Feature flags, all 2,081.

tengu_osier_pylon_trace ◇ Claude could read console errors and failed loads from real viewers of an artifact, but it's switched off.v2.1.238 → v2.1.308 ◆in the build AND v2.1.280 tengu_pale_kestrel ◇ Permission prompts for inline shell commands such as bash -c now depend on whether the command contains rm or rmdirv2.1.299 → v2.1.308 ◆in the build AND v2.1.299 tengu_paper_halyard ◇ Internal rewrite of how instruction files get filtered from context, no visible change.v2.1.138 or earlier → v2.1.308 ◆in the build AND v2.1.252 tengu_parsed_eagle ◇ Claude Code can re-encode images as WebP when that is smaller, and now refuses images that are too costly to decodev2.1.295 → v2.1.308 ◆in the build AND v2.1.295 tengu_parsed_willow ◇ New tengu_parsed_willow gate (defaulting on) added with CLAUDE_CODE_PARSED_WILLOW env overridev2.1.277 → v2.1.308 ◆in the build AND v2.1.281 tengu_patient_finch ◇ If a URL failed earlier in the session, WebFetch skips a quick second attempt and tells the model to say the page could not be readv2.1.300 → v2.1.308 ◆in the build AND v2.1.300 tengu_patient_meteor ◇ When a subagent edits a file, the Edit tool can attach a compact summary result, off unless switched on remotelyv2.1.295 → v2.1.308 ◆in the build AND v2.1.295 tengu_peaceful_leaf ◇ A remote switch can add guidance on files and folders on your computer to the browser and computer-use tool descriptionsv2.1.288 → v2.1.308 ◆in the build AND v2.1.299 tengu_penguins_off ◇ Fast mode ('penguin mode') gains an admin setting to require per-session opt-in or disable it, plus an org-check bypass env varv2.1.138 or earlier → v2.1.308 ◆in the build AND v2.1.299 tengu_per_turn_effort ◇ Per-family capability gates now use a name-to-flag table instead of matching skill and tool name strings.never mined ·never mined AND v2.1.308 tengu_pewter_avocet ◇ Claude Code now re-sends a goal check-in prompt if a turn ends before the goal could be evaluatedv2.1.269 → v2.1.308 ◆in the build AND v2.1.269 tengu_pewter_kite_ms ◇ A background query could narrate what Claude is doing under the spinner, off unless enabled.v2.1.234 → v2.1.283 ○gone AND v2.1.234 tengu_pewter_lark ◇ Default "off". Controls the goal-command-nudge tip for /goal (3-session cooldown) for users signed in with a claude.ai subscription: any value other than off makes the tip eligible, and the value picks the copy: copy_orig ("Set an objective with /goal"), copy_b ("Use /goal all tests pass"), or any other value for the default "/goal keeps Claude working until a condition is met" text. On the first-party API without a claude.ai subscription login the tip is eligible with the copy_orig text whatever the flag says, and in other setups, such as third-party providers, it stays off.v2.1.141 → v2.1.179 ○gone AND v2.1.308 tengu_pewter_ledger ◇ No description from anybody.never mined ·never mined AND v2.1.122 tengu_pewter_lintel ◇ Several keypresses arriving at once can now be drawn one by one without setting CLAUDE_CODE_COMMIT_BETWEEN_KEYSv2.1.283 ○gone AND v2.1.284 tengu_pewter_orchard ◇ No description from anybody.v2.1.210 → v2.1.231 ○gone AND v2.1.308 tengu_pewter_summit ◇ No description from anybody.v2.1.145 → v2.1.308 ◆in the build AND v2.1.308 tengu_plaid_harbor ◇ The names "claude-ai" and "anthropic-skills" are now reserved for skills synced from claude.ai, and local skills or MCP servers using them are refusedv2.1.283 → v2.1.308 ◆in the build AND v2.1.282 tengu_plain_kestrel ◇ Claude Code now uses a call's origin to tell whether a plugin's own tool call was rewritten by another hookv2.1.300 → v2.1.308 ◆in the build AND v2.1.300 tengu_plank_river_frost ◇ No description from anybody.never mined ·never mined AND v2.1.14 tengu_playful_hare ◇ /recap now refuses requests relayed from Slack, Teams, project threads, routines or other programs, and only answers you directlyv2.1.284 → v2.1.308 ◆in the build AND v2.1.284 tengu_playful_lobster ◇ Read-permission checks now detect hardlinked files