| `CLAUDE_CODE_SIMPLE` | Enable simple mode (Bash + Edit tools only). Also configurable via `env` key — see [Settings Reference](./claude-settings.md#environment-variables-via-env) |
| `CLAUDE_BASH_NO_LOGIN=1` | Skip login shell for BashTool |
For environment variables configurable via the `"env"` key in `settings.json` (including `MAX_THINKING_TOKENS`, `CLAUDE_CODE_SHELL`, `CLAUDE_CODE_ENABLE_TASKS`, `CLAUDE_CODE_DISABLE_BACKGROUND_TASKS`, `CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS`, and more), see the [Claude Settings Reference](./claude-settings.md#environment-variables-via-env).
Claude Code commands — frontmatter fields and official built-in slash commands.
@@ -83,7 +83,7 @@ Claude Code commands — frontmatter fields and official built-in slash commands
| 43 | `/plan [description]` |  | Enter plan mode directly from the prompt. Pass an optional description to enter plan mode and immediately start with that task, for example `/plan fix the auth bug` |
| 44 | `/add-dir <path>` |  | Add a new working directory to the current session |
| 45 | `/diff` |  | Open an interactive diff viewer showing uncommitted changes and per-turn diffs. Use left/right arrows to switch between the current git diff and individual Claude turns, and up/down to browse files |
| 46 | `/init` |  | Initialize project with a `CLAUDE.md` guide. Set `CLAUDE_CODE_NEW_INIT=true` for an interactive flow that also walks through skills, hooks, and personal memory files |
| 46 | `/init` |  | Initialize project with a `CLAUDE.md` guide. Set `CLAUDE_CODE_NEW_INIT=1` for an interactive flow that also walks through skills, hooks, and personal memory files |
| 47 | `/pr-comments [PR]` |  | Fetch and display comments from a GitHub pull request. Automatically detects the PR for the current branch, or pass a PR URL or number. Requires the `gh` CLI |
| 49 | `/security-review` |  | Analyze pending changes on the current branch for security vulnerabilities. Reviews the git diff and identifies risks like injection, auth issues, and data exposure |
A comprehensive guide to all available configuration options in Claude Code's `settings.json` files. As of v2.1.88, Claude Code exposes **60+ settings** and **100+ environment variables** (use the `"env"` field in `settings.json` to avoid wrapper scripts).
A comprehensive guide to all available configuration options in Claude Code's `settings.json` files. As of v2.1.89, Claude Code exposes **60+ settings** and **170+ environment variables** (use the `"env"` field in `settings.json` to avoid wrapper scripts).
<table width="100%">
<tr>
@@ -68,7 +68,7 @@ Within the managed tier, precedence is: server-managed > MDM/OS-level policies >
| `model` | string | `"default"` | Override default model. Accepts aliases (`sonnet`, `opus`, `haiku`) or full model IDs |
| `agent` | string | - | Set the default agent for the main conversation. Value is the agent name from `.claude/agents/`. Also available via `--agent` CLI flag |
| `cleanupPeriodDays` | number | `30` | Sessions inactive longer than this are deleted at startup. Setting to `0` deletes all existing transcripts and disables session persistence entirely (no `.jsonl` files written, `/resume` shows no conversations, hooks receive an empty `transcript_path`) |
| `cleanupPeriodDays` | number | `30` | Sessions inactive longer than this are deleted at startup. As of v2.1.89, setting to `0` is rejected with a validation error *(note: official settings page still describes the old `0`-value behavior — "deletes all existing transcripts and disables session persistence entirely" — but the changelog confirms this is now invalid)* |
| `alwaysThinkingEnabled` | boolean | `false` | Enable extended thinking by default for all sessions |
| `skipWebFetchPreflight` | boolean | `false` | Skip WebFetch blocklist check before fetching URLs *(in JSON schema, not on official settings page)* |
@@ -79,6 +79,7 @@ Within the managed tier, precedence is: server-managed > MDM/OS-level policies >
| `voiceEnabled` | boolean | - | Enable push-to-talk voice dictation. Written automatically when you run `/voice`. Requires a Claude.ai account |
| `showClearContextOnPlanAccept` | boolean | `false` | Show the "clear context" option on the plan accept screen. Set to `true` to restore the option (hidden by default since v2.1.81) |
| `disableDeepLinkRegistration` | string | - | Set to `"disable"` to prevent Claude Code from registering the `claude-cli://` protocol handler with the operating system on startup. Deep links let external tools open a Claude Code session with a pre-filled prompt via `claude-cli://open?q=...`. Useful in environments where protocol handler registration is restricted or managed separately |
| `showThinkingSummaries` | boolean | `false` | Show thinking summaries in interactive sessions. As of v2.1.89, thinking summaries are no longer generated by default — set to `true` to restore them *(in v2.1.89 changelog, not yet on official settings page)* |
| `feedbackSurveyRate` | number | - | Probability (0–1) that the session quality survey appears when eligible. Enterprise admins can control how often the survey is shown. Example: `0.05` = 5% of eligible sessions |
**Example:**
@@ -221,6 +222,7 @@ Control what tools and operations Claude can perform.
