You can now move a session on your computer into a new cloud session. Claude Code copies the newest turns after checking the move is safe. A new /project command names the claude.ai Project for usage reports, and --save keeps that choice. Admins can set disableReplyCheck in managed settings to stop Claude being re-prompted when a turn ends without a reply. Setting CLAUDE_CODE_ULTRAREVIEW_LAUNCH_RECORD to a file path makes each successful ultrareview launch write its session id there. This release has 425 entries, and 240 of them are usable today.
26 entries are in the build but not switched on yet. The biggest is an experimental REPL mode, where Claude writes JavaScript that calls its tools. It comes with an opt-in dialog and limits on resuming and moving sessions. More pieces of a code-running mode landed this time, including script tracking and a tool that waits while idle. A check that can skip compacting a long conversation is also here, but it stays off unless switched on remotely. Session-long "always allow" choices for MCP connector tools, which link Claude to outside services, are recorded only when a remote switch is on.
Some of this release's fixes make broken config files safer to deal with. If the config file is corrupt or empty, Claude Code looks for the best backup and tells you what it did. A global config file that cannot be read is now left alone instead of being overwritten. A failed message send in a remote session no longer leaves it stuck sending. Cloud sessions that fail during setup now wait for the refusal message, so you can see why the launch failed. Saving over a file mounted into a container now works when the file is busy.
Written by our agent from the shipped bundle, not by Anthropic.
A new path moves a local session to a cloud session by copying its recent turns, via /teleport and claude --cloud, with many safety checks
Sessions & agentsSelf-hosted runner reply-check behaviour can now be turned off with a managed disableReplyCheck setting or shaped by a CLAUDE_CODE_REPLY_GATE_POLICY variable
ElsewhereSetting CLAUDE_CODE_EXPERIMENTAL_CODE_MODE asks Claude Code to switch on an experimental Code Mode, which it can refuse for several reasons
A new /project command, gated by tengu_gentle_cupcake, sets attributionProject for usage reports; IDs may start with chan_ and --settings is a source
ExtensionsCLAUDE_CODE_MEMORY_PUSH_DELETE_MODE now accepts recorded, and Claude is told how to delete a shared-memory file with a plain rm command
Sessions & agentsWant the reasoning? Read walks the 46 entries that probably matter to you, each one opening to what changed and why.
Read this release → Every row →Anything you can use today, anything that visibly changes, and anything worth poking at. One line each, open for detail.
A new path moves a local session to a cloud session by copying its recent turns, via /teleport and claude --cloud, with many safety checks
Unclear It is not clear whether /teleport or --cloud reach this move for anyone using this version, since the feature depends on a remote switch.
Self-hosted runner reply-check behaviour can now be turned off with a managed disableReplyCheck setting or shaped by a CLAUDE_CODE_REPLY_GATE_POLICY variable
Setting CLAUDE_CODE_EXPERIMENTAL_CODE_MODE asks Claude Code to switch on an experimental Code Mode, which it can refuse for several reasons
Unclear It is not stated whether setting the variable is enough on its own, since a server-side check also decides who is eligible.
An experimental REPL mode lets Claude write JavaScript that calls its tools, with an opt-in dialog and limits on resuming and moving sessions
Unclear It is not clear how the mode is switched on or which environment variable controls it.
A new /project command, gated by tengu_gentle_cupcake, sets attributionProject for usage reports; IDs may start with chan_ and --settings is a source
Unclear It is unclear whether the /project command is actually available to type in this release.
CLAUDE_CODE_MEMORY_PUSH_DELETE_MODE now accepts recorded, and Claude is told how to delete a shared-memory file with a plain rm command
Unclear Which mode is used when the variable is not set is not stated.
Resumed remote sessions can send a short background query to warm the prompt cache, controlled by CLAUDE_CODE_RESUME_CACHE_WARM or remote config
Unclear What setting CLAUDE_CODE_RESUME_CACHE_WARM does, and what happens when it is not set, is not known.
Two new environment variables, one controlling whether a claimed spare process keeps its boot session id, the other about stopping held messages
Unclear What either environment variable changes when it is set is not known.
A new classifierMayAllow flag lets auto mode's classifier settle some permission asks, such as uploads, that used to always need a person
Unclear It is not clear what decides when the safety check may approve asks meant for a person.
A largely hidden REPL mode, where Claude runs scripts that call tools, gains reporting, display, resume and compaction support across Claude Code
Unclear It is not clear whether REPL mode is something you can switch on yourself or is controlled some other way.
