commit-push skill running." : 'Your full conversation. /compact is available to run.';
626624 const mono = 'var(--font-mono, ui-monospace, SFMono-Regular, Menlo, monospace)';
627625 const renderWithCode = s => s.split('`').map((part, i) => i % 2 === 1 ? <code key={i} style={{
628626 fontFamily: mono,
from line 946
948946 lineHeight: 1.5,
949947 marginTop: 4
950948 }}>
951 This is what's left in context: startup content, which lives outside the message history and reloads after compaction, a structured summary of the entire conversation, the files modified most recently, which Claude Code re-reads along with the rules that match them, and the body of each skill you invoked. Skill descriptions don't reload.
949 This is what's left in context: startup content, which lives outside the message history and reloads after compaction, a structured summary of the entire conversation, the files modified most recently, and the body of each skill you invoked. Skill descriptions don't reload.
952950 </div>
953951 </div>}
954952 {time > 0 && visible.length > 0 && <div style={{
from line 1603
16051603| Context that hooks added earlier | Summarized with the rest of the conversation |
16061604| [SessionStart hooks](/docs/en/hooks-guide#re-inject-context-after-compaction) that match the `compact` source | Claude Code runs them and adds their output to the compacted context |
16071605
1608Path-scoped rules and nested CLAUDE.md files load into message history when their trigger file is read, so compaction summarizes them away with everything else. Right after compaction, Claude Code re-reads up to five of the files Claude has read or edited in the session, choosing the ones modified most recently, and reloads the rules and nested CLAUDE.md files that apply to those files. A file over 5,000 tokens comes back as a path reference without its content, shown as `Referenced file` instead of `Read`. Its rules still reload. If a rule must persist across compaction, drop the `paths:` frontmatter or move it to the project-root CLAUDE.md.
1606Right after compaction, Claude Code re-reads up to five of the files Claude has read or edited in the session, choosing the ones modified most recently. A file over 5,000 tokens comes back as a path reference without its content, shown as `Referenced file` instead of `Read`.
1607
1608Path-scoped rules and nested CLAUDE.md files load into message history when their trigger file is read, so compaction summarizes them away with everything else. If a rule must persist across compaction, drop the `paths:` frontmatter or move it to the project-root CLAUDE.md.
16091609
16101610Skill bodies are re-injected after compaction, but large skills are truncated to fit the per-skill cap, and the oldest invoked skills are dropped once the total budget is exceeded. Truncation keeps the start of the file, so put the most important instructions near the top of `SKILL.md`.
16111611
hooks Changed · +6 / -2 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.
from line 1750
17501750| `subagent_type` | string | `"Explore"` | Type of specialized agent to use |
17511751| `model` | string | `"sonnet"` | Optional model alias to override the default |
17521752
1753When a foreground Agent call completes, your [PostToolUse hook](#posttooluse) receives the subagent's final text and run telemetry in `tool_response`. Read these fields to inspect the run; for token and cost rollups across subagents, use the [token and cost counters](/docs/en/monitoring-usage#token-counter) filtered to `query_source` `"subagent"`, since `totalTokens` and `usage` cover the final request only:
1753When a foreground Agent call completes, your [PostToolUse hook](#posttooluse) receives the subagent's result and run telemetry in `tool_response`. Read these fields to inspect the run; for token and cost rollups across subagents, use the [token and cost counters](/docs/en/monitoring-usage#token-counter) filtered to `query_source` `"subagent"`, since `totalTokens` and `usage` cover the final request only:
17541754
17551755| Field | Type | Example | Description |
17561756| :------------------ | :----- | :---------------------------------------------------- | :------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
17571757| `status` | string | `"completed"` | `"completed"` for foreground subagents, `"async_launched"` for background subagents. As of v2.1.198, subagents run in the background by default, so an omitted `run_in_background` also produces `"async_launched"` |
17581758| `agentId` | string | `"a4d2c8f1e0b3a297"` | Identifier for the subagent run |
1759| `content` | array | `[{"type": "text", "text": "Found 12 endpoints..."}]` | The subagent's final text blocks |
