Complete reference for settings.json configuration and environment variables. Covers confirmed settings through Claude Code v2.1.310 (last reviewed October 27, 2026). The official settings reference remains authoritative.
Claude Code uses four settings scopes, applied from highest to lowest priority:
Priority
Scope
Location
Shared?
Purpose
1
Managed
Server, MDM profile, registry, or system managed-settings.json
Yes (IT-deployed)
Org-wide policies, cannot be overridden
2
Command line
-- flags at startup
No
Temporary session overrides
3
Local
.claude/settings.local.json
No (gitignored)
Personal project-specific
4
Project
.claude/settings.json
Yes (committed)
Team-shared settings
5
User
~/.claude/settings.json
No
Global personal defaults
Array merging: Settings like permissions.allow, sandbox.filesystem.allowWrite, and allowedHttpHookUrls are concatenated and deduplicated across scopes, not replaced.
Deny precedence:permissions.deny rules always take effect regardless of allow/ask rules at any scope.
Managed settings delivery methods:
Server-managed (Claude.ai admin console)
macOS MDM: com.anthropic.claudecode plist
Windows registry: HKLM\SOFTWARE\Policies\ClaudeCode
Other config:~/.claude.json stores OAuth session, MCP server configs, per-project trust state, and preferences like editorMode. Do not put ~/.claude.json keys into settings.json: it will trigger schema validation errors.
JSON Schema URL for IDE validation and autocomplete. Add "https://json.schemastore.org/claude-code-settings.json" to enable inline validation in VS Code, Cursor, and other editors.
Override the default model for all sessions. Accepts aliases ("sonnet", "opus", "haiku", "opusplan") or full model IDs like "claude-sonnet-4-6". The ANTHROPIC_MODEL environment variable takes precedence.
Run the main thread as a named subagent. Applies that agent’s system prompt, tool restrictions, and model. Value must match an agent defined in .claude/agents/. Also available via --agent CLI flag.
Sessions inactive longer than this number of days are deleted at startup. Setting to 0 deletes all existing transcripts at startup and disables session persistence entirely: no .jsonl files are written, /resume shows no conversations, and hooks receive an empty transcript_path.
Include built-in commit and PR workflow instructions and a git status snapshot in the system prompt. Set to false when using a custom git workflow skill. The CLAUDE_CODE_DISABLE_GIT_INSTRUCTIONS env var takes precedence.
Restrict which models users can select via /model, --model, Config tool, or ANTHROPIC_MODEL. Does not affect the Default option unless enforceAvailableModels is enabled.
When true, fast mode does not persist across sessions. Each session starts with fast mode off, requiring users to enable it with /fast. The user’s preference is still saved.
Environment variables applied to every session. Use this instead of wrapper scripts to set variables. See the Environment Variables section for all supported keys.
Type: string
Scope: user / local / managed
Default: none
Custom directory for auto-memory storage. Accepts ~/-expanded paths. Not accepted in project settings (.claude/settings.json) to prevent shared repos from redirecting memory writes to sensitive locations.
Type: string
Scope: all
Default: unset (built-in starting mode for the surface: auto in interactive terminal and VS Code sessions since v2.1.283)
Values:"default" (alias "manual") | "acceptEdits" | "plan" | "auto" | "dontAsk" | "bypassPermissions"
Default permission mode when opening Claude Code. auto and bypassPermissions take effect only from user or managed settings (or --permission-mode), not from project or local settings. In Remote environments, only "acceptEdits" and "plan" are honored.
Type: string
Scope: all
Default: none
Values:"disable"
Set to "disable" to prevent bypassPermissions mode from being activated. Disables the --dangerously-skip-permissions flag. Most useful in managed settings.
Rules follow the format Tool or Tool(specifier). Evaluation order: deny first, then ask, then allow. The first matching rule wins.
Tool
Pattern
Example
Bash
Command pattern with wildcards
Bash(npm run *), Bash(git *)
Read
File path pattern
Read(.env), Read(./secrets/**)
Edit
File path pattern
Edit(src/**), Edit(*.ts)
Write
File path pattern
Write(*.md)
WebFetch
domain:hostname
WebFetch(domain:example.com)
WebSearch
No specifier
WebSearch
Task
Agent name
Task(Explore)
Agent
Agent name
Agent(researcher)
MCP
mcp__server__tool or MCP(server:tool)
mcp__memory__*
Path prefixes for Read/Edit rules:
Prefix
Meaning
//
Absolute path from filesystem root
~/
Relative to home directory
/
Relative to project root
./ or none
Relative path (current directory)
Bash wildcard notes:* matches at any position. Bash(ls *) (space before *) matches ls -la but NOT lsof. Bash(*) is equivalent to Bash (matches all commands). The legacy :* suffix (e.g., Bash(npm:*)) is deprecated.
{
"permissions": {
"allow": ["Edit(*)", "Bash(npm run *)", "Bash(git *)"],
Configure custom commands to run at lifecycle events. See the hooks documentation and §7.1 of the guide for the full set of 33 hook events, exit codes, and environment variables.
Type: array of strings
Scope: all
Default: none (no restriction)
Allowlist of URL patterns that HTTP hooks may target. Supports * as a wildcard. When defined, hooks with non-matching URLs are silently blocked. Empty array blocks all HTTP hooks. Arrays merge across settings sources.
Type: array of strings
Scope: all
Default: none (no restriction)
Allowlist of environment variable names that HTTP hooks can interpolate into header values. Each hook’s effective allowedEnvVars is the intersection with this list. Arrays merge across settings sources.
Type: array
Scope: managed only
Default: none (no restrictions)
Allowlist of MCP servers users can configure. Each entry matches by serverName, serverCommand, or serverUrl. Undefined = no restrictions, empty array = lockdown.
When true, only allowedMcpServers from managed settings are respected. Users can still add MCP servers, but only admin-defined servers are usable. deniedMcpServers still merges from all sources.
Allow channels for claude.ai Team and Enterprise users. When unset or false, channel message delivery is blocked regardless of what users pass to --channels. Console API-key deployments use Console defaults and managed controls instead; do not infer their policy from this Team/Enterprise-only setting.
Allowlist of channel plugins that may push messages. Replaces the default Anthropic allowlist when set. Requires channelsEnabled: true. Empty array blocks all channel plugins.
Security boundary: this allowlist controls which Channel plugins may deliver messages. It does not grant a delivered message, its sender, or the plugin any Bash, filesystem, GitHub, or other tool permission. Gate senders and treat Channel content as untrusted input, including when a plugin can relay a permission prompt.
Exit with an error at startup if sandbox.enabled is true but the sandbox cannot start (missing dependencies, unsupported platform). When false, a warning is shown and commands run unsandboxed. Useful in managed deployments that require sandboxing as a hard gate.
Auto-approve bash commands when sandboxed. When the sandbox is active, bash commands that would normally require confirmation are automatically approved.
Commands that bypass the sandbox and run directly in your environment.
Use the glob form, not the bare name. A bare entry such as "git" matches only the zero-argument string git, so it never fires on a real invocation and the command stays sandboxed. Write "git *" instead. The bare form is what the published JSON schema suggests, which is why this is easy to get wrong: you configure something that does nothing, notice the command is still sandboxed, and only then find the glob form (anthropics/claude-code#10524). Verified on 2.1.220: "git" had no effect, "git *" worked immediately.