New, not yet connected code lets Remote Control find a replacement Claude Code program if the one it started from was removed by an update
claude mcp add --web can add a claude.ai connector from the terminal#A hidden --web option on claude mcp add adds an https server as a claude.ai connector, but it is off unless switched on remotely
claude.ai connector configs gain an optional urlReportable field, passed through when the server reports url_reportable as true
The plugin model.fork call now takes an optional effort level and reports which one was used, and hook matchers can match full MCP tool names
In REPL mode, plugins that check tool calls can block REPL-made calls, queued scripts are judged by auto mode, and long output is cut to a size limit
Unclear It is not clear when Claude Code runs in REPL mode, so whether these changes reach you is unknown.
New code-mode pieces: a setup opt-in step, warning banners, MCP tools reached inside the REPL, and file reads refused in code mode
Unclear It does not say what decides whether the code mode opt-in step appears.
Claude Code now scores each prompt 0 to 2 for security-audit intent and stores it in _securityCheckRequestScores; hook docs mention inner REPL calls
Unclear It is not known where the stored scores are used, or whether they are sent anywhere.
Claude Code adds CLAUDE_CODE_STREAM_TIMER_SLACK_MS, CLAUDE_CODE_MAIN_REPLY_HOLD_MS and CLAUDE_CODE_MCP_SERVE_TUNNEL_BINARY_FRAMES
Unclear The condition under which CLAUDE_CODE_MAIN_REPLY_HOLD_MS takes effect is not stated.
Setting CLAUDE_CODE_ULTRAREVIEW_LAUNCH_RECORD to a file path makes a successful ultrareview launch write its session id to that file
A setting lets a message sent with Enter join the turn already running, shown with the hint 'enter joins the running turn'
Unclear It is not clear whether this feature is available outside Anthropic's internal builds or what switches it on.
mcp_tool hooks can run on another machine, and hook conditions work more widely#An mcp_tool hook can now make its call on the machine that holds the MCP server, and hook "if" checks no longer require project settings
Unclear It is not clear when an mcp_tool hook runs on the other machine rather than locally.
Before --remote creates a cloud session, Claude Code can check whether moving your local session is allowed and stop if it is not
Unclear It is not clear what causes this check to run, so it is unknown which --remote launches it applies to.
Several internal changes prepare for a REPL mode, including a filter on REPL mode messages and a new usage data field
Claude Code has unfinished code for a REPL script mode that refuses Claude in Chrome calls and skips auto mode checks on calls inside an approved script
Unclear Whether this REPL mode can be reached at all in released builds.
Claude Code adds settings for reporting where tools come from, and now reads two remote switches through a stored-state check
Unclear It is not clear whether tool-provenance information is sent in any session a reader runs.
Claude Code gains more parts of an opt-in mode for running code, including script tracking and an idle-wait tool, all behind a feature check
Unclear It is not clear whether this mode can be switched on by anyone yet, or how.
A new set_attached_project command sets or clears CLAUDE_PROJECT_UUID on a live session, if the session gets its Project from its host
CLAUDE_CODE_PROGRAM_USE_CASE_STAMPED variable for programs that launch Claude Code#Claude Code now reads CLAUDE_CODE_PROGRAM_USE_CASE_STAMPED, which a launching program can set; the value follow picks a separate way of working out the use case
Unclear It is not stated what the program use case controls, so the effect of setting the variable is unknown.
Setting CLAUDE_CODE_DISABLE_AGENT_PROGRESS_SUMMARIES stops Claude Code from writing progress summaries for agents running in the background
Claude Code has text for an experimental mode turned on by an environment variable, listing the many reasons it can be refused
Unclear The mode's name and the environment variable that turns it on are not shown.
A new check can skip compaction based on token estimate, model and context window, but it stays off unless switched on remotely
Unclear What happens when the check decides to skip is not clear.
In some multi-machine sessions, Claude's instructions now say the machines tool can be called inside the REPL
Unclear When these instructions apply is not known.
Claude Code lists two new kinds of background model request: one for hook conditions and one for permission requests from an attached computer
Unclear It is not clear whether either is in use yet.
After first-run setup, a code-mode opt-in step can restart Claude Code, and MCP setup now records how long loading took
Unclear What code mode is and when the opt-in step appears are not known.
When matching permission rules, Claude Code now also checks a second list of rules after the first
Unclear It is not clear which rules the second list holds, so the effect on your permissions is unknown.
If adding or removing an MCP server can't save the config, the error now blames a lock and suggests CLAUDE_CODE_CONFIG_LOCK_WAIT_MS=60000
A new SDK field gives the date and time your claude.ai login ends for good, possibly up to 3 days earlier than shown
Unclear Which SDK message or result carries this field is not stated.