1759| `content` | array | `[{"type": "text", "text": "Found 12 endpoints..."}]` | The subagent's final text blocks, or, for a subagent whose report goes through `SubagentHandback`, a short note about that hand-back in their place |
17601760| `resolvedModel` | string | `"claude-sonnet-4-5"` | Model the subagent started on, which may differ from the requested model |
17611761| `modelsUsed` | array | `["claude-sonnet-4-5", "claude-haiku-4-5"]` | Models used in order, with consecutive repeats collapsed; set only when the model was swapped mid-run. Requires Claude Code v2.1.212 or later |
17621762| `totalTokens` | number | `12450` | Token count from the subagent's final API request: input, output, and cache tokens combined. This isn't a total across the whole run |
from line 1764
17641764| `totalToolUseCount` | number | `7` | Count of tool calls the subagent made |
17651765| `usage` | object | `{"input_tokens": 8320, ...}` | Per-type token breakdown of the final API request: `input_tokens`, `output_tokens`, `cache_creation_input_tokens`, `cache_read_input_tokens` |
17661766
1767On Claude Code v2.1.271 or later, a subagent that runs with the [`SubagentHandback`](/docs/en/tools-reference) tool, which Claude Code provides in [auto mode](/docs/en/permission-modes#eliminate-prompts-with-auto-mode), delivers its report through that tool rather than returning it as text. The `content` field of its `completed` result then carries a short note about that hand-back rather than the report itself. To read the report, match a `PreToolUse` or `PostToolUse` hook on `SubagentHandback` and read `tool_input.message`.
1768
17671769For background subagents, the tool returns when the task moves to the background, so `tool_response` carries no usage fields: a background launch returns immediately, and a foreground task that Claude Code backgrounds mid-run returns at that transition. It has `status: "async_launched"`, `agentId`, `description`, `prompt`, `outputFile`, and `resolvedModel`.
17681770
17691771On a `completed` response, `resolvedModel` names the model the subagent started on, which can differ from the `model` value in `tool_input`, such as when `availableModels` or another override applies. On an `async_launched` response, `resolvedModel` names the model in use when the agent moved to the background, so a swap that happened before backgrounding is reflected there. `modelsUsed` and the backgrounding-time `resolvedModel` behavior require Claude Code v2.1.212 or later.
from line 2390
23882390
23892391In addition to the [common input fields](#common-input-fields), SubagentStop hooks receive `stop_hook_active`, `agent_id`, `agent_type`, `agent_transcript_path`, and `last_assistant_message`. The `agent_type` field is the value used for matcher filtering. The `transcript_path` is the main session's transcript, while `agent_transcript_path` is the subagent's own transcript stored in a nested `subagents/` folder. The `last_assistant_message` field contains the text content of the subagent's final response, so hooks can access it without parsing the transcript file.
23902392
2393On Claude Code v2.1.271 or later, a subagent that runs with the [`SubagentHandback`](/docs/en/tools-reference) tool delivers its report through that tool before it stops. The `last_assistant_message` field then holds the subagent's closing text, if any, which is not the delivered report. The report is that call's `message` input, which a `PreToolUse` or `PostToolUse` hook matched on `SubagentHandback` receives as `tool_input.message`.
2394
23912395SubagentStop hooks also receive the `background_tasks` and `session_crons` arrays described under [Stop input](#stop-input). Both arrays are scoped to the parent session, not the subagent.
23922396
23932397```json theme={null}
from line 2714
27102714| `project_settings` | `.claude/settings.json` changes |
27112715| `local_settings` | `.claude/settings.local.json` changes |
27122716| `policy_settings` | `managed-settings.json` or a file in `managed-settings.d/` changes |
2713| `skills` | A skill file in `.claude/skills/` changes |
2714
2715This example logs all configuration changes for security auditing:
2716
2717```json theme={null}
2718{
2719 "hooks": {
2720 "ConfigChange": [
2721 {
2722 "hooks": [
2723 {
2724 "type": "command",
2725 "command": "${CLAUDE_PROJECT_DIR}/.claude/hooks/audit-config-change.sh",
2726 "args": []
2727 }
2728 ]
2729 }
2730 ]
2731 }
2732}
2733```
2734
2735#### ConfigChange input
2736
2737In addition to the [common input fields](#common-input-fields), ConfigChange hooks receive `source` and optionally `file_path`. The `source` field indicates which configuration type changed, and `file_path` provides the path to the specific file that was modified.
