Feature flags, all 2,095.

tengu_settings_mode_widening_carry_ignored ◇ Project or local settings can no longer widen the effective permission mode via defaultMode.never mined ·never mined never read tengu_shale_finch ◇ Default false. When on, subagents that are not teammates and are not spawned with an exact tool list lose the TodoWrite, TaskCreate, TaskGet, TaskUpdate and TaskList tools.v2.1.138 or earlier → v2.1.285 ○gone AND v2.1.308 tengu_shell_allow_rules_at_init ◇ An analytics event, not a feature flag. Claude Code logs it once during session initialization with counts of the always-allow permission rules for its shell tools, keyed <source>_<tool>_<kind> where the source is userSettings, projectSettings, localSettings, flagSettings, cliArg or session and the kind is bare, scoped or dangerous_prefix, plus total_shell_allow_rules.never mined ·never mined never read tengu_shell_set_cwd ◇ Directory tracking keeps working for backgrounded Bash commands instead of breaking partway through.never mined ·never mined never read tengu_shimmering_cherny ◇ The staleness cutoff for auto-resuming an interrupted turn is now configurable, defaulting to 6 hoursv2.1.269 → v2.1.309 ◆in the build never read tengu_shoji_engine ◇ No description from anybody.v2.1.193 → v2.1.202 ○gone AND v2.1.308 tengu_shutdown_pending_state ◇ Quitting mid-question leaves that prompt pending for your next session instead of denying it.never mined ·never mined never read tengu_signed_cache_identity ◇ Signed-cache validation now also resolves an identity and reports it via a new telemetry eventnever mined ·never mined never read tengu_signed_cache_shadow ◇ Managed-settings and policy-limits caches are signature-checked against embedded trust roots, but only report telemetry for now.never mined ·never mined never read tengu_silk_ledger ◇ Default false. Gate for synced project memory writes: while it is off, the orgMemoryWrites setting does not enable writes, turning writes on is refused as refused_gates, and the /config toggle Synced project memory writes is offered only if orgMemoryWrites is already true. When on, writes also need a signed-in account with an organization, CLAUDE_CODE_DISABLE_ORG_MEMORY unset and CLAUDE_MEMORY_STORES empty.v2.1.211 → v2.1.222 ○gone never read tengu_silver_lantern ◇ No description from anybody.never mined ·never mined never read tengu_skill_improvement_detected ◇ No description from anybody.never mined ·never mined never read tengu_skill_loaded ◇ Skill usage stats include hashed skill names only for plugins from outside marketplaces.never mined ·never mined never read tengu_skills_sync_manifest_failed ◇ An analytics event, not a feature flag. Claude Code logs it when fetching the sync manifest fails while listing synced skill entries (for plugins it logs tengu_plugins_sync_manifest_failed instead), with the field unavailable (whether the failure reason is unavailable).never mined ·never mined never read tengu_skills_sync_resync_skipped ◇ Skills-sync resync ticks now record why they fired and report when one is skipped after the machine slept too long.never mined ·never mined never read tengu_sl_gate_config ◇ The same shell preamble would install a bq shim on PATH, and cannot be turned on here.v2.1.239 → v2.1.291 ○gone never read tengu_slate_fern ◇ Default false. Turns on the unused plugin check: user-installed marketplace plugins not used for at least days days and sessions sessions feed the plugin-disuse-review startup tip (You haven't used the ... plugin in a while) and the days-since-last-use shown in the plugin list. Plugins that provide LSP servers, themes, output styles, monitors or workflows are never flagged, and the check is off when the strictKnownMarketplaces policy is set. true uses 14 days and 10 sessions; an object can set numeric days and sessions, each falling back to those values.v2.1.172 → v2.1.186 ○gone AND v2.1.308 tengu_slate_harbor_experiment ◇ Default false. When on, or when CLAUDE_CODE_NEW_INIT is set, /init sends the alternate multi-phase prompt instead of the original one that begins Please analyze this codebase and create a CLAUDE.md file. The /init command description follows CLAUDE_CODE_NEW_INIT alone.v2.1.138 or earlier → v2.1.309 ◆in the build AND v2.1.308 tengu_slate_heron ◇ No description from anybody.never mined ·never mined never read tengu_slate_kestrel ◇ Default false. When on, a claude.ai OAuth login with no known