Result messages in the SDK have an optional credential_expires_at field
Claude Code can list files changed in a launch folder that isn't a git repository, using a temporary repository; needs git 2.32+
onProcessSpawned callback now reports how long the start took#The SDK now passes spawnSyncMs, the time taken to start the Claude Code process, as a second argument to onProcessSpawned
What would fix it?
Smaller changes and internals, grouped as the pipeline found them. Nothing is dropped, it is only further down.
9 more of these are in What probably matters to you, above.
security-review now ends with a markdown-only report and may add a closing tip, and a spinner tip can suggest the Claude Security plugin
mcp_tool hooks now wait for servers, report clear failures, can block actions, and are handled more strictly when they come from project settings
Unclear It is not clear under what condition hooks from project settings are skipped when user settings exist.
Claude Code's check that asks before an MCP tool sends session files elsewhere moved into new logic, with a separate check for unasked turns
Unclear It is not clear whether Claude Code still asks before using MCP tools that can send this session's files elsewhere.
The off-by-default request-reuse mode now keeps a saved request for each subagent and can place a block boundary near the end of the tools list
Connected apps can now draw backdrops on mobile as well as desktop, backdrops can carry a color scheme, and Input and Select join the known draws
Unclear It is not clear whether anything switches the backdrop feature itself on or off.
The ui.backdrop theme argument now accepts colorScheme light or dark, alongside new upload tools and claude.ai connector URL helpers
A new connectorCardAllows permission state lets connector tools allowed from a card run for the session unless destructive, with a new MCP ask-rule step
Unclear It is not clear whether this is switched on for users or where a connector card appears.
Behind a new remote switch, non-interactive sessions trim Project CLAUDE.md files only when all instructions together exceed a budget
/send-to-project can turn Remote Control on, opens a compose dialog when no message was typed, and explains more clearly why nothing was sent
Unclear Whether /send-to-project is available to users in this release is not clear.
In acceptEdits mode, Claude Code can lift policy rules that would ask before using an MCP tool, unless CLAUDE_CODE_EVAL_CONFINED is set
Unclear It is not clear which MCP ask rules are lifted, or whether this is active yet.
When a project file is too large to fetch directly, Claude Code can now download it as a local file with a notice, if a server switch allows it
With the stream-json output format, Claude Code now has a way to write display rows to its output stream
WebFetch can return a page as a citable search result for subagents, but only with CLAUDE_CODE_WEBSEARCH_CITATIONS set and a remote switch that defaults off
Permission prompts gain an outcome that allows a connector tool for the rest of the current turn once a person approves it
Unclear It does not say how this option is offered or whether it is switched on for everyone.
In auto mode, Claude Code can now refuse REPL scripts that a hook rewrote, that contain \uXXXX escapes, or that start or message other agents
Unclear These refusals only apply when a further condition holds, and what that condition is, and whether it is met by default, is not stated.
Claude Code has code to resend a request once after it is rejected for hitting a five-hour or seven-day usage limit, behind a remote switch
Unclear It is not stated how the remote switch controls the retry, so whether the retry happens for anyone is unclear.
Shared memory sync now syncs deletes of named files at once, restores files it could not delete, and tells Claude when a delete is held
Remote Control now adds REPL mode's file and shell tools to what the Claude apps can show, so a script's tool calls appear there
The SDK's decoded body field on peer messages is now also present when Claude Desktop declares it and the text matches
Tool search can leave out matched tools that are already loaded or too large for the remaining space, and tell Claude why
Reading an artifact's database into files is now refused when the rows may have been written by other people; leaving out out_dir still works
In REPL mode the Background tasks dialog gains a "scripts" group, and clicking a tool block in the conversation can show a script's source
Unclear It is not clear whether REPL mode, and so this section, is available to readers by default.
The spinner shown while Claude works can now display a waiting message when no other message replaces it
Claude Code can now display calls to a script tool, labelled script: with a line count, and hides tool errors that repeat a script's error
Unclear It is not stated what switches the script tool on, so it is unclear whether any reader can use it.
A message from a declared peer session that includes a disallowed_tools list is now refused instead of accepted
A check on public changes now asks for permission when an item came from another organisation's invite or reaches outside the audience
Unclear Which actions in Claude Code this check applies to is not clear.
primitiveToolGrant can now remove a tool from the session#If primitiveToolGrant is set and leaves out a certain tool, Claude Code no longer treats a related tool as available
Unclear It is not clear which tools this check covers.
Plugin interface drawing now tracks mobile separately, and plugin backdrops are drawn on surfaces other than desktop
Unclear Whether any mobile app shows plugin-drawn interface yet is not clear.