Every command in a call must match (v2.1.277+). Claude Code checks entries against the whole command line it sends. npm ci && docker compose build stays sandboxed unless another entry covers npm ci, and a redirect to a file, a cd, a loop or a $(...) substitution keeps the whole call sandboxed. Before v2.1.277, one matching command unsandboxed the entire call, including commands that ran before it (anthropics/claude-code#81157). Entries match the text of the call: git -C <path> push, /usr/local/bin/docker and a script that calls the tool internally do not match.
An excluded command runs with no filesystem limit and no proxy, so it can reach any host, not only allowedDomains. Scope entries to the subcommands that genuinely need to leave the sandbox rather than the whole binary. Git over SSH on macOS is the common case (on Linux and WSL2 it works sandboxed once the host is allowed): only the network operations need an exception, so list those and leave local git confined.
When the sandbox is admin-required (allowUnsandboxedCommands: false from managed settings or --settings, or managed allowManagedDomainsOnly), entries in a repository’s .claude/settings.json and .claude/settings.local.json are ignored (v2.1.282+).
Allow commands to opt out of the sandbox via the dangerouslyDisableSandbox parameter. Set to false for strict sandboxing where all commands must run inside the sandbox or be in excludedCommands. The /sandbox Overrides tab shows false as Strict sandbox mode.
Since v2.1.285, a false in user settings, --settings or managed settings holds even when a project sets true (before, the project’s true won). A false from managed settings or --settings also makes the sandbox admin-required: Claude Code then ignores the repository settings that would loosen it, excludedCommands included. See admin-required sandbox.
(macOS only) Allow access to the system TLS trust service (com.apple.trustd.agent). Required for Go-based tools like gh, gcloud, and terraform when using httpProxyPort with a MITM proxy and custom CA. Reduces security.
Specific Unix socket paths accessible in the sandbox on macOS (for SSH agents, local databases, etc.). Ignored on Linux and WSL2, where the seccomp filter cannot inspect socket paths; use allowAllUnixSockets only after accepting its broader boundary. The cross-session inbox case does not require this exception for normal SendMessage calls.
Allow all Unix socket connections in the sandbox. Overrides allowUnixSockets. On Linux and WSL2, this is the only way to permit Unix sockets because it skips the seccomp filter that otherwise blocks socket(AF_UNIX, ...) calls.
Allow binding to localhost ports (macOS only). Sandboxed commands can then listen on network ports and connect to any port on localhost, which includes every other service listening there; an unauthenticated local service such as a debugger can then act for the command outside the sandbox. On Linux and WSL2 a sandboxed command’s localhost is private to it and this key has no effect.
When true, only allowedDomains and WebFetch(domain:...) allow rules from managed settings are respected. Non-allowed domains are blocked without prompting. Denied domains are still respected from all sources.
Domains sandboxed commands can never reach, even when a broader allowedDomains wildcard would permit them. A denied host is refused in every permission mode, and a per-command allowed_domains list in auto mode cannot open it. Set in managed settings or --settings, it also stops a repository from setting proxy ports (v2.1.285+). IPv6 addresses use the bracketed form, "[::1]" or "[::1]:443" (v2.1.229+).
Additional paths where sandboxed commands can write. Merged across all settings scopes. Also merged with paths from Edit(...) allow permission rules.
Path prefix conventions:/ = absolute, ~/ = home-relative, ./ or no prefix = project-relative in project settings / ~/.claude-relative in user settings.
Paths to re-allow read access within denyRead regions. Takes precedence over denyRead. Arrays merge across all settings scopes.
When read rules overlap, the more specific path wins in both directions: denyRead: ["~/"] with allowRead: ["~/projects"] opens only that subtree, and allowRead: ["~/"] with denyRead: ["~/.env"] keeps that one file blocked. A broad allow cannot silently re-expose a secret an exact deny covers.
Skip filesystem isolation while keeping network isolation. Sandboxed commands get unrestricted read and write access to the host, and their egress stays confined to allowedDomains. Use when the point of sandboxing is controlling where commands connect rather than what they write.
Turning the layer off also disables filesystem.denyRead, credentials.files, and the protection on Claude Code’s own settings files, and stops the $TMPDIR override. credentials.envVars still applies, since environment scrubbing is independent of the filesystem layer.
Project settings cannot set it, so a checked-out repository cannot switch filesystem isolation off. When managed settings configure sandbox.filesystem at all, or list any credentials.files entry, only managed settings can set it. CLAUDE_CODE_SUBPROCESS_ENV_SCRUB makes Claude Code ignore it from every source.
With the filesystem layer off and commands auto-allowed, a sandboxed command can write shell startup files, executables on $PATH, or ~/.claude/settings.json, and use them to widen its own access on the next run. Set it only for workloads you trust not to escalate.
Type: array of { path, mode }Scope: all (deny narrows only, so any scope may add one)
Default:[]Since: v2.1.187
Credential files to hide from sandboxed commands. mode accepts "deny", which blocks reads the same way filesystem.denyRead does, and "mask" (v2.1.221+). With mask, on Linux and WSL2 sandboxed commands read a sentinel copy and the proxy substitutes the real value on egress; on macOS the file is simply unreadable, the same effect as deny, so a tool that authenticates with it stops working inside the sandbox. Partial masking uses extract (a regex marking the secret), decode: "jwt" with maskClaims, onExtractNoMatch and maskDuplicates (v2.1.224+). The default onExtractNoMatch: "warn" skips the entry when the pattern matches nothing, leaving the real file readable; use deny when the secret may be present but the pattern may miss it. A mask entry on a directory, a glob, a file over 8 MiB or a non-UTF-8 file falls back to deny. mask is honored only from user, managed and --settings settings.
The sandbox’s default read policy covers the entire machine, and there is no built-in credential denylist, so ~/.ssh and ~/.aws/credentials are readable by every sandboxed command until you list them. Paths follow the same prefix rules as sandbox.filesystem.*, and deny entries merge across every scope: any scope can add one, no scope can remove one another scope added.
{
"sandbox": {
"credentials": {
"files": [
{ "path": "~/.ssh", "mode": "deny" },
{ "path": "~/.aws", "mode": "deny" },
{ "path": "~/.gnupg", "mode": "deny" },
{ "path": "~/.config/gh", "mode": "deny" }
]
}
}
}
Verified on 2.1.220: with the entry in place, ls ~/.ssh from a sandboxed command returns Operation not permitted while the directory still appears in a listing of the home directory.
Type: array of { name, mode, injectHosts? }Scope:deny from all scopes; mask from user, managed, --settings only
Default:[]Since: v2.1.187 (deny), v2.1.199 (mask)
Environment variables to withhold from sandboxed commands. Sandboxed commands otherwise inherit the parent environment as-is, credentials included, which is the gap a credentials.files entry alone leaves open.
deny unsets the variable before each sandboxed command runs. It affects sandboxed Bash only, so MCP servers keep their credentials: they run as separate processes, not as sandboxed commands. Wildcards such as AWS_* are not supported; list each variable. A shell startup file can undo it: zsh reads ~/.zshenv at every start, so a variable exported there is set again inside the sandboxed command, as measured on v2.1.302 (see Default behavior for the probe and the CLAUDECODE guard).