and log it as shadow telemetry, without changing the decisionv2.1.235 → v2.1.308 ◆in the build AND v2.1.235 tengu_playful_panda ◇ Attachment uploads can now send only the original file, and chat uploads report whether the copy is an image and its file typev2.1.285 → v2.1.308 ◆in the build AND v2.1.292 tengu_plugin_autoupdate_allow_credential_helper ◇ Plugin autoupdate now always disables the git credential helper during background refreshv2.1.174 → v2.1.278 ○gone AND v2.1.280 tengu_plugin_command_source_refresh ◇ Command-built plugins refresh on their own background schedule, and block if the command changed.v2.1.229 → v2.1.308 ◆in the build AND v2.1.300 tengu_plugin_hooks_modules ◇ Tools a plugin registers reach the model through a local loopback MCP server.v2.1.242 → v2.1.308 ◆in the build AND v2.1.283 tengu_plugin_input_refusal ◇ Plugin calls that submit a prompt or run a command are refused if you stopped the session and haven't typed anything sincev2.1.299 → v2.1.308 ◆in the build AND v2.1.299 tengu_plugins_npm_marketplace ◇ Plugins can now be installed directly from the npm registryv2.1.275 → v2.1.308 ◆in the build AND v2.1.275 tengu_polished_dewdrop ◇ A new hidden switch, settable via env var, can make certain responses silently drop or error instead.v2.1.263 → v2.1.308 ◆in the build AND v2.1.263 tengu_polished_lagoon ◇ One session could inject a coordination notice into another, but three separate gates keep it dark.v2.1.234 → v2.1.308 ◆in the build AND v2.1.308 tengu_polished_reef ◇ Remote SDK sessions can prepare attached files while still waiting to take over a turn, behind a switch that is off by defaultv2.1.282 → v2.1.286 ○gone AND v2.1.286 tengu_porch_bell_9f ◇ Default "". A non-empty string is shown as a high-priority warning notification for 30000 ms, and is re-checked when flag values update; an empty value removes the notification. Not shown when the workspace is remote.v2.1.138 or earlier → v2.1.139 ○gone AND v2.1.308 tengu_pr_footer_surface_suffix ◇ Default false. When true, the default PR attribution footer Generated with Claude Code gets via <surface> appended, with the surface named from CLAUDE_CODE_ENTRYPOINT: Claude Desktop, Mobile, Cowork, Claude Tag in Slack, Claude Tag in Teams or GitHub Actions. Other entrypoints get no suffix, and an attribution.pr setting still replaces the default footer.v2.1.211 → v2.1.308 ◆in the build AND v2.1.211 tengu_precious_koala ◇ Claude Code has a new remotely controlled switch, off by default, that can turn on an unnamed feature everywhere or only in Coworkv2.1.299 → v2.1.308 ◆in the build AND v2.1.299 tengu_print_ceiling_stop_agents ◇ Behind a flag, exceeding the print-mode background wait ceiling can also stop the running agents, not just report them.v2.1.257 → v2.1.308 ◆in the build AND v2.1.295 tengu_proactivity_selector ◇ A --proactivity flag picks the proactivity level when Claude Code starts, and CLAUDE_CODE_DISABLE_PROACTIVITY turns the selector offv2.1.290 → v2.1.308 ◆in the build AND v2.1.290 tengu_projects_thread_recap ◇ A richer "recap" format for project task status is being built, gated off by defaultv2.1.265 → v2.1.308 ◆in the build AND v2.1.265 tengu_projects_thread_recap_prompt ◇ A richer "recap" format for project task status is being built, gated off by defaultv2.1.265 → v2.1.308 ◆in the build AND v2.1.265 tengu_projects_thread_reply_lead_killswitch ◇ Claude Code Projects threads can now label replies as sent directly, relayed, or from the project timelinev2.1.274 → v2.1.308 ◆in the build AND v2.1.274 tengu_prompt_cache_1h_config ◇ Whether a one-hour prompt cache is allowed can now be overridden per agent and can ignore overage.v2.1.138 or earlier → v2.1.308 ◆in the build AND v2.1.293 tengu_prompt_suggestion ◇ Matching your reply to a suggested prompt is now case-insensitivenever mined ·never mined AND v2.1.268 tengu_propose_goal ◇ A tool letting Claude propose a session goal you approve with one keypress is present but gated off.v2.1.227 → v2.1.308 ◆in the build AND v2.1.227 tengu_propose_skills ◇ An env var turns on the tool that proposes new skills; otherwise it stays off in this build.v2.1.212 → v2.1.250 ○gone AND v2.1.251 tengu_proud_clover ◇ A new "max effort" mode appears with a footer hint pointing to the /effort commandv2.1.280 → v2.1.308 ◆in the build AND v2.1.292 tengu_ptc_enabled ◇ Internal cleanup of the SDK control-message handlers, no functional change foundv2.1.206 → v2.1.308 ◆in the build AND v2.1.301 tengu_purring_pine ◇ Behind a server-controlled switch, the tool that delivers a report to its caller can now make the agent stop straight afterv2.1.285 → v2.1.308 ◆in the build AND v2.1.285 tengu_purring_yeti ◇ A list Claude Code fetches from its remote configuration is now read under a different name, so values set under the old name no longer applynever mined ·never mined AND v2.1.290 tengu_quartz_pipit ◇ Default true. When a main-thread turn under an active /goal ends in a way Claude Code classifies as unrecoverable (a context limit such as prompt_too_long or blocking_limit), the goal is cleared, its Stop hook removed, and the warning Goal cleared after an unrecoverable error (...) is shown. When false, that case returns without clearing the goal.v2.1.234 → v2.1.308 ◆in the build AND v2.1.308 tengu_quiet_dolphin ◇ The /diff panel is now prototyped internally as a plugin pane rather than built-in UInever mined ·never mined AND v2.1.308 tengu_quiet_elephant ◇ A new generator builds 5-character random names and retries up to 10 times to avoid a list of banned wordsv2.1.290 → v2.1.308 ◆in the build AND v2.1.290 tengu_quiet_heron ◇ Continuing a server-side conversation thread can now skip resending the system prompt and tools list, with a fallback if the server disagreesv2.1.269 → v2.1.308 ◆in the build AND v2.1.269 tengu_quiet_lantern ◇ A new gate, tengu_quiet_lantern, controls whether the server classifier falls back to local when unsupportedv2.1.278 → v2.1.308 ◆in the build AND v2.1.278 tengu_quill_marten ◇ Claude Code can fetch context for a claude.ai Project into the session, behind a switch that is off unless turned on remotelyv2.1.299 → v2.1.308 ◆in the build AND v2.1.299 tengu_quirky_kazoo ◇ A list Claude Code fetches from its remote configuration is now read under a different name, so values set under the old name no longer applynever mined ·never mined AND v2.1.292 tengu_quirky_shore ◇ A new switch, off by default, makes computer-use actions in auto mode go through further permission checks even when an allow rule covers the whole toolv2.1.284 → v2.1.308 ◆in the build AND v2.1.284 tengu_quirky_sketch ◇ Claude Code can now single out its device tools for special handling, under a server flag that is off unless switched onnever mined ·never mined AND v2.1.290 tengu_quirky_teacup ◇ New warning notification when the OpenTelemetry headers helper fails, pointing to /statusv2.1.275 → v2.1.308 ◆in the build AND v2.1.275 tengu_quizzical_giraffe ◇ A remotely controlled switch can stop Claude Code from fetching managed settings early at startup, and a message covers approvals it cannot ask forv2.1.292 → v2.1.308 ◆in the build AND v2.1.292 tengu_quizzical_thimble ◇ Terminal title can now be suppressed for the SDK-CLI entrypoint, behind a new flagv2.1.277 → v2.1.308 ◆in the build AND v2.1.277 tengu_quizzical_toucan ◇ In remote sessions, queued notifications now carry a timestamp of when they arrivedv2.1.292 → v2.1.308 ◆in the build AND v2.1.292 tengu_quizzical_wirth ◇ Claude Code now counts turns since a person last wrote and can act once a limit is reached, but this is off unless switched on remotelyv2.1.296 → v2.1.306 ○gone AND v2.1.296 tengu_radiant_island ◇ A one-time prompt offering to switch your default effort to medium exists but has no text.v2.1.240 → v2.1.261 ○gone AND v2.1.263 tengu_radiant_teapot ◇ A new internal helper preserves the permission mode a tool request was made under when replying to it in 'arbiter' modev2.1.275 → v2.1.308 ◆in the build