Resumed conversations now restore REPL mode, and one bridge-session record is no longer written to the current transcript at exit
Unclear It is not clear what REPL mode is or what restoring it changes for someone resuming a conversation.
/clear formally ends the conversation, and commands can run while a question waits#/clear now ends the main conversation, some commands can run while a question is open at idle, and SDK-sent messages carry an empty blocked-tool list
Unclear It is not clear what the built-in check tests, so it is unknown who gets the /clear and open-question changes.
The commit, push and pull request command now first looks for a replacement prompt command and runs it instead of its own
Unclear It is not clear where the replacement command comes from or whether users can provide one.
classic.SessionStart hook on pre-started sessions#If a host started the session early, some mod calls made from a classic.SessionStart hook are now refused instead of hanging
Unclear It is not clear which mod calls this applies to.
Resuming a subagent can now wake it for queued results, failing if none are queued, and it keeps more detail from earlier runs
Unclear It is not clear what starts this wake-up path or when it applies.
When two tools from one MCP server get the same name in Claude Code, allow rules are refused for them with a warning, and connector servers are labelled
Remote session spawns now pass a spawn_reason, and the self-hosted runner's spawn-hook log lines include it
Re-login to the same organisation can keep the stored login, and sign-in now tells you when a previous account or API key may still be stored
The usage scan now skips sessions from other accounts or organizations, and usage guides must say how many sessions were left out
Unclear It is not clear which command produces this guide or whether it is available to all users.
After an API error that stopped on a stop sequence, Claude Code now treats the turn as waiting for a tool, so Claude keeps going
When a check is on, agent status events include the id of the enclosing tool call, and that id is passed to the work it starts
A tool can now start the two-minute move to the background on its own, without CLAUDE_AUTO_BACKGROUND_TASKS being set
Unclear Which tools start the timer themselves.
The "Run a dynamic workflow?" dialog now cycles between summary, raw script and arguments, and trims long scripts
A redesigned /powerup layout hides the title row and shows key hints for selecting, opening, closing and marking lessons done
Unclear It is not clear what switches the redesigned layout on, or whether anyone sees it yet.
A file mentioned with @ that is too large to attach now becomes a reference, and Claude is told to read all of it in portions
Claude Code now warns when your sign-in is about to expire or has expired, and points you to claude auth login or claude setup-token
Unclear It is not clear when, or whether, Claude Code shows this warning in practice.
When a config file is not valid JSON, Claude Code leaves it untouched, prints advice, and refuses non-terminal starts with config_file_invalid
Bug reports filed to GitHub now add a line saying a certain mode is on, when it is active
Unclear The name of the mode is not known.
Remote and cloud sessions now explain when a chat's claude.ai Project changed or was left, or when trusted devices can't be confirmed
When you interrupt, the list of still-waiting messages now includes ones held back behind a model change or start-up work
For models on a built-in list, agent progress summaries now always use the short form instead of the configured length
Claude Code adds refresh reasons for changing or leaving a project, notices for MCP sign-in and unsent media, and marks paused remote sessions
When attachments are withdrawn or deleted in a Remote Control session, files, PDFs, @-mention references and Read results are now replaced too
When you branch a conversation with /branch, the new copy now keeps the original session's REPL mode instead of losing it
The /mcp list of cloud MCP servers shows a second count per server, and failed forked slash commands now return their result text
Unclear What the extra per-server number in the /mcp list counts is not stated.
Claude Code can take image size limits from the model's own limits, and an ID check now accepts a chan_ prefix as well as claude_proj_
In headless use, Claude Code can now send an empty success result with zero turns and the session's cost totals while no turn is in progress
Unclear It is not clear in which situations Claude Code sends this empty result.
In some cases a sandboxed command's request to reach a website is now checked automatically with a time limit, and denied if no answer comes
Unclear It is not clear which condition triggers the time-limited check or whether it applies to ordinary sessions.
In REPL mode, pressing stop while code is running now interrupts it and ends the turn after a short wait instead of straight away
Unclear It is not clear whether REPL mode is switched on by default or how a reader turns it on.
git commit -F -#Under a server switch, Claude's commit instructions use git commit -F - with the message fed in, instead of the -m form
When pulling shared team memory, Claude Code now skips writing held-back files and replays recorded removals before writing a conflict notice
Published verbatim by Anthropic for v2.1.318. Text is unmodified from the upstream changelog. Everything else on this page came out of the bundle instead, which is why the two lists don't match.