For structured values such as a DATABASE_URL or a JWT, extract, decode, maskClaims and onExtractNoMatch keep parsing tools working (v2.1.224+). AWS SigV4 requests need AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY masked together so the proxy can re-sign them (v2.1.221+); group differently named variables with credentials.awsPairs, and use credentials.sigv4 to forward streaming uploads, presigned URLs or SigV4A requests the proxy cannot re-sign (both v2.1.224+). claude doctor warns when TLS termination is unavailable or an injectHosts entry can never match.
mask protects the credential while keeping the tool that authenticates with it working. The command sees a per-session sentinel; when a request leaves the sandbox for one of the credential’s injectHosts, the proxy substitutes the real value. The command and anything it logs never hold the real credential. Each injectHosts entry must itself be covered by allowedDomains, and network.tlsTerminate is required because the proxy has to see request contents to substitute. Without it, masking fails closed: the sentinel reaches the server unchanged and authentication fails, and Claude Code reports the misconfiguration at startup.
deny wins when the same variable appears with both modes in any scope. Start with deny and move a variable to mask only when a CLI actually breaks without it, since mask requires TLS termination and authorizes the proxy to send the real credential to the listed hosts.
To strip Anthropic and cloud provider credentials from all subprocesses regardless of sandboxing, set CLAUDE_CODE_SUBPROCESS_ENV_SCRUB instead.
Make the built-in proxy terminate TLS itself. Required by mask entries in credentials.envVars and credentials.files. It enables credential substitution; it does not add content filtering.
Deny sandboxed commands any host outside the allowlist instead of prompting. The allowlist is allowedDomains plus domains from WebFetch(domain:...) allow rules, or only the managed entries when allowManagedDomainsOnly is set. Applies to sandboxed commands only: in-process tools such as WebFetch still follow their permission rules.
Enable it last, once the domain list has survived a week of real work. Before that it converts every missing domain from a one-time prompt into a hard failure.
Allow Apple Events on macOS, which the sandbox blocks by default. Fixes open, osascript, and browser-based auth flows failing with error -600.
Enabling it removes code-execution isolation: sandboxed commands can launch other applications unsandboxed with no prompt, and send AppleScript to running applications, subject to the macOS automation-consent prompt. Prefer adding the specific command to excludedCommands.
Unix-socket exceptions are intentionally absent from this baseline. Granting /var/run/docker.sock would give sandboxed code control of the Docker daemon and an effective path to the host.
For marketplace publication and the <claude-code-hint /> recommendation boundary, see Plugin Distribution and Recommendation Hints. The settings below govern local and managed marketplace controls.
Custom message appended to the plugin trust warning shown before installation. Use for org-specific context like confirming plugins from an internal marketplace are vetted.
Fallback effort for a model without a saved choice. Defaults depend on the model: Opus 5.5, Sonnet 5.5 and Haiku 5.5 use medium; Fable 5.1 uses high. The legacy Haiku 4.5 has no effort parameter. Opus 5.5 ignores this legacy key in user settings, but explicit project, local, managed, and --settings values still apply.
Since v2.1.251, /effort saves the active model’s level here. Confirm with Enter to save or s for this session only. max is session-only and cannot be stored as effortLevel.
{
"modelSettings": {
"claude-opus-5-5": { "effortLevel": "high" }
}
}
Within one settings file, a model-specific level wins over the fallback. Across files, normal settings precedence applies. CLAUDE_CODE_EFFORT_LEVEL overrides the chosen level, subject to organization caps. /effort auto clears the saved choice for the active model.
Pair with availableModels to keep the /model Default choice inside that allowlist. Without this setting, the allowlist alone does not constrain Default. See the official model settings reference.
Map Anthropic model IDs to provider-specific model IDs (e.g., Bedrock inference profile ARNs). Each key is a model picker entry name; each value is the provider model ID.
Configure a custom status line. The command receives a JSON object on stdin with fields like context_window.used_percentage, rate_limits.five_hour.used_percentage, etc.
Configure a custom script for @ file path autocomplete. The command receives JSON on stdin with a query field and outputs newline-separated file paths (max 15).
Controls how Claude communicates throughout the session. Select with /output-style or /config. The setting value is case-sensitive; the command accepts case-insensitive names.
Built-in values:
"Default": no style instructions, Claude Code’s standard system prompt
"Proactive": starts implementing right away and makes reasonable assumptions on routine decisions; permission mode unchanged
"Concise": leads with the result, no preamble, narration, or closing recap
"Explanatory": adds Insight blocks explaining the choices behind the code
"Learning": adds Insight blocks and leaves design-heavy pieces for you at TODO(human) markers
A value that does not match a style name exactly (for example explanatory) falls back to Default. /output-style and the /config menu save the choice to .claude/settings.local.json.
Custom styles: reference the style name (the file name without .md, unless the frontmatter sets name) from .claude/output-styles/ for a project or ~/.claude/output-styles/ for a user-wide style. Custom styles omit the built-in Claude Code software engineering instructions unless their YAML frontmatter sets keep-coding-instructions: true. Style selection applies to the next message. Restart to discover newly created style files.
{ "outputStyle": "Explanatory" }
{ "outputStyle": "strict-reviewer" }
Setting persists across sessions. Explanatory and Learning increase output tokens; prompt caching can reduce repeated input cost but does not discount generated output. See Section 9.7 for full documentation and custom style examples.
Customize the action verbs shown in the spinner and turn duration messages. Set mode to "replace" to use only your verbs, or "append" to add to defaults.
Shell script path (executed in /bin/sh) that outputs an auth token sent as X-Api-Key and Authorization: Bearer headers for model requests. Useful for short-lived credentials.
Keys confirmed in the JSON schema not covered in the sections above:
Key
Type
Description
claudeMdExcludes📋 Schema only
array
Glob patterns for CLAUDE.md files to exclude from loading
allowManagedMcpServersOnly
boolean
(Managed) Only managed MCP servers are usable
allowManagedHooksOnly
boolean
(Managed) Only managed and SDK hooks are loaded
allowManagedModsOnly
boolean
(Managed) Stop user-installed mods from loading; their plugins stay installed and their skills, commands, agents, and MCP servers still load. Built-in mods keep running
Set in your shell before launching claude, or configure under the env key in settings.json to apply to every session. When an env var and an equivalent settings field both apply, the env var takes precedence (e.g. ANTHROPIC_MODEL overrides the model setting). Changes take effect on the next claude launch.