2738
2739```json theme={null}
2740{
2741 "session_id": "abc123",
2742 "transcript_path": "/Users/.../.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",
2743 "cwd": "/Users/...",
2744 "hook_event_name": "ConfigChange",
2745 "source": "project_settings",
2746 "file_path": "/User
2717| `skills` | A sk
overview Changed · +5 / -5 lines
from line 18
1818 <Tab title="Native Install (Recommended)">
1919 **macOS, Linux, WSL:**
2020
21 ```bash theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null}
21 ```bash theme={null}
2222 curl -fsSL https://claude.ai/install.sh | bash
2323 ```
2424
2525 **Windows PowerShell:**
2626
27 ```powershell theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null}
27 ```powershell theme={null}
2828 irm https://claude.ai/install.ps1 | iex
2929 ```
3030
3131 **Windows CMD:**
3232
33 ```batch theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null}
33 ```batch theme={null}
3434 curl -fsSL https://claude.ai/install.cmd -o install.cmd && install.cmd && del install.cmd
3535 ```
3636
from line 46
4646 </Tab>
4747
4848 <Tab title="Homebrew">
49 ```bash theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null}
49 ```bash theme={null}
5050 brew install --cask claude-code
5151 ```
5252
from line 58
5858 </Tab>
5959
6060 <Tab title="WinGet">
61 ```powershell theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null}
61 ```powershell theme={null}
6262 winget install Anthropic.ClaudeCode
6363 ```
6464
setup Changed · +5 / -5 lines
from line 37
3737 <Tab title="Native Install (Recommended)">
3838 **macOS, Linux, WSL:**
3939
40 ```bash theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null}
40 ```bash theme={null}
4141 curl -fsSL https://claude.ai/install.sh | bash
4242 ```
4343
4444 **Windows PowerShell:**
4545
46 ```powershell theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null}
46 ```powershell theme={null}
4747 irm https://claude.ai/install.ps1 | iex
4848 ```
4949
5050 **Windows CMD:**
5151
52 ```batch theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null}
52 ```batch theme={null}
5353 curl -fsSL https://claude.ai/install.cmd -o install.cmd && install.cmd && del install.cmd
5454 ```
5555
from line 65
6565 </Tab>
6666
6767 <Tab title="Homebrew">
68 ```bash theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null}
68 ```bash theme={null}
6969 brew install --cask claude-code
7070 ```
7171
from line 77
7777 </Tab>
7878
7979 <Tab title="WinGet">
80 ```powershell theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null} theme={null}
80 ```powershell theme={null}
8181 winget install Anthropic.ClaudeCode
8282 ```
8383
sub-agents Changed · +3 / -3 lines
from line 414
414414* `WaitForMcpServers`
415415* `Workflow`
416416
417The second filter applies to subagents running in the background. Apart from `Agent` and `ExitPlanMode`, which follow the first filter's conditions wherever the subagent runs, a background subagent keeps every MCP tool but only these built-in tools: `Read`, `Grep`, `Glob`, `Bash`, `PowerShell`, `Edit`, `Write`, `NotebookEdit`, `WebFetch`, `WebSearch`, `TodoWrite`, `Skill`, `ToolSearch`, `EnterWorktree`, `ExitWorktree`, `Monitor`, `TaskStop`, `SendMessage`, and `Artifact`. Claude Code removes every other built-in tool from a background subagent, whether inherited or listed in the `tools` field, so the same definition can resolve to different tools in the foreground and the background. The removal reports no error unless it leaves the `tools` list [resolving to nothing](/docs/en/errors#agent-would-be-spawned-with-zero-tools).
417The second filter applies to subagents running in the background. Apart from `Agent` and `ExitPlanMode`, which follow the first filter's conditions wherever the subagent runs, a background subagent keeps every MCP tool but only these built-in tools: `Read`, `Grep`, `Glob`, `Bash`, `PowerShell`, `Edit`, `Write`, `NotebookEdit`, `WebFetch`, `WebSearch`, `TodoWrite`, `Skill`, `ToolSearch`, `EnterWorktree`, `ExitWorktree`, `Monitor`, `TaskStop`, `SendMessage`, and `Artifact`, plus [`SubagentHandback`](/docs/en/tools-reference) for a subagent that reports through it. Claude Code removes every other built-in tool from a background subagent, whether inherited or listed in the `tools` field, so the same definition can resolve to different tools in the foreground and the background. The removal reports no error unless it leaves the `tools` list [resolving to nothing](/docs/en/errors#agent-would-be-spawned-with-zero-tools).