| `permissions.additionalDirectories` | array | Extra directories Claude can access |
| `permissions.defaultMode` | string | Default permission mode. In Remote environments, only `acceptEdits` and `plan` are honored (v2.1.70+) |
| `permissions.skipDangerousModePermissionPrompt` | boolean | Skip the confirmation prompt shown before entering bypass permissions mode via `--dangerously-skip-permissions` or `defaultMode: "bypassPermissions"`. Ignored when set in project settings (`.claude/settings.json`) to prevent untrusted repositories from auto-bypassing the prompt |
| `allowManagedPermissionRulesOnly` | boolean | **(Managed only)** Only managed permission rules apply; user/project `allow`, `ask`, `deny` rules are ignored |
| `allow_remote_sessions` | boolean | **(Managed only)** Allow users to start Remote Control and web sessions. Defaults to `true`. Set to `false` to prevent remote session access *(not in official docs — official permissions page states "Access to Remote Control and web sessions is not controlled by a managed settings key." On Team and Enterprise plans, admins enable/disable via [Claude Code admin settings](https://claude.ai/admin-settings/claude-code))* |
| `autoMode` | object | Customize what the [auto mode](/en/permission-modes#eliminate-prompts-with-auto-mode) classifier blocks and allows. Contains `environment` (trusted infrastructure descriptions), `allow` (exceptions to block rules), and `soft_deny` (block rules) — all arrays of prose strings. **Not read from shared project settings** (`.claude/settings.json`) to prevent repo injection. Available in user, local, and managed settings. Setting `allow` or `soft_deny`**replaces** the entire default list for that section. Run `claude auto-mode defaults` to see built-in rules before customizing |
@@ -237,7 +239,7 @@ Control what tools and operations Claude can perform.
| `"dontAsk"` | Auto-denies tools unless pre-approved via `/permissions` or `permissions.allow` rules |
| `"viewOnly"` | Read-only mode, no modifications *(not in official docs — unverified)* |
| `"bypassPermissions"` | Skip all permission checks (dangerous) |
| `"auto"` | Background classifier replaces manual prompts. Research preview — requires Team plan + Sonnet/Opus 4.6. Classifier auto-approves read-only and file edits; sends everything else through a safety check. Falls back to prompting after 3 consecutive or 20 total blocks. Configure with `autoMode` setting |
| `"auto"` | Background classifier replaces manual prompts (`--enable-auto-mode`). Research preview — requires Team plan + Sonnet/Opus 4.6. Classifier auto-approves read-only and file edits; sends everything else through a safety check. Falls back to prompting after 3 consecutive or 20 total blocks. Configure with `autoMode` setting |
| `"plan"` | Read-only exploration mode |
### Tool Permission Syntax
@@ -677,6 +679,10 @@ Set environment variables for all Claude Code sessions.
| `ANTHROPIC_API_KEY` | API key for authentication |
| `ANTHROPIC_VERTEX_PROJECT_ID` | GCP project ID for Vertex AI |
| `ANTHROPIC_CUSTOM_MODEL_OPTION` | Model ID to add as a custom entry in the `/model` picker. Use to make a non-standard or gateway-specific model selectable without replacing built-in aliases |
| `ANTHROPIC_CUSTOM_MODEL_OPTION_NAME` | Display name for the custom model entry in the `/model` picker. Defaults to the model ID when not set |
| `ANTHROPIC_CUSTOM_MODEL_OPTION_DESCRIPTION` | Display description for the custom model entry in the `/model` picker. Defaults to `Custom model (<model-id>)` when not set |
@@ -696,6 +702,7 @@ Set environment variables for all Claude Code sessions.