DEPRECATED. Use ANTHROPIC_DEFAULT_HAIKU_MODEL instead
ANTHROPIC_SMALL_FAST_MODEL_AWS_REGION
AWS region for the Haiku-class model on Bedrock when ANTHROPIC_DEFAULT_HAIKU_MODEL is also set
CLAUDE_CODE_SUBAGENT_MODEL
Override model for subagents (e.g. haiku)
CLAUDE_CODE_SUBAGENT_MODEL_FORCE
Set to 1 (v2.1.257+) to force every subagent, teammate and workflow agent onto CLAUDE_CODE_SUBAGENT_MODEL, or onto the lead’s model when that variable is unset or inherit, ignoring the spawn prompt and the agent definition’s model (Agent Teams)
CLAUDE_CODE_EFFORT_LEVEL
Effort level: low, medium, high, xhigh, max, or auto. Takes precedence over /effort and the effortLevel setting
FALLBACK_FOR_ALL_PRIMARY_MODELS⚠️ Not in official list (checked 2026-10-27)
Any non-empty value triggers fallback to --fallback-model after repeated overload errors on any primary model, not just Opus
CLAUDE_CODE_DISABLE_LEGACY_MODEL_REMAP
Set to 1 to prevent automatic remapping of Opus 4.0 and 4.1 to the current Opus version on the Anthropic API
API request timeout in milliseconds (default: 600000). Maximum: 2147483647
BASH_DEFAULT_TIMEOUT_MS
Default bash command timeout in milliseconds (default: 120000)
BASH_MAX_TIMEOUT_MS
Maximum bash command timeout in milliseconds (default: 600000)
BASH_MAX_OUTPUT_LENGTH
Maximum characters in bash output before saving to a file and sending the path
MAX_THINKING_TOKENS
Fixed budget on older extended-thinking models. 0 disables thinking where supported, but not on Opus 5.5, Sonnet 5.5, Haiku 5.5 or Fable. The fixed-budget compatibility switch applies only to Opus 4.6 and Sonnet 4.6
CLAUDE_CODE_MAX_OUTPUT_TOKENS
Output limit; default and cap depend on the model. Unrecognized gateway IDs default to 32,000. Values above the known model cap are reduced to that cap
CLAUDE_CODE_FILE_READ_MAX_OUTPUT_TOKENS
Override default file read token limit
CLAUDE_CODE_MAX_CONTEXT_TOKENS
Override context window size Claude Code assumes for the active model. Only takes effect when DISABLE_COMPACT is also set
CLAUDE_CODE_MAX_TURNS
Cap the number of agentic turns per session. Equivalent to --max-turns (the flag takes precedence when both are set)
CLAUDE_CODE_MAX_RETRIES
Override the number of retries for failed API requests (default: 10)
CLAUDE_CODE_MAX_TOOL_USE_CONCURRENCY⚠️ Not in official list (checked 2026-10-27)
Maximum read-only tools and subagents executing in parallel (default: 10)
CLAUDE_ASYNC_AGENT_STALL_TIMEOUT_MS
Stall timeout in milliseconds for background subagents (default: 600000)
TASK_MAX_OUTPUT_LENGTH
No effect since TaskOutput was removed in v2.1.278
MAX_STRUCTURED_OUTPUT_RETRIES
Retries when model response fails --json-schema validation in non-interactive mode (default: 5)
CLAUDE_CODE_STOP_HOOK_BLOCK_CAP
Maximum consecutive times a Stop hook may block the turn from ending before Claude Code overrides it (default: 8)
CLAUDE_CODE_SESSIONEND_HOOKS_TIMEOUT_MS
SessionEnd hook time budget in ms. Default: 1.5s, raised to the highest configured per-hook timeout up to 60s
CLAUDE_CODE_API_KEY_HELPER_TTL_MS
Credential refresh interval in ms for apiKeyHelper
MCP_TIMEOUT
MCP server startup timeout in ms (default: 30000)
MCP_TOOL_TIMEOUT
MCP tool execution timeout in ms (default: 100000000)
Set to 1 in subprocesses Claude Code spawns (Bash/PowerShell tools, tmux sessions, hook commands, status line commands, stdio MCP subprocesses). Use to detect when a script runs inside Claude Code
DEBUG
Set to 1 (or true/yes/on) to enable debug mode. Logs written to ~/.claude/debug/<session-id>.txt. Namespace patterns like DEBUG=express:* do not trigger it
CLAUDE_CODE_DEBUG_LOGS_DIR
Override the debug log file path (file path, not a directory). Requires debug mode enabled separately
CLAUDE_CODE_DEBUG_LOG_LEVEL
Minimum log level for the debug file: verbose, debug (default), info, warn, error
CLAUDE_CODE_SHELL
Override automatic shell detection
CLAUDE_CODE_SHELL_PREFIX
Command prefix prepended to all shell commands Claude Code spawns
CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR
Return to the original working directory after each Bash command in the main session (1 to enable)
CLAUDE_CODE_NEW_INIT
Set to 1 to make /init run an interactive setup flow asking which files to generate
CLAUDE_CODE_SIMPLE
Run with minimal system prompt and only Bash, file read, and file edit tools. Equivalent to --bare
CLAUDE_CODE_SIMPLE_SYSTEM_PROMPT
Set to 1 to use a shorter system prompt and abbreviated tool descriptions. Set to 0/false/no/off to opt out
CLAUDE_CODE_SUBPROCESS_ENV_SCRUB
Set to 1 to strip Anthropic and cloud provider credentials from subprocess environments (Bash, hooks, MCP stdio servers)
CLAUDE_CODE_ADDITIONAL_DIRECTORIES_CLAUDE_MD
Set to 1 to load CLAUDE.md files from directories specified with --add-dir
CLAUDE_CODE_TMPDIR
Override the temp directory used for internal temp files
CLAUDE_CODE_GIT_BASH_PATH
Windows only: path to Git Bash executable when not in PATH
CLAUDE_CODE_GLOB_HIDDEN
Set to false to exclude dotfiles from Glob tool results (dotfiles included by default)
CLAUDE_CODE_GLOB_NO_IGNORE
Set to false to make the Glob tool respect .gitignore patterns
CLAUDE_CODE_GLOB_TIMEOUT_SECONDS
Glob tool file discovery timeout in seconds (default: 20, 60 on WSL)
CLAUDE_CODE_USE_NATIVE_FILE_SEARCH⚠️ Not in official list (checked 2026-10-27)
Set to 1 to discover commands/subagents using Node.js file APIs instead of ripgrep
CLAUDE_CODE_PERFORCE_MODE
Set to 1 to enable Perforce-aware write protection (edits fail on files lacking owner-write bit)
CLAUDE_CODE_POWERSHELL_RESPECT_EXECUTION_POLICY
Set to 1 to stop Claude Code from passing -ExecutionPolicy Bypass when spawning PowerShell
CLAUDE_CODE_FORK_SUBAGENT
Controls fork mode, which is on by default: Claude can start forks that inherit the full conversation context. See the official variable list for the off value and per-surface behavior
CLAUDE_AUTO_BACKGROUND_TASKS
Set to 1 to force-enable automatic backgrounding of long-running agent tasks: a subagent moves to the background after about two minutes. CLAUDE_CODE_WORKER_CHECKIN_SCHEDULE sets the check-in reminders
CLAUDE_CODE_REMOTE
Set automatically to true when running as a cloud session
CLAUDE_CODE_REMOTE_SESSION_ID
Set automatically in cloud sessions to the current session ID
CLAUDE_CODE_SESSION_ID
Set automatically to the current session ID in Bash/PowerShell subprocesses, hook commands, and MCP stdio subprocesses
CLAUDE_EFFORT
Set automatically in Bash/hook subprocesses to the active effort level: low, medium, high, xhigh, or max
CLAUDE_ENV_FILE
Path to a shell script run before each Bash command. Also populated dynamically by SessionStart, Setup, CwdChanged, and FileChanged hooks
USE_BUILTIN_RIPGREP
Set to 0 to use system-installed rg instead of the one bundled with Claude Code
Auto-compact trigger threshold as a percentage (1-100). Default ~95%. Lower values trigger compaction earlier
CLAUDE_CODE_AUTO_COMPACT_WINDOW
Context capacity in tokens used for compaction calculations. Defaults to model’s context window (200K standard, 1M for extended). Lower value on a 1M model treats it as smaller for compaction purposes
CLAUDE_CODE_DISABLE_1M_CONTEXT
Set to 1 to disable 1M token context window support
DISABLE_AUTO_COMPACT
Set to 1 to disable automatic compaction when approaching the context limit. Manual /compact remains available
DISABLE_COMPACT
Set to 1 to disable all compaction, including manual /compact
CLAUDE_CODE_DISABLE_THINKING
Set to 1 to force-disable extended thinking regardless of model support
DISABLE_INTERLEAVED_THINKING
Set to 1 to prevent sending the interleaved-thinking beta header. Use when your gateway does not support it
Standard OTEL exporter variables (OTEL_METRICS_EXPORTER, OTEL_LOGS_EXPORTER, OTEL_EXPORTER_OTLP_ENDPOINT, OTEL_EXPORTER_OTLP_PROTOCOL, OTEL_EXPORTER_OTLP_HEADERS, OTEL_METRIC_EXPORT_INTERVAL, OTEL_RESOURCE_ATTRIBUTES, and signal-specific variants) are supported. CLAUDE_CODE_ENABLE_TELEMETRY=1 is required to activate OTel collection.