418418
419419[`ListAgents`](/docs/en/cross-session-messaging) follows these filters like any built-in tool: a foreground subagent inherits it in sessions where cross-session messaging is enabled, and a background subagent doesn't keep it.
420420
from line 547
547547
548548The main conversation's permission mode decides whether Claude Code uses the value you set:
549549
550* When the main conversation is in `bypassPermissions`, `acceptEdits`, or [auto mode](/docs/en/permission-modes#eliminate-prompts-with-auto-mode), the subagent runs in that same mode and Claude Code ignores the `permissionMode` you set. Under auto mode, the classifier evaluates the subagent's tool calls with the main conversation's block and allow rules.
550* When the main conversation is in `bypassPermissions`, `acceptEdits`, or [auto mode](/docs/en/permission-modes#eliminate-prompts-with-auto-mode), the subagent runs in that same mode and Claude Code ignores the `permissionMode` you set. Under auto mode, the classifier evaluates the subagent's tool calls with the main conversation's block and allow rules. When the subagent finishes, the classifier also reviews its work and its final report before the report is delivered, as [How auto mode handles subagents](/docs/en/permission-modes#eliminate-prompts-with-auto-mode) describes.
551551* When the main conversation is in `default`, `dontAsk`, or `plan` mode, the subagent runs in the permission mode you set, except `bypassPermissions`. A subagent that declares `bypassPermissions` keeps the main conversation's mode instead. The `bypassPermissions` exception requires Claude Code v2.1.267 or later.
552552
553553`permissionMode` accepts these values, and `manual` as an alias for `default`:
from line 986
986986
987987By default, a subagent can spawn subagents of its own, up to three layers below the main conversation. At the depth limit, Claude Code withholds the `Agent` tool from every subagent except a [fork](#fork-the-current-conversation), so a subagent at the limit does its delegated work itself and returns one summary. A fork at the limit keeps `Agent` in its inherited tool list, but the tool returns an error instead of spawning.
988988
989Nested subagents suit a delegated task that itself splits into parallel subtasks, such as a reviewer subagent that dispatches a verifier per finding, so the intermediate output never reaches your main conversation. Only the top-level subagent's summary returns to you.
989Nested subagents suit a delegated task that itself splits into parallel subtasks, such as a reviewer subagent that dispatches a verifier per finding. In an interactive session, only the top-level subagent's summary returns to you and the intermediate output stays out of your main conversation: a subagent that launches background subagents waits for their results before it finishes. In [non-interactive mode](/docs/en/headless) and the Agent SDK, the launching subagent doesn't wait, so a nested background subagent that finishes after its launcher has ended reports to your main conversation instead.
990990
991991To change the limit, set [`CLAUDE_CODE_MAX_SUBAGENT_SPAWN_DEPTH`](/docs/en/env-vars) to the number of subagent layers you want below your main conversation. For example, this entry in [`settings.json`](/docs/en/settings) caps nesting at two layers:
992992
tools-reference Changed · +3 / -2 lines
from line 46
4646| `SendUserFile` | Sends files from the session to you with an optional caption, so a generated report, diagram, screenshot, or built artifact reaches your device instead of only being mentioned in the transcript. As of v2.1.196, the optional `display` input controls presentation: `render` opens the file inline in the client, `attach` shows a download card only, and when unset the client decides by file type. Available when a [Remote Control](/docs/en/remote-control) client is connected or the session runs in a managed cloud environment such as [Claude Code on the web](/docs/en/claude-code-on-the-web). Delivery runs through Anthropic-hosted infrastructure, so the tool is not available on Amazon Bedrock, Google Cloud's Agent Platform, or Microsoft Foundry | No |
4747| `ShareOnboardingGuide` | Uploads `ONBOARDING.md` and returns a share link teammates can open in Claude Code. Called from `/team-onboarding` after the guide is written. Available to claude.ai subscribers on Pro, Max, Team, and Enterprise plans | Yes |
4848| `Skill` | Executes a [skill](/docs/en/skills#control-who-invokes-a-skill) within the main conversation | Yes |
49| `SubagentHandback` | Delivers a subagent's final report to whichever conversation receives that subagent's result. Provided only in [auto mode](/docs/en/permission-modes#eliminate-prompts-with-auto-mode), to subagents that the Agent tool runs locally other than [forks](/docs/en/sub-agents#fork-the-current-conversation), and available in the terminal CLI, IDE extensions, web sessions, and the Agent SDK; the classifier reviews the report before it's delivered. Requires Claude Code v2.1.271 or later | No |