| `DISABLE_TELEMETRY` | Disable telemetry (`1` to disable) |
| `MCP_TIMEOUT` | MCP startup timeout in ms |
| `MAX_MCP_OUTPUT_TOKENS` | Max MCP output tokens (default: 25000). Warning displayed when output exceeds 10,000 tokens |
| `API_TIMEOUT_MS` | Timeout in ms for API requests (default: 600000) |
| `BASH_MAX_TIMEOUT_MS` | Bash command timeout |
| `BASH_MAX_OUTPUT_LENGTH` | Max bash output length |
| `CLAUDE_AUTOCOMPACT_PCT_OVERRIDE` | Auto-compact threshold percentage (1-100). Default is ~95%. Set lower (e.g., `50`) to trigger compaction earlier. Values above 95% have no effect. Use `/context` to monitor current usage. Example: `CLAUDE_AUTOCOMPACT_PCT_OVERRIDE=50 claude` |
@@ -706,12 +713,19 @@ Set environment variables for all Claude Code sessions.
| `CLAUDE_CODE_GLOB_HIDDEN` | Include dotfiles in Glob results (`1` to enable) |
| `CLAUDE_CODE_GLOB_NO_IGNORE` | Make Glob ignore `.gitignore` patterns (`1` to enable) |
| `CLAUDE_CODE_GLOB_TIMEOUT_SECONDS` | Timeout in seconds for Glob file discovery |
| `CLAUDE_CODE_ENABLE_TASKS` | Set to `true` to enable task tracking in non-interactive mode (`-p` flag). Tasks are on by default in interactive mode |
| `CLAUDE_CODE_SIMPLE` | Set to `1` to run with a minimal system prompt and only the Bash, file read, and file edit tools |
| `CLAUDE_CODE_ACCOUNT_UUID` | Override account UUID for authentication |
| `CLAUDE_CODE_DISABLE_GIT_INSTRUCTIONS` | Disable git-related system prompt instructions |
@@ -728,7 +742,13 @@ Set environment variables for all Claude Code sessions.
| `CLAUDE_CODE_MAX_OUTPUT_TOKENS` | Max output tokens per response. Default: 32,000 (64,000 for Opus 4.6 as of v2.1.77). Upper bound: 64,000 (128,000 for Opus 4.6 and Sonnet 4.6 as of v2.1.77) |
| `CLAUDE_CODE_DISABLE_FAST_MODE` | Disable fast mode entirely (`1` to disable) |
| `CLAUDE_CODE_DISABLE_NONSTREAMING_FALLBACK` | Set to `1` to disable the non-streaming fallback when a streaming request fails mid-stream. Streaming errors propagate to the retry layer instead. Useful when a proxy or gateway causes the fallback to produce duplicate tool execution (v2.1.83) |
| `CLAUDE_ENABLE_STREAM_WATCHDOG` | Abort stalled streams (`1` to enable) |
| `CLAUDE_CODE_SUBAGENT_MODEL` | Override model for subagents (e.g., `haiku`, `sonnet`) |
| `CLAUDE_CODE_SUBPROCESS_ENV_SCRUB` | Set to `1` to strip Anthropic and cloud provider credentials from subprocess environments (Bash tool, hooks, MCP stdio servers). Use for defense-in-depth when subprocesses should not inherit API keys (v2.1.83) |
| `CLAUDE_CODE_MAX_TOOL_USE_CONCURRENCY` | Max parallel read-only tools (default: 10) |
| `CLAUDE_AGENT_SDK_DISABLE_BUILTIN_AGENTS` | Disable built-in subagent types in SDK mode (`1` to disable) |
| `CLAUDE_AGENT_SDK_MCP_NO_PREFIX` | Skip `mcp__<server>__` prefix for MCP tools in SDK mode (`1` to enable) |
| `MCP_CONNECTION_NONBLOCKING` | Set to `true` in `-p` mode to skip the MCP connection wait entirely. Bounds `--mcp-config` server connections at 5s instead of blocking on the slowest server *(in v2.1.89 changelog, not yet on official env-vars page)* |
| `CLAUDE_CODE_SESSIONEND_HOOKS_TIMEOUT_MS` | SessionEnd hook timeout in ms (replaces hard 1.5s limit) |
| `CLAUDE_CODE_DISABLE_TERMINAL_TITLE` | Disable terminal title updates (`1` to disable) |