Variable
Description
OTEL_LOG_RAW_API_BODIES
Emit API request/response JSON as log events. Set to 1 for inline bodies (truncated at 60 KB), or file:<dir> to write full bodies to disk
OTEL_LOG_TOOL_CONTENT
Set to 1 to include tool input/output content in OTel span events. Disabled by default
OTEL_LOG_TOOL_DETAILS
Set to 1 to include tool input arguments, MCP server names, and raw error strings in OTel traces. Disabled by default
OTEL_LOG_USER_PROMPTS
Set to 1 to include user prompt text in OTel traces. Disabled by default
OTEL_METRICS_INCLUDE_ACCOUNT_UUID
Set to false to exclude account UUID from metrics attributes (default: included)
OTEL_METRICS_INCLUDE_ENTRYPOINT
Set to true to include session entrypoint in metrics attributes (default: excluded). Added in v2.1.152
OTEL_METRICS_INCLUDE_RESOURCE_ATTRIBUTES
Set to false to exclude OTEL_RESOURCE_ATTRIBUTES keys from metric datapoint labels (default: included). Added in v2.1.161
OTEL_METRICS_INCLUDE_SESSION_ID
Set to false to exclude session ID from metrics attributes (default: included)
OTEL_METRICS_INCLUDE_VERSION
Set to true to include Claude Code version in metrics attributes (default: excluded)
CLAUDE_CODE_OTEL_FLUSH_TIMEOUT_MS
Timeout in ms for flushing pending OTel spans on shutdown (default: 5000)
CLAUDE_CODE_OTEL_SHUTDOWN_TIMEOUT_MS
Timeout in ms for the OTel exporter to finish on exit (default: 2000)
CLAUDE_CODE_OTEL_HEADERS_HELPER_DEBOUNCE_MS
Refresh interval in ms for dynamic OTel headers (default: 1740000, 29 minutes)
CLAUDE_CODE_ENABLE_FEEDBACK_SURVEY_FOR_OTEL
Set to 1 to route session quality survey ratings to your OTel collector instead of Anthropic
CLAUDE_CODE_PROPAGATE_TRACEPARENT
Set to 1 to propagate W3C trace context when ANTHROPIC_BASE_URL points at a custom proxy. Added in v2.1.152
false turns off idle compaction, which compacts a 200K+ token conversation left idle about 54 minutes before its one-hour prompt cache expires (v2.1.299). CLAUDE_CODE_IDLE_COMPACT_MIN_TOKENS changes the threshold (minimum 100,000)
syncClaudeAiSkills, syncClaudeAiPlugins
User, local, managed, --settings
false stops syncing the skills or plugins enabled on your claude.ai account into terminal sessions; a repository cannot turn it off for you (v2.1.275)
maxProseWidth
Any file
Caps response prose at this many columns (minimum 40); tables and code keep the full width (v2.1.282)
deniedModels
Managed
Blocks models by family alias or ID, even when availableModels permits them (v2.1.283)
availableModelsMatch
Managed
"exact" keeps a newer version blocked until it is listed; the default "prefix" lets "claude-opus-5" permit Opus 5.5
allowedProviders
Managed
Lists the services a machine may reach Claude through (anthropic, bedrock, gateways…); other providers are refused at startup and login (v2.1.285)
sandbox.allowNestedUserNamespaces
User, managed
Lets sandboxed Linux/WSL2 commands create user namespaces (Chromium, nested bubblewrap). Turns off a security boundary: sandboxed code can end up running outside the sandbox (v2.1.305)
CLAUDE_CODE_AUTO_MODE_SERVER
Environment
0 uses Claude Code’s own classifier requests instead of server-side auto mode review
CLAUDE_CODE_AUTO_MODE_MODEL
Environment
Model for the classifier requests Claude Code sends in auto mode, as a full model ID (v2.1.302)
CLAUDE_CODE_DISABLE_DANGEROUS_RM_TIMEOUT
Environment
1 turns off the time limit and auto-denial on critical-path removals in auto and bypassPermissions (v2.1.281)
CLAUDE_CODE_DISABLE_WEB_FETCH
Environment
1 turns off WebFetch; WebSearch stays available (v2.1.285)
CLAUDE_CODE_DISABLE_FILE_MENTIONS
Environment
1 stops expanding @ file mentions; Claude gets the paths and reads them with its tools (v2.1.299)
CLAUDE_CODE_DISABLE_ULTRAREVIEW_TRUST_CHECK
Environment
1 lets claude ultrareview run in an untrusted folder when it cannot ask, such as CI; or pass --trust-workspace (v2.1.309)
MCP_PROTOCOL_NEGOTIATION
Environment
legacy skips the probe for MCP protocol revision 2026-07-28 on the v2 MCP client runtime
These keys appear in the official settings reference (read October 27, 2026) without a dedicated entry above. Descriptions are the first sentence of the official entry; read it for types, defaults and precedence.
Key
Scope
Purpose
advisorModel
Any file
Pick which model answers when Claude calls the server-side advisor tool.
agentPushNotifEnabled
Any file
Allow Claude to send a push notification to your phone when it decides one is worth sending, for example when a long task finishes.
appendPlugins
User or managed
List the managed plugins whose mods run after every mod a user installs, in the listed order.
askUserQuestionTimeout
User or managed
Let an unanswered AskUserQuestion dialog auto-continue after a period of idle time, submitting whatever options you had already selected.
autoContinueAtUsageLimit
User or managed
After a claude.ai usage limit stops your session, wait in the open session and continue the task automatically after the reset.
awaySummaryEnabled
Any file
Show a one-line session recap when you return to the terminal after a few minutes away.
browserExternalPageTools
Managed
Stop Claude from using its tools to read or act on external pages in the desktop app’s Browser pane.
claudeInChromeDefaultEnabled
Global config
Start every interactive CLI session with Chrome integration on, without passing --chrome each time.
copyFullResponse
Global config
Make /copy copy the full response every time, without showing the picker.
copyOnSelect
Global config
Copy text to your clipboard automatically when you finish selecting it with the mouse in fullscreen rendering or agent view.
defaultToAgentsView
Global config
Open agent view instead of a new conversation when you run claude with no arguments.
diffTool
Global config
Choose where Claude Code shows the diff of an Edit or Write change it proposes when a VS Code or JetBrains IDE is connected: "auto" opens it in the IDE’s diff viewer, "terminal" keeps it in the terminal.
disableAgentView
Any file
Turn off background agents and agent view: claude agents, --bg, /background, and the on-demand supervisor.
disableArtifact
Any file
> [!WARNING]
disableBrowserExternalNavigation
Managed
Turn off external browsing in the desktop app’s Browser pane for people and Claude alike.
disableCommandPluginSources
Managed
Block the command plugin source, which installs a plugin by running a marketplace-declared command on the user’s machine.
disableDesktopLocalSessions
Managed
Turn off Code sessions that run on the device in the desktop app, for deployments where developers should work on remote machines over SSH.
disableMobileSimulatorTools
Managed
Block Claude’s tools for the desktop app’s iOS Simulator pane.