AND v2.1.275 tengu_rate_limit_promo_notices ◇ Default []. A list of { bar, text, variant } notices for the usage limit bars (Current session, Current week (all models) and the rest): each notice whose bar matches a bar's key case-insensitively, such as five_hour, seven_day or seven_day_sonnet, is rendered as an extra line under that bar. variant must be a key of the dark theme or it is dropped; a list that fails validation shows nothing.v2.1.208 → v2.1.308 ◆in the build AND v2.1.308 tengu_rc_long_turn_nudge ◇ The long-turn nudge feature no longer accepts a simple on/off setting; it now requires a full config objectv2.1.208 → v2.1.308 ◆in the build AND v2.1.280 tengu_rc_permission_nudge ◇ An object {afterPromptCount, probability, maxImpressions} configuring the "Approve tool calls from your phone" hint, with code default {afterPromptCount: 5, probability: 0, maxImpressions: 3}. The hint can appear only while Remote Control is active with a session URL, after afterPromptCount permission prompts, with random chance probability, and at most maxImpressions times; with the default probability of 0 it never shows. The CLAUDE_CODE_RC_PERMISSION_NUDGE environment variable (JSON) overrides it.v2.1.208 → v2.1.308 ◆in the build AND v2.1.308 tengu_reactive_quilt ◇ Cloud sessions gain a way to hand answered file writes to the next worker and place them in the home folder, off unless switched on remotelyv2.1.288 → v2.1.308 ◆in the build AND v2.1.288 tengu_record_created_pr_to_ccr ◇ Pull requests you create could be reported to a cloud endpoint, but nothing is sent here.v2.1.234 → v2.1.308 ◆in the build AND v2.1.234 tengu_recursive_lobster ◇ Claude Code registers a new remote switch whose purpose is not yet visiblev2.1.307 → v2.1.308 ◆in the build AND v2.1.307 tengu_reflective_bee ◇ A new dialog can offer to connect GitHub to claude.ai after a git push, not yet confirmed livev2.1.267 → v2.1.308 ◆in the build AND v2.1.308 tengu_reflective_wolf ◇ A new gated mechanism lets Claude Code tell the model when an 'advisor tool' becomes available or unavailablev2.1.268 → v2.1.308 ◆in the build AND v2.1.265 tengu_remote_backend ◇ Joining an existing cloud session, trusted-device enrollment and the folder sync offer can now be turned on by two new remote switches, off by defaultv2.1.138 or earlier → v2.1.308 ◆in the build AND v2.1.305 tengu_repl_mcp_error_throw ◇ The gate behind MCP error-throwing now reads under the name tengu_tranquil_fern instead of tengu_repl_mcp_error_throwv2.1.217 → v2.1.274 ○gone AND v2.1.275 tengu_resume_open_live_bg ◇ --resume on a session running in the background now attaches to it and hands over your prompt instead of refusingv2.1.285 → v2.1.308 ◆in the build AND v2.1.285 tengu_review_bughunter_config ◇ The cloud review poll timeout is no longer hardcoded at 30 minutes and now comes from remote config.v2.1.138 or earlier → v2.1.308 ◆in the build AND v2.1.259 tengu_rewind_first_message ◇ Default false. Changes the SDK rewind_conversation control request: when false, the rewind is anchored at the nearest assistant message before the target and fails with no preceding assistant if there is none. When true, it is anchored at the nearest user or assistant message before the target, that error is not returned (so the first message can be the target), and the anchor is persisted with { rewound: true }.v2.1.191 → v2.1.245 ○gone AND v2.1.308 tengu_rosy_donut ◇ A new internal flag, on by default, gates some turn and line attribution behavior.v2.1.257 → v2.1.308 ◆in the build AND v2.1.257 tengu_rosy_finch ◇ An MCP tool call's timeout can now pause while it waits on a question to you, but only if this is switched on remotelyv2.1.305 → v2.1.308 ◆in the build AND v2.1.305 tengu_rosy_flute ◇ WebFetch now measures how often the same web page is fetched again, without changing what it returnsv2.1.281 → v2.1.308 ◆in the build AND v2.1.307 tengu_rosy_pine ◇ A prompt can offer to continue a resumed old session on a newer model of the same