subscription type (null) counts as eligible for organization policy limits; otherwise an OAuth login is eligible only on an enterprise or team subscription. A login with an API key is eligible either way, and the other checks (first-party API, the token's scope) still apply.v2.1.138 or earlier → v2.1.148 ○gone AND v2.1.308 tengu_slate_lantern_hush ◇ Artifact live-connection keepalive pings can now stretch out during idle periodsv2.1.265 → v2.1.309 ◆in the build never read tengu_slate_meadow ◇ Default false. When true, enables the background-agents fleet gate (isAgentsFleetEnabled) and adds the built-in Claude agent to the built-in agent list, unless agent view is disabled: by CLAUDE_CODE_DISABLE_AGENT_VIEW, the disableAgentView setting, or a third-party API provider. Both are also on without the flag when Claude Code runs in essential-traffic-only mode (for example with CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC set).v2.1.138 or earlier → v2.1.140 ○gone AND v2.1.308 tengu_slate_moth ◇ Default false. Makes the context tip large-context-stale-files eligible: when the context is over 300,000 tokens and more than 100,000 of them are stale file tokens, the tip suggests /compact.v2.1.191 → v2.1.211 ○gone AND v2.1.308 tengu_slate_prism ◇ Default true. When an SDK initialize control request sets agentProgressSummaries, Claude Code turns the agent progress summary setting on only while this flag is true; with the flag false that request field has no effect at this site.v2.1.138 or earlier → v2.1.245 ○gone never read tengu_slate_quill ◇ Default false. One of the gates for the DesignSync tool and the /design, /design-consent, /design-revoke and /design-login commands. They are off when the allow_design_sync policy check fails or another internal check disables them, on when a further internal check passes, and otherwise follow this flag.v2.1.160 → v2.1.245 ○gone AND v2.1.308 tengu_slate_reef ◇ No description from anybody.never mined ·never mined never read tengu_slim_subagent_claudemd ◇ Default true. When on, subagents whose definition sets omitClaudeMd (the built-in Explore and Plan agents in this version) start without the CLAUDE.md content in their user context, unless the caller passes its own user context. When off, they receive the CLAUDE.md content like other subagents.v2.1.138 or earlier → v2.1.195 ○gone AND v2.1.308 tengu_sm_compact ◇ No description from anybody.never mined ·never mined never read tengu_snug_wall ◇ Claude Code registers two new switches that Anthropic's servers control; what they turn on is not visible yetv2.1.309 ●new never read tengu_sonnet45_to_46_migration ◇ An analytics event, not a feature flag. Claude Code logs it when its startup migration rewrites a user-settings model of Sonnet 4.5 (claude-sonnet-4-5-20250929 or sonnet-4-5-20250929, with or without [1m]) to sonnet or sonnet[1m], which it does only on the first-party provider for Pro, Max or Team (Max 5x tier) subscribers. Fields: from_model and has_1m.never mined ·never mined never read tengu_sorrel_trellis_finial ◇ Artifact comment replies get grouped wakes and a stand-down when everything was already answered, both switched off.never mined ·never mined never read tengu_sorrel_trellis_lintel ◇ A comment sent moments before a watch starts could get a reply, but the flag is off.v2.1.236 → v2.1.267 ○gone AND v2.1.308 tengu_sorrel_trellis_weir ◇ How many turns Claude takes automatically can now be set from the server, still defaulting to 60.v2.1.227 → v2.1.233 ○gone never read tengu_sorted_truffle ◇ Default true, read once per session. When on, the SuggestPluginInstall tool fetches details for each requested plugin ID it has not already resolved before it renders install cards; IDs whose lookup fails are counted as unchecked. When off, no lookup runs.v2.1.283 → v2.1.299 ○gone never read tengu_sparrow_ledger ◇ Default false. When on, adds a Work step by step: section to the system prompt telling the model to reproduce the symptom before editing, edit the source, then re-observe the symptom until it is gone. The CLAUDE_CODE_VERIFY_PROMPT environment variable turns on the same section.v2.1.138 or earlier → v2.1.205 ○gone AND v2.1.308 tengu_speculation ◇ An analytics event, not a feature flag. Claude Code logs it when a speculation run (a forked agent run ahead on a suggestion) ends with outcome accepted, aborted or error. Fields include