disableRemoteControl
Any file
Turn off Remote Control: Claude Code then refuses claude remote-control, the --remote-control flag, auto-start, and the in-session toggle, and reports that your organization’s policy disabled it.
disableSideloadFlags
Managed
Reject the --plugin-dir, --plugin-url, --agents, and --mcp-config CLI flags at startup, which users could otherwise pass to bypass strictKnownMarketplaces for a single run.
disableWorkflows
Any file
Turn off dynamic workflows and the bundled workflow commands for everyone your settings reach, such as an organization through managed settings.
enableArtifact
Any file
Turn off the Artifact tool, which publishes session output as a private web page on claude.ai.
enableWorkflows
Any file
Turn dynamic workflows on or off for yourself when your plan’s default isn’t what you want.
externalEditorContext
Global config
When you press Ctrl+G, Claude Code opens the prompt you’re typing in your external editor.
fileCheckpointingEnabled
Any file
Have Claude Code snapshot files before each edit so /rewind can restore them.
gatewayInternalNetworks
Managed
Declare the public IPv4 blocks that your organization numbers its internal network from, so /login accepts a cloud gateway there.
gcpAuthRefresh
Any file
Run your own command to refresh Google Cloud Application Default Credentials when Claude Code finds they’ve expired or can’t be loaded, so Google Cloud’s Agent Platform requests keep working without you re-authenticating by hand.
inputNeededNotifEnabled
Any file
Get a push notification on your phone when a permission prompt or question is waiting for your input.
leftArrowOpensAgents
Global config
Press ← on an empty prompt to background the session and open agent view.
managedSourcesBehavior
Managed
Choose whether Claude Code applies only the highest-priority managed source your organization delivers, or combines every admin source it delivers.
minimumVersion
Any file
Keep background auto-updates and claude update from installing any version below this one, so moving to the "stable" channel doesn’t downgrade you from a newer "latest" build.
permissionExplainerEnabled
Global config
> [!WARNING]
policyHelper
Managed
Run an executable you deploy that computes managed settings at startup, so you can derive policy from device posture, identity, or a remote service instead of a static file.
policyHelper.path
Managed
Name the helper executable Claude Code runs.
policyHelper.refreshIntervalMs
Managed
Have Claude Code re-run the helper in the background on an interval so policy changes reach a running session.
policyHelper.timeoutMs
Managed
Set how long Claude Code waits for the helper before treating the run as failed.
prStatusFooterEnabled
Global config
Show a badge in the prompt footer for the current branch’s open pull request or merge request, with a colored underline that shows its status.
preferredNotifChannel
Any file
Choose how Claude Code notifies you when a task completes or a permission prompt is waiting.
prependPlugins
User or managed
List the managed plugins whose mods run before every mod a user installs, in the listed order.
processWrapper
User or managed
On macOS and Linux, place a corporate launcher command in front of the background processes Claude Code starts.
promptSuggestionEnabled
Any file
Show or hide prompt suggestions, the grayed-out predictions that appear in your prompt input.
remote.defaultEnvironmentId
Any file
Pick the default cloud environment for cloud sessions you create from the CLI, such as with claude --cloud.
remoteControlAtStartup
Any file
Connect Remote Control automatically when each interactive session starts, instead of waiting for /remote-control.
sandbox.credentials.allowPlaintextInject
User or managed
Allow mask substitution on plain HTTP requests as well as TLS-terminated HTTPS.
sandbox.credentials.awsPairs
User or managed
Group masked environment variables that form one AWS credential for SigV4 re-signing when your credential lives in variables with non-standard names.
sandbox.credentials.sigv4
User or managed
Choose what the sandbox proxy does with AWS request forms it can’t re-sign: streaming for aws-chunked streaming uploads, presigned for presigned URLs, and sigv4a for SigV4A asymmetric signatures.
sandbox.ignoreViolations
Any file
Silence sandbox violation reports for paths you expect a command to probe and be refused, such as a tool that checks /etc/hosts on startup, so those denials don’t show up as violations or in what Claude sees.
skillListingBudgetFraction
Any file
Each turn, Claude sees a listing of your skills with their descriptions, and Claude Code caps that listing at a share of the context window.
skillListingMaxDescChars
Any file
Each turn, Claude sees a listing of your skills that shows each skill’s description and when_to_use text.
skipAutoPermissionPrompt
User or managed
Skip the one-time notice describing auto mode that Claude Code shows when you first enter auto mode yourself, for example through your own settings or the mode selector, rather than when the built-in default starts a session in it.
skipDangerousModePermissionPrompt
User, local, or managed
Skip the confirmation dialog Claude Code shows before a session enters bypassPermissions mode, whether from --dangerously-skip-permissions or from defaultMode: "bypassPermissions".
sshConfigs
User or managed
Add SSH connections to the Desktop environment selector.
sshHostAllowlist
Managed
Limit the hosts a Desktop SSH session can connect to.
sshHostCommands
User or managed
Add SSH connections whose host comes from a command to the Desktop environment selector.
strictPluginOnlyCustomization
Managed
Block skills, agents, hooks, and MCP servers from user and project sources, so they can come only from plugins or managed settings.
strictPluginOnlyCustomization.agents
Managed
Lock the agents surface.
strictPluginOnlyCustomization.hooks
Managed
Lock the hooks surface:
strictPluginOnlyCustomization.mcp
Managed
Lock the mcp surface.
strictPluginOnlyCustomization.skills
Managed
Lock the skills surface.
switchModelsOnFlag
Any file
Choose what happens when a safety classifier flags a request: switch to the fallback model and continue, or pause so you can choose between switching and editing the prompt.
syntaxHighlightingDisabled
Any file
Claude Code colors code by language in the diffs, code blocks, and file previews it shows in the terminal, with its built-in highlighter; no plugin or language server is involved.
teammateDefaultModel
Global config
> [!WARNING]
terminalTitleFromRename
Any file
Claude Code sets your terminal tab’s title.
timeFormat
Any file
Choose how Claude Code writes the times it shows in the interface, such as the done 6:05 PM at the end of each turn duration message and the timestamps in the transcript viewer.
timeZone
Any file
Show the times in the interface in a time zone other than your system’s.
viewMode
Any file
Set the transcript view Claude Code starts in: "default", "verbose", or "focus".
workflowKeywordTriggerEnabled
Any file
Choose whether typing the keyword ultracode in a prompt triggers a dynamic workflow.
worktree.location
User
Choose the folder on a remote machine where Desktop SSH sessions create their worktrees, instead of <project-root>/.claude/worktrees/.
Further environment variables in the official list
These variables appear in the official environment variable list (read October 27, 2026) without a row above. Descriptions are the first sentence of the official entry.
Variable
Purpose
ANTHROPIC_DEFAULT_FABLE_MODEL_DESCRIPTION
Display description for the pinned Fable model in the /model picker.
ANTHROPIC_DEFAULT_FABLE_MODEL_NAME
Display name for the pinned Fable model in the /model picker.
Comma-separated list of capabilities the pinned Fable model supports, for example effort,thinking.
ANTHROPIC_FEDERATION_RULE_ID
Federation rule ID for Workload Identity Federation.
ANTHROPIC_FOUNDRY_AUTH_TOKEN
Bearer token for Microsoft Foundry authentication, such as a Microsoft Entra access token.
ANTHROPIC_ORGANIZATION_ID
Organization ID for Workload Identity Federation.
ANTHROPIC_PROFILE
Name of the Anthropic profile to authenticate with, such as one created by ant auth login or by signing in to a Console account without an API key.
API_FORCE_IDLE_TIMEOUT
Override the 5-minute body idle timeout that aborts a streaming model response when no bytes arrive.