4950| `TaskCreate` | Creates a new task in the task list. Provided by default only on the models listed under [Task tool availability](#task-tool-availability), and on other models when you opt in | No |
5051| `TaskGet` | Retrieves full details for a specific task. Provided by default only on the models listed under [Task tool availability](#task-tool-availability), and on other models when you opt in | No |
5152| `TaskList` | Lists all tasks with their current status. Provided by default only on the models listed under [Task tool availability](#task-tool-availability), and on other models when you opt in | No |
from line 92
9192
9293## Agent tool behavior
9394
94The Agent tool spawns a subagent in a separate context window. The subagent works through its task autonomously, then returns a single text result to the parent conversation. The parent doesn't see the subagent's intermediate tool calls or outputs, only that final result. With [agent teams](/docs/en/agent-teams) enabled, a call that carries a `name` can launch a [teammate](/docs/en/agent-teams#how-claude-starts-agent-teams) instead, which reports back through team messages rather than by returning a result.
95The Agent tool spawns a subagent in a separate context window. The subagent works through its task autonomously, then returns its result to the parent conversation. The parent doesn't see the subagent's intermediate tool calls or outputs, only that final result. With [agent teams](/docs/en/agent-teams) enabled, a call that carries a `name` can launch a [teammate](/docs/en/agent-teams#how-claude-starts-agent-teams) instead, which reports back through team messages rather than by returning a result.
9596
9697To cap how many turns a subagent runs, set `maxTurns` in the [subagent definition](/docs/en/sub-agents#supported-frontmatter-fields). When the subagent reaches the limit, Claude Code marks the returned result as partial output, and Claude can [resume the subagent](/docs/en/sub-agents#resume-subagents) to continue.
9798
from line 105
104105* **`disallowedTools` only**: the subagent gets every parent tool except the listed ones.
105106* **Both set**: `disallowedTools` takes precedence. A tool listed in both is removed.
106107
107In every case, the resolved set is limited to the [tools available to subagents](/docs/en/sub-agents#available-tools): a tool that isn't available to subagents is never granted, even when listed in `tools`.
108In every case, the resolved set is limited to the [tools available to subagents](/docs/en/sub-agents#available-tools): a tool that isn't available to subagents is never granted, even when listed in `tools`. Where the conditions in the `SubagentHandback` tools-table entry hold, Claude Code also gives the subagent that tool, even if you leave it out of `tools` or list it in `disallowedTools`.
108109
109110If every entry in a subagent's `tools` list fails to match a usable tool, the Agent tool usually returns an error naming the entries instead of launching the subagent; see [Agent would be spawned with zero tools](/docs/en/errors#agent-would-be-spawned-with-zero-tools) for the message and how to fix each entry.
110111
permission-modes Changed · +1 / -1 lines
from line 474
474474
475475 1. Before a subagent starts, the delegated task description is evaluated, so a dangerous-looking task is blocked at spawn time.
476476 2. While the subagent runs, each of its actions goes through the classifier with the same rules as the parent session, and any `permissionMode` in the subagent's frontmatter is ignored.
477 3. When the subagent finishes, the classifier reviews its full action history; if that return check flags a concern, a security warning is prepended to the subagent's results. When a separate API safety check refuses the review request itself, Claude Code still returns the subagent's results, prepended with a warning that the work is unreviewed and should be treated as untrusted.
477 3. When the subagent finishes, the classifier reviews its work and its final report before the parent reads the report. When the classifier flags the subagent's work or report, or a separate API safety check refuses the review, the report is still delivered, prepended with a security warning. When the classifier is unavailable for the review, the report arrives with a note to verify the subagent's work before acting on it.
478478
479479 Step 1 requires Claude Code v2.1.178 or later. Earlier versions applied the classifier at steps 2 and 3, but did not evaluate the task description before the subagent started.
480480 </Accordion>