speculation_id, outcome, duration_ms, suggestion_length, tools_executed (successful tool results), completed and the boundary_type, boundary_tool and boundary_detail where it stopped, plus abort_reason or error_type, error_message and error_phase where they apply.never mined ·never mined never read tengu_spicy_pearl ◇ No description from anybody.v2.1.299 → v2.1.309 ◆in the build AND v2.1.308 tengu_spinner_stall_cleared ◇ Spinner stall detection now resets when the active agent changes instead of carrying state across agents.never mined ·never mined never read tengu_sprightly_lagoon ◇ Default false. When on, the checks for per-turn effort and per-turn timing both report unsupported, so requests carry no per-turn effort or now values and those checks add no per-turn-control-2026-07-01 beta.v2.1.260 → v2.1.301 ○gone AND v2.1.308 tengu_starry_waterfall ◇ No description from anybody.v2.1.304 → v2.1.309 ◆in the build AND v2.1.308 tengu_startup_flag_fetch_wait ◇ New telemetry tracks how long startup waits on remote feature-flag fetchesnever mined ·never mined never read tengu_startup_notice ◇ Default "". A non-empty string is shown as a high-priority, warning-colored notification for 30 seconds. The value is re-read when flag values update: a new string replaces the notification and an empty one removes it.v2.1.140 → v2.1.309 ◆in the build AND v2.1.289 tengu_startup_telemetry ◇ Startup telemetry now reports theme as one of a known set of names, labeling anything else 'custom'never mined ·never mined never read tengu_stateful_mitten ◇ No description from anybody.v2.1.296 → v2.1.309 ◆in the build never read tengu_stdout_backpressure ◇ New telemetry tracks stdout backpressure stalls and dropped bytes.never mined ·never mined never read tengu_steady_plum ◇ The effort-downgrade nudge now works per model instead of just one, but each model needs its own flag turned on.never mined ·never mined AND v2.1.308 tengu_stop_hook_added ◇ Goals now record whether you set them explicitly or they came from somewhere else.never mined ·never mined never read tengu_stop_hook_command ◇ No description from anybody.never mined ·never mined never read tengu_streaming_tool_execution ◇ No description from anybody.never mined ·never mined never read tengu_streaming_tool_execution2 ◇ Default false. Turns on streaming tool execution in the query loop: tool calls are handed to an executor as their tool_use blocks arrive while the model's response is still streaming, completed results are emitted during the stream, and the remaining results are collected after it ends. The executor is discarded and recreated when a response is retried or restarted.v2.1.138 or earlier → v2.1.153 ○gone AND v2.1.308 tengu_structured_output_retraction_exhausted ◇ A missing timing field in a stream event can no longer crash a headless turn.never mined ·never mined never read tengu_structured_output_schema_lint ◇ StructuredOutput schemas are now linted for internally-contradictory constraints.never mined ·never mined never read tengu_subagent_cache_evict ◇ Default false. When on, model requests marked evictCacheOnComplete that carry sticky betas add the prompt-caching-evict-2026-05-12 beta. Provider checks come first: the request must go to the first-party API at a first-party base URL (or Claude Platform on AWS with no ANTHROPIC_AWS_BASE_URL), with neither CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS set nor HIPAA mode active, and under those conditions a truthy CLAUDE_CODE_SUBAGENT_CACHE_EVICT turns it on without consulting the flag.v2.1.198 → v2.1.309 ◆in the build AND v2.1.308 tengu_subagent_output_flagged ◇ An analytics event, not a feature flag. Claude Code logs it when a finished subagent's output, scanned against a list of patterns at the finalize step, has at least one reportable match. Fields: agent_id, surface, patterns, categories and match_count.never mined ·never mined never read tengu_subagent_resume_sidecar_read ◇ Resuming a subagent now logs diagnostic details about how it's being resumednever mined ·never mined never read tengu_subagent_steer_applied ◇ Subagent delegation nudges are now switched off for Opus 5; you can also control them by env var.never mined ·never mined never read tengu_subagent_type_miss ◇ The Task tool now rejects launches missing subagent_type instead of silently picking a default agentnever mined ·never mined never read tengu_subagent_type_normalized ◇ Subagent-type mismatch telemetry now