BETA_TRACING_ENDPOINT
OTLP/HTTP endpoint for detailed beta tracing: with ENABLE_BETA_TRACING_DETAILED=1, logs and traces go there instead of to the configured exporters.
CLAUDE_AUTO_BACKGROUND_TASKS
Set to 1 to force-enable automatic backgrounding of long-running agent tasks.
CLAUDE_AX_PREPARK_MS
In screen reader mode, how many milliseconds Claude Code waits before it writes a new or changed line.
CLAUDE_AX_STARTUP_QUIET_MS
In screen reader mode, how many milliseconds Claude Code holds the first interface render after the startup confirmation line.
CLAUDE_BYTE_STREAM_IDLE_TIMEOUT_MS
Timeout in milliseconds for the byte-level streaming idle watchdog; when set, it takes precedence over CLAUDE_STREAM_IDLE_TIMEOUT_MS for that watchdog and leaves the event-level watchdog unchanged.
CLAUDE_CODE_ARTIFACT_AUTO_OPEN
Set to 0 to stop Claude Code from opening the browser automatically when a new artifact is published
CLAUDE_CODE_ARTIFACT_COMMENTS
Set to 0 to stop Claude reading and replying to comments on an artifact.
CLAUDE_CODE_ARTIFACT_COMMENTS_AUTOREACT
Set to 0 to stop Claude replying on its own to comments sent to it.
CLAUDE_CODE_ARTIFACT_PUBLISH_EDITS
Set to 0 to stop Claude from applying small changes to an artifact’s file as part of publishing it.
Time in milliseconds Claude Code waits for the AWS default credential provider chain to produce credentials before the request fails with AWS default-chain credential resolve timed out (default: 60000).
CLAUDE_CODE_BASH_EDIT_DIFF
Set to 0 to turn off the diff of the files that changed while a Bash command ran, or 1 to record it in every permission mode.
CLAUDE_CODE_BG_TASKS_REPORT_RUNNING
Set to 0 to make a non-interactive session report an idle status to its host at every turn end, even while background work is still running.
CLAUDE_CODE_BRIDGE_SESSION_ID
Set automatically in Bash tool and hook command subprocesses while the session has an active Remote Control connection, and removed when the connection ends.
CLAUDE_CODE_BS_AS_CTRL_BACKSPACE
Set to 0 to make Claude Code read the 0x08 byte, also written ^H, as plain Backspace, or 1 to read it as Ctrl+Backspace.
CLAUDE_CODE_CONNECT_TIMEOUT_MS
Removed in v2.1.186 and now a no-op.
CLAUDE_CODE_DISABLE_ADVISOR_TOOL
Set to 1 to disable the advisor tool.
CLAUDE_CODE_DISABLE_ARTIFACT
Set to 1 to turn off the Artifact tool, which publishes session output as a private web page on claude.ai.
CLAUDE_CODE_DISABLE_AUTH_REFRESH_LOCK
Set to 1 to make a Claude Code process run its gcpAuthRefresh or awsAuthRefresh command itself instead of waiting while another process runs it.
CLAUDE_CODE_DISABLE_BEDROCK_CONTENT_TYPE_DEFAULT
Set to 1 to stop Claude Code from treating an Amazon Bedrock streaming response with a missing or empty Content-Type header as Amazon Bedrock’s binary event stream.
CLAUDE_CODE_DISABLE_BEDROCK_CONTENT_TYPE_GUARD
Set to 1 to skip the check that an Amazon Bedrock streaming response carries the application/vnd.amazon.eventstream content-type.
CLAUDE_CODE_DISABLE_BG_EXIT_HANDOFF
Set to 1 to stop a background session’s in-flight work when the supervisor stops, restarts, or updates that session’s process, instead of handing it to the session’s next process.
CLAUDE_CODE_DISABLE_CFC_PROMPT
Set to 1 to keep the Claude in Chrome browser tools available while omitting the Chrome section of the system prompt and the /claude-in-chrome bundled skill.
CLAUDE_CODE_DISABLE_EXPLORE_PLAN_AGENTS
Set to 1 to disable the built-in Explore and Plan subagents.
CLAUDE_CODE_DISABLE_INLINE_SHELL_RM_PROMPT
Set to 1 to stop Claude Code from reading scripts passed to a shell with -c, such as bash -c 'rm -rf ~', for critical-path removals.
CLAUDE_CODE_DISABLE_MCP_TASK_BACKGROUND
Set to 1 so that Claude waits for the result of an MCP task in an interactive session.
CLAUDE_CODE_DISABLE_MODEL_ACCESS_FALLBACK
Set to 1 to stop Claude Code on Amazon Bedrock and Google Cloud’s Agent Platform from switching to an older model when your account loses access to a session’s model mid-session; the refused request fails at once instead.
CLAUDE_CODE_DISABLE_MOUNTED_DRIVE_RM_PROMPT
Set to 1 to turn off the mounted drive prompt for paths such as /Volumes/USB.
Set to 1 to stop Claude Code from re-reading the mTLS client certificate and key when an API request fails with a connection-level error, such as a connection reset or a TLS handshake error.
CLAUDE_CODE_DISABLE_NOTIFICATION_PRESENCE_CHECK
Set to 1 to send the PushNotification tool’s desktop notification even while you are typing in or focused on the terminal.
Set to 1 to stop Claude Code from running your Notification hooks for unanswered permission requests in sessions where Claude Code sends them to the Agent SDK’s canUseTool callback, which is how Claude Desktop and the VS Code extension host Claude Code.
CLAUDE_CODE_DISABLE_POWERSHELL_CMD_RM_DENY
Set to 1 to turn off the PowerShell tool check that denies the cmd built-ins rd, rmdir, del, and erase on a system path, such as a drive root or your home directory.
CLAUDE_CODE_DISABLE_REFUSAL_FALLBACK
Set to 1 to turn off the automatic model switch when a safety classifier flags a request, the behavior the switchModelsOnFlag setting controls
CLAUDE_CODE_DISABLE_STRUCTURED_OUTPUTS
Set to 1 to stop Claude Code from sending the structured-output output_config.format field and the anthropic-beta value that pairs with it, for an LLM gateway whose upstream rejects them.
CLAUDE_CODE_DISABLE_WINDOWS_SHELL_LAUNCHER
Set to 1 to start PowerShell tool commands on Windows directly instead of through the cmd.exe launcher.
CLAUDE_CODE_EMIT_SESSION_STATE_EVENTS
Set to 1 to add session_state_changed messages, which carry the session’s state, to the message stream.
CLAUDE_CODE_ENABLE_CFC
Set to 1 to start a CLI session with Chrome integration on, or 0 to start with it off.
CLAUDE_CODE_ENABLE_OPUS_4_7_FAST_MODE
Removed in v2.1.142, when the fast mode default moved from Opus 4.6 to Opus 4.7
CLAUDE_CODE_ENABLE_SHARE_BUNDLE
Set to 1 to make /share save its report and the whole session transcript to a local archive under ~/.claude/feedback-bundles/ instead of sending the report to Anthropic.
CLAUDE_CODE_FORCE_SESSION_PERSISTENCE
Set to 1 to force transcript persistence, prompt history, and claude agents registration even when this claude was launched from inside another Claude Code session.
CLAUDE_CODE_FORCE_STRIKETHROUGH
Set to 1 to force strikethrough rendering for ~~text~~ in Claude’s responses when your terminal supports it but is not auto-detected, such as over SSH without TERM_PROGRAM forwarded.