| `CLAUDE_CODE_NO_FLICKER` | Set to `1` to enable flicker-free alt-screen rendering. Eliminates visual flicker during fullscreen redraws (v2.1.88) |
| `CLAUDE_CODE_SCROLL_SPEED` | Mouse wheel scroll multiplier for fullscreen rendering. Increase for faster scrolling, decrease for finer control |
| `CLAUDE_CODE_DISABLE_MOUSE` | Set to `1` to disable mouse tracking in fullscreen rendering. Useful when mouse events interfere with terminal multiplexers or accessibility tools |
| `CLAUDE_CODE_ACCESSIBILITY` | Set to `1` to keep native terminal cursor visible for screen readers and accessibility tools |
| `CLAUDE_CODE_SYNTAX_HIGHLIGHT` | Set to `0` to disable syntax highlighting in diff output |
| `CLAUDE_CODE_IDE_SKIP_AUTO_INSTALL` | Skip automatic IDE extension installation (`1` to skip) |
| `CLAUDE_CODE_AUTO_CONNECT_IDE` | Override auto IDE connection behavior |
| `CLAUDE_CODE_IDE_HOST_OVERRIDE` | Override IDE host address for connection |
| `CLAUDE_CODE_IDE_SKIP_VALID_CHECK` | Skip IDE lockfile validation (`1` to skip) |
| `CLAUDE_CODE_OTEL_HEADERS_HELPER_DEBOUNCE_MS` | Debounce interval in ms for OTel headers helper script |
| `CLAUDE_CODE_OTEL_FLUSH_TIMEOUT_MS` | Timeout in ms for OpenTelemetry flush |
| `CLAUDE_CODE_OTEL_SHUTDOWN_TIMEOUT_MS` | Timeout in ms for OpenTelemetry shutdown |
| `CLAUDE_STREAM_IDLE_TIMEOUT_MS` | Timeout in ms before the streaming idle watchdog closes a stalled connection. Default: `90000` (90 seconds). Increase if long-running tools or slow networks cause premature timeout errors |
| `OTEL_LOG_TOOL_DETAILS` | Set to `1` to include `tool_parameters` in OpenTelemetry events. Omitted by default for privacy *(in v2.1.85 changelog, not yet on official env-vars page)* |
| `CLAUDE_CODE_MCP_SERVER_NAME` | Name of the MCP server, passed as an environment variable to `headersHelper` scripts so they can generate server-specific authentication headers *(in v2.1.85 changelog, not yet on official env-vars page)* |
@@ -780,6 +828,8 @@ Set environment variables for all Claude Code sessions.
| `ANTHROPIC_DEFAULT_SONNET_MODEL_SUPPORTED_CAPABILITIES` | Override capability detection for a pinned Sonnet model. Comma-separated values (e.g., `effort,thinking`). Required when the pinned model supports features the auto-detection cannot confirm |
| `MAX_THINKING_TOKENS` | Maximum extended thinking tokens per response |
| `CLAUDE_CODE_AUTO_COMPACT_WINDOW` | Set the context capacity in tokens used for auto-compaction calculations. Defaults to the model's context window (200K standard, 1M for extended context models). Use a lower value (e.g., `500000`) on a 1M model to treat it as 500K for compaction. Capped at actual context window. `CLAUDE_AUTOCOMPACT_PCT_OVERRIDE` is applied as a percentage of this value. Setting this decouples the compaction threshold from the status line's `used_percentage` |
| `DISABLE_AUTO_COMPACT` | Disable automatic context compaction (`1` to disable). Manual `/compact` still works |
| `DISABLE_COMPACT` | Disable all compaction — both automatic and manual (`1` to disable) |
Claude Code subagents — frontmatter fields and official built-in agent types.
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Shayan Rais