records the length of the requested name; forked teammates flag an inherited system promptnever mined ·never mined never read tengu_sumi ◇ No description from anybody.never mined ·never mined never read tengu_summarize_tool_results ◇ No description from anybody.never mined ·never mined never read tengu_swift_garden ◇ Default true. When on, opening the /model picker can start an entitlement probe for the Fable model (a forced server-side validation of the model, at most 2 attempts, with the picker waiting up to 15000 ms for it), subject to further conditions such as the first-party API provider. The result is recorded as accepted or refused. false skips the probe.v2.1.260 → v2.1.276 ○gone never read tengu_swinburne_dune ◇ No description from anybody.never mined ·never mined never read tengu_switch_to_subscription_notice_shown ◇ An analytics event, not a feature flag. Claude Code logs it each time it records an impression of the subscription-switch notification, incrementing that notification's seen count in the global config; the notice stays eligible while the count is below 3. It carries no fields.never mined ·never mined never read tengu_sync_plugin_install_timeout ◇ Internal plugin sync and MCP relay functions were renamed with no behavior changenever mined ·never mined never read tengu_tallow_quoit_gate ◇ No description from anybody.v2.1.307 → v2.1.309 ◆in the build AND v2.1.308 tengu_team_mem_basis_unusable ◇ Team memory sync adds structured failure reasons for loading its sync baseline, plus a new account-wide memory scopenever mined ·never mined never read tengu_team_mem_conflict_recovered ◇ Shared memory conflict notices now distinguish a write that never saw the shared copy.never mined ·never mined never read tengu_team_mem_foreign_partition_recovered ◇ Shared team memory mounts stop clobbering each other and switch back on once a conflict clears.never mined ·never mined never read tengu_team_mem_multistore_sync ◇ You now get told when a memory file is too big to sync to shared memory.never mined ·never mined never read tengu_team_mem_push_delete_deferred ◇ Team memory sync can now skip an individual refused file without breaking the whole sync, and reports how many files were refusednever mined ·never mined never read tengu_team_mem_push_suppressed ◇ Failed memory pushes no longer stay suppressed for your whole session; retries resume normally.never mined ·never mined never read tengu_team_onboarding_generated ◇ Team onboarding still pulls its prompt and lookback window from remote config.never mined ·never mined never read tengu_teleport_bundle_read_rules ◇ Before packaging your checkout, you are told if broken settings mean only committed files go up.never mined ·never mined never read tengu_teleport_bundle_started ◇ Handing work to a remote session now shows bundling progress and names the fallback on failure.never mined ·never mined never read tengu_teleport_client_directory_sync ◇ Cloud session creation records whether directory sync was declared to the server.never mined ·never mined never read tengu_teleport_client_directory_sync_retracted ◇ Cloud sessions can ask the server to drop their directory-sync tag.never mined ·never mined never read tengu_teleport_first_message_error ◇ An analytics event, not a feature flag. In a teleported session, Claude Code logs it when the first model request fails, then marks the first message as logged. Fields: session_id and error_type.never mined ·never mined never read tengu_teleport_first_message_success ◇ An analytics event, not a feature flag. In a teleported session, Claude Code logs it when the first model request succeeds, then marks the first message as logged. Field: session_id.never mined ·never mined never read tengu_teleport_menu ◇ The /teleport menu was rebuilt and now tracks what you chose, but stays gated off.never mined ·never mined never read tengu_teleport_permission_mode_refused ◇ No description from anybody.v2.1.299 ○gone never read tengu_teleport_resume_error ◇ 'Claude Code on the web' is renamed to 'cloud session(s)' across error messages, menus, and commandsnever mined ·never mined never read tengu_teleport_source_decision ◇ Teleport checks whether your local checkout diverged from the remote before deciding what to seed into the container.never mined ·never mined never read tengu_teleport_to_cloud ◇ /teleport can hand your current session to the cloud if your git tree is