family; it is off unless switched on remotelyv2.1.286 ○gone AND v2.1.286 tengu_russet_linnet ◇ Default false. When on, the Skill tool's prompt uses the shorter Invoke a skill. text in place of the default Execute a skill within the main conversation text. Setting CLAUDE_CODE_SKILL_DESC_REFRAME to true turns this on regardless of the flag.v2.1.196 → v2.1.215 ○gone AND v2.1.196 tengu_russet_whimbrel ◇ No description from anybody.v2.1.295 → v2.1.308 ◆in the build AND v2.1.308 tengu_rustling_crescent ◇ Claude Code can list the subagents that ran earlier in a session, off unless CLAUDE_CODE_SUBAGENT_HISTORY is set or it is switched on remotelyv2.1.307 → v2.1.308 ◆in the build AND v2.1.307 tengu_rustling_pixel ◇ Claude Code can now keep extended-thinking content across any model switch, not just upgradesv2.1.280 → v2.1.308 ◆in the build AND v2.1.280 tengu_rustling_sunset ◇ Claude Code can now show a heartbeat message while waiting on a rate-limit retryv2.1.280 → v2.1.308 ◆in the build AND v2.1.280 tengu_sable_orrery ◇ Behind a server-controlled switch, the auto mode startup dialog now changes some state before it shows and restores it afterv2.1.305 → v2.1.308 ◆in the build AND v2.1.305 tengu_saffron_credits_only_tiers ◇ No description from anybody.v2.1.190 → v2.1.308 ◆in the build AND v2.1.190 tengu_saffron_ferry ◇ Claude Code reads a new on/off switch from the server, and treats it as on when the server sends nothingv2.1.300 → v2.1.308 ◆in the build AND v2.1.300 tengu_saffron_ladder ◇ The remote setting behind the Fable upgrade offer now only turns a fixed "Upgrade to Max" option on or off, with no custom textv2.1.216 → v2.1.308 ◆in the build AND v2.1.282 tengu_saffron_lattice ◇ Whether a model needs consent or usage credits can now be decided per model rather than for the whole planv2.1.170 → v2.1.308 ◆in the build AND v2.1.299 tengu_saffron_picker_dim ◇ Default false. When on, and the cached extra-usage disabled reason is org_level_disabled or overage_not_provisioned (along with further account checks), model picker options whose value is a Fable model are shown disabled with the label Fable (disabled), and their description can gain a requires usage credits note.v2.1.190 → v2.1.308 ◆in the build AND v2.1.190 tengu_sage_compass ◇ No description from anybody.never mined ·never mined AND v2.1.98 tengu_sage_compass2 ◇ An internal advisor model concept exists, with sub-agents told never to consult it, and only runs under first-party auth plus a flag.v2.1.138 or earlier → v2.1.308 ◆in the build AND v2.1.282 tengu_sedge_lantern ◇ /recap and the Session recap settings entry now work for everyone without an experiment flag.v2.1.138 or earlier → v2.1.222 ○gone AND v2.1.223 tengu_send_file ◇ Sending files between sessions now fails with a clear "not available in this session" message.v2.1.227 → v2.1.308 ◆in the build AND v2.1.290 tengu_send_user_file ◇ Under zero data retention, the file-send tool no longer appears at all.v2.1.142 → v2.1.308 ◆in the build AND v2.1.300 tengu_sepia_cormorant ◇ Default null. A list of model name substrings: when the session's current model contains one (case-insensitive) and tengu_umber_petrel is also on, auto memory defaults to off for that model and its status reads unavailable for current model with the toggle disabled. CLAUDE_CODE_DISABLE_AUTO_MEMORY set to a false value still forces auto memory on; a non-array or empty value has no effect.v2.1.138 or earlier → v2.1.308 ◆in the build AND v2.1.308 tengu_sequential_cloud ◇ Claude Code can ping Anthropic's servers to keep a one-hour prompt cache alive during idle pauses, but it is off unless switched on remotelyv2.1.286 → v2.1.308 ◆in the build AND v2.1.286 tengu_sequential_puffin ◇ Resuming a session another process still controls can decline the takeover and tell you why.v2.1.232 → v2.1.308 ◆in the build AND v2.1.232 tengu_serene_teacup ◇ A new step at session start refreshes an internal setting called the atis pinv2.1.282 → v2.1.308 ◆in the build AND v2.1.282