CLAUDE_CODE_FORCE_TERMINAL_IMAGES
Set to 1 to draw mod Image elements as pictures when your terminal draws kitty graphics protocol images with Unicode placeholders and you still see an Image’s dimmed alt text.
CLAUDE_CODE_GATEWAY_MODEL_DISCOVERY_TIMEOUT_MS
Timeout in milliseconds for the gateway model discovery request that CLAUDE_CODE_ENABLE_GATEWAY_MODEL_DISCOVERY turns on (default: 3000).
CLAUDE_CODE_NONBLOCKING_STDOUT
Set to 1 to write terminal output through a second non-blocking file descriptor, so a terminal that stops reading, such as a paused tmux control-mode pane or a stalled SSH connection, can’t freeze Claude Code mid-session.
CLAUDE_CODE_OTEL_DIAG_STDERR
Set to 1 to write OpenTelemetry exporter diagnostic errors to stderr.
CLAUDE_CODE_PLUGIN_DIRS
Plugin directories to load for the session, each loaded the way a --plugin-dir flag loads it.
CLAUDE_CODE_PLUGIN_DIR_WATCH
Controls whether Claude Code reloads a mod when the mod’s files change.
CLAUDE_CODE_PRINT_BG_WAIT_CEILING_MS
Ceiling in milliseconds on idle waiting for background work, such as subagents and workflows, after the final turn in non-interactive mode with the -p flag.
CLAUDE_CODE_PROMPT_CACHE_TTL
Set 5m or 1h, the only values Claude Code accepts, to choose the prompt cache TTL for the main conversation: your interactive, -p, and SDK turns, plus the helpers that run inline with them.
CLAUDE_CODE_RESUME_INTERRUPTED_TURN_MAX_AGE_MS
Maximum age in milliseconds of the last transcript message for a session that ended mid-turn to continue automatically on resume.
CLAUDE_CODE_SEND_FEEDBACK
Set to 0 to turn off Claude-drafted feedback for a session.
CLAUDE_CODE_SKIP_AWS_CRED_CACHE
Set to 1 to turn off the in-process cache of credentials resolved from the AWS default credential provider chain, so Claude Code resolves the chain on every API request.
CLAUDE_CODE_SKIP_FAST_MODE_ORG_CHECK
Set to 1 to skip the client-side fast mode availability check, for proxies that intercept the check’s request rather than refuse it.
CLAUDE_CODE_SKIP_MODEL_ACCESS_MEMORY
The startup model checks on Amazon Bedrock and Google Cloud’s Agent Platform remember on this machine which models they found your account can’t invoke, for up to a day.
CLAUDE_CODE_STARTUP_FAILURE_RESULTS
Set to 1 to have a session started with --output-format stream-json write a result message naming why Claude Code refused to start for startup failures that otherwise end with stderr alone.
CLAUDE_CODE_SUBAGENT_PROMPT_CACHE_TTL
Set 5m or 1h, the only values Claude Code accepts, to choose the prompt cache TTL for requests outside the main conversation, such as subagents, workflows, and background work.
CLAUDE_CODE_SYNC_SKILLS_INSTALL_TIMEOUT_MS
Timeout in milliseconds for the skills resync that runs mid-session when an app built on the Agent SDK reloads skills (default: 30000).
CLAUDE_CODE_TOOL_MEMORY_CGROUP_EXCLUDE
On Linux and WSL, set to a comma-separated list of the kinds of processes Claude Code excludes from the tool memory cap, such as mcp or lsp.
CLAUDE_CODE_TRANSCRIPT_LOCAL_GC
Set to 1 to limit how large the transcript file of a long -p or Agent SDK session grows.
CLAUDE_CODE_USER_DIALOG_TIMEOUT_MS
Deadline in milliseconds before Claude Code cancels a dialog it forwards to a remote client such as a Remote Control or SDK host, or the approval dialog for a held cross-session message; permission prompts and AskUserQuestion questions use their own…
CLAUDE_CODE_WARN_DISK_FREE_GIB
Set to a whole number of GiB, such as 10, to have Claude Code warn Claude when free space on the working directory’s filesystem falls below that amount.
CLAUDE_CODE_WARN_MEMORY_FREE_GIB
On Linux and WSL, set to a whole number of GiB such as 4 to have Claude Code warn Claude when available memory falls below that amount.
CLAUDE_CODE_WORKER_CHECKIN_SCHEDULE
When CLAUDE_AUTO_BACKGROUND_TASKS is set to 1, how long Claude Code waits before each reminder to Claude to check on background subagents that are still running.
CLAUDE_CODE_WORKFLOW_MAX_CONCURRENT_AGENTS
How many agents a single workflow run executes at once, from 1 to 256.
CLAUDE_CODE_YOU_SHOULD_KNOW_DEBUG
Set to 1 to show, above the prompt, a line for each of the last three checks the You should know mod made in this session.
CLAUDE_JOB_DIR
Set by Claude Code in each background session to that session’s ~/.claude/jobs/<id> directory.
CLAUDE_STREAM_FIRST_BYTE_TIMEOUT_MS
Deadline in milliseconds for the first response byte of a streaming request, on the connections where the first-byte deadline runs.
CLAUDE_SUBAGENT_BG_SHELL_MAX_MS
Removed in v2.1.260 and now a no-op.
DISABLE_PROMPT_CACHING_FABLE
Set to 1 to disable prompt caching for Fable models
ENABLE_BETA_TRACING_DETAILED
Set to 1, and set BETA_TRACING_ENDPOINT to your OTLP/HTTP collector endpoint, to turn on detailed beta tracing, which adds content-bearing span attributes and the claude_code.hook span.
MCP_DISCOVERY_CACHE
Turns the MCP discovery cache on or off.
MCP_DISCOVERY_CACHE_MAX_STALE_S
Maximum age, in seconds, of a discovery-cache entry (default: 14400, or 4 hours).
MCP_DISCOVERY_CACHE_STRIKES
At a start where a discovery-cache entry is older than MCP_DISCOVERY_CACHE_TTL_S, Claude Code refreshes it in the background.
MCP_DISCOVERY_CACHE_TTL_S
Seconds for which Claude Code uses a discovery-cache entry without refreshing it (default: 900).
OTEL_ATTRIBUTE_VALUE_LENGTH_LIMIT
Standard OpenTelemetry SDK limit on attribute value length.
OTEL_METRICS_INCLUDE_REPOSITORY
Set to true to tag OpenTelemetry metrics and events with vcs.* attributes identifying the session’s repository (default: excluded).
VERTEX_REGION_CLAUDE_5_5_OPUS
Override region for Claude Opus 5.5 when using Google Cloud’s Agent Platform.
VERTEX_REGION_CLAUDE_5_5_SONNET
Override region for Claude Sonnet 5.5 when using Google Cloud’s Agent Platform.
VERTEX_REGION_CLAUDE_5_OPUS
Override region for Claude Opus 5 when using Google Cloud’s Agent Platform.
VERTEX_REGION_CLAUDE_5_SONNET
Override region for Claude Sonnet 5 when using Google Cloud’s Agent Platform.
VERTEX_REGION_CLAUDE_FABLE_5
Override region for Claude Fable 5 when using Google Cloud’s Agent Platform.
VERTEX_REGION_CLAUDE_FABLE_5_1
Override region for Claude Fable 5.1 when using Google Cloud’s Agent Platform.
VERTEX_REGION_CLAUDE_HAIKU_5_5
Override region for Claude Haiku 5.5 when using Google Cloud’s Agent Platform.