clean and pushed.never mined ·never mined never read tengu_terminal_probe ◇ Claude Code now identifies your terminal emulator and its version and reports it in a startup telemetry eventnever mined ·never mined never read tengu_tern_alloy ◇ Default "off". Any other value makes the subagent-fanout-nudge tip eligible, with a 3-session cooldown, on the first-party API for users not signed in with a claude.ai subscription. The value copy_b shows For big tasks, tell Claude to use subagents. They work in parallel and keep your main thread clean.; any other non-off value shows Say "fan out subagents" and Claude sends a team. Each one digs deep so nothing gets missed.v2.1.138 or earlier → v2.1.140 ○gone AND v2.1.308 tengu_tether_decision ◇ New gated 'tether' system can thread API requests together server-side instead of resending full context each turnnever mined ·never mined never read tengu_tether_echo_audit ◇ New telemetry events track the tether feature's threading decisions, outcomes, and echo auditsnever mined ·never mined never read tengu_tether_live_outcome ◇ New gated 'tether' system can thread API requests together server-side instead of resending full context each turnnever mined ·never mined never read tengu_tether_unrecognised_400 ◇ Claude Code added a new telemetry event, tengu_tether_unrecognised_400, for the thread-continuation systemnever mined ·never mined never read tengu_thinking_drop ◇ New telemetry event tengu_thinking_drop logs when thinking blocks get stripped from a requestnever mined ·never mined never read tengu_thinking_toggled_hotkey ◇ An analytics event, not a feature flag. Claude Code logs it when the thinking toggle hotkey turns extended thinking on or off, after updating thinkingEnabled and sending a set_max_thinking_tokens control request. Field: enabled.never mined ·never mined never read tengu_tide_elm ◇ Default "off". Any other value makes the effort-high-nudge tip, which suggests /effort high and has a 3-session cooldown, eligible on the first-party API for users not signed in with a claude.ai subscription, when the current model supports effort, no policy effortLevel is set, CLAUDE_CODE_EFFORT_LEVEL is unset or invalid (its values auto and unset also hide the tip), and the effortLevel setting is not high, xhigh or max. The value copy_b shows Use /effort high for better one-shot answers. Claude thinks it through first.; any other non-off value shows Working on something tricky? /effort high gives better first answers.v2.1.138 or earlier → v2.1.140 ○gone AND v2.1.308 tengu_timber_lark ◇ Default "off". Controls the loop-command-nudge tip for /loop: "off" hides it, "copy_b" shows Use /loop 5m check the deploy to run any prompt on a schedule. Set it and forget it., and any other value shows /loop runs any prompt on a recurring schedule. Great for monitoring deploys, babysitting PRs, or polling status. The tip also needs cron scheduling to be enabled (CLAUDE_CODE_DISABLE_CRON not set true and tengu_kairos_cron on) plus a further eligibility check, and has a cooldown of 3 sessions.v2.1.138 or earlier → v2.1.140 ○gone AND v2.1.308 tengu_time_based_microcompact ◇ An analytics event, not a feature flag. Claude Code logs it when a keep-recent microcompaction with the context_hint trigger clears older tool results, keeping the most recent ones, and saves at least a minimum token count. Fields: toolsCleared, toolsKept, keepRecent, tokensSaved and trigger.never mined ·never mined never read tengu_time_format_setting_changed ◇ A new Time format setting in /config controls whether UI clocks use 12-hour, 24-hour, UTC or a custom pattern.never mined ·never mined never read tengu_timer_delay_clamped ◇ Claude Code now clamps overly long setTimeout/setInterval delays instead of letting them silently misbehavenever mined ·never mined never read tengu_tingly_hopper ◇ Always-loaded claude.ai MCP tools are held out of the declared tool set and loaded lazily until the tool-use decision settles.v2.1.257 → v2.1.267 ○gone AND v2.1.308 tengu_tip_shown ◇ Tip telemetry now labels plugin-marketplace tips by their actual marketplace instead of a generic bucketnever mined ·never mined never read tengu_toasty_breeze ◇ Hitting the 5-hour limit could offer to keep working at reduced priority, but it is off for everyone.v2.1.239 → v2.1.309 ◆in the build AND v2.1.308 tengu_toasty_fiddle ◇ No description from anybody.v2.1.305 → v2.1.309 ◆in the build never read