Of these 84 bullets, 26 name something an entry on this page also names, 22 name something no entry here does, and 36 name nothing specific enough to line up either way. The pairings are made on names both sides wrote down, a flag or a setting or a slash command, so read one as probably the same thing rather than as a fact, and read the middle number as candidates rather than as a miss count.
assume_role on mantle upstreams of the Claude apps gateway, for a role in another AWS account and for per-developer AWS cost attribution
Probably bedrock-style-assume-rolebearer-token-check-extended-to-ma, mantlebedrock-gateway-attribution-warning-reworded-and-made effort option to a mod's $.model.fork, so a quick side job can run at low whatever effort the session uses; the result reports the effort sent
Probably hook-matchers-mcp-tool-name-matching-and-modelfork-effort claude auth status (loginExpiresAt) and to claude -p JSON results (credential_expires_at)
Probably stop-message-uuid-description-extended-to-claude-p, result-messages-gain-credential-expires-at claude -p text output, starting 10 days before a stored claude.ai sign-in ends
Probably stop-message-uuid-description-extended-to-claude-p --continue sometimes losing the earlier conversation after a malformed tool call's retries ran out (regression in 2.1.314)
No entry names this #96628[BUG] After a tool call fails to parse, the retry notice can point `parentUuid` at a row that is never written, so `--resume` loses everything before it Open
/logout, claude auth logout and claude mcp remove reporting success, and changes to trust, consent or MCP settings being dropped with no word, when the config file could not be updated on a shared network home directory
No entry names this Not logged in when another terminal or host signs in again to the same account and organization on the same home
Probably auth-status-output-gains-an-extra-line claude mcp add and remove writing the config without its lock: they now wait for it or exit 1, though another job's save can still replace a saved server
Probably claude-mcp-add-web-add-a-claudeai-connector-from-the-cli claude plugin commands exiting before a slow token refresh was saved, which could sign you out on the next run
No entry names this #102265[BUG] Short-lived subcommand (claude plugin validate) exits mid OAuth refresh: the rotated refresh token is never saved, the next process gets invalid_grant and blanks the login Open
claude plugin validate passing a custom-named marketplace file that claude plugin marketplace add rejects
No entry names this allowedProviders list that another login on the same machine had stored
No entry names this claude -p and SDK sessions
Probably stop-message-uuid-description-extended-to-claude-p claude -p and Agent SDK sessions reporting success when an API error cut off a reply while a mod hooked turn.step
Probably stop-message-uuid-description-extended-to-claude-p claude remote-control hosts, the background daemon and self-hosted runners failing to start new sessions (spawn ENOENT) after an update or version cleanup removed the binary of a still-running process
No entry names this sudo claude) being treated as dead by version cleanup, which could delete its binary
No entry names this /remote-control showing a session as available after a failed connection
Probably remote-control-failure-messages-tell-the-user-how-to-recover, remote-control-callout-flow-refactored-remote-control-dial /design revoke reporting success when the request failed with a 404
No entry names this Setup hook writes to CLAUDE_ENV_FILE not reaching the Bash tool in some cloud sessions
No entry names this NODE_EXTRA_CA_CERTS file is empty, not PEM, or starts with a damaged certificate: readable certificates are now used, with a warning on stderr
Probably much-longer-node-extra-ca-certs-diagnostics-two-new-cases, node-extra-ca-certs-partial-load-of-unparseable-ca-files, ca-cert-drop-message-expanded /mcp in cloud sessions counting tools the model can't use; a server's row now shows how many of its tools are available
Probably mcp-listing-and-ultrareview-credits-refactor-forked-slash turn.step hook making a stopped reply finish as a success in headless runs
No entry names this claude -p with no ~/.claude.json backup: it now takes a fresh backup before deleting the last good one
Probably config-lock-only-saves-and-lock-wait-env-var, stop-message-uuid-description-extended-to-claude-p, config-file-backup-and-empty-file-recovery setup-token being refused at start on machines with forceLoginOrgUUID, when the token belongs to a permitted organization
Probably warning-when-the-oauth-sign-in-is-near-its-end, expired-login-messages-login-reached-its-time-limit #106324[BUG] forceLoginOrgUUID is incompatible with setup-token OAuth tokens Open
~/.claude.json that isn't valid JSON with defaults; the start now fails, says how to restore the file, and leaves it as is (anthropics/claude-code#87551)
Probably config-lock-only-saves-and-lock-wait-env-var, config-file-backup-and-empty-file-recovery ~/.claude.json with no backup being reported as corrupted; Claude Code now starts as it does when the file is missing
Probably config-lock-only-saves-and-lock-wait-env-var, config-file-backup-and-empty-file-recovery ~/.claude.json naming an empty or cut-off backup, or failing on a read-only file
Probably config-lock-only-saves-and-lock-wait-env-var, config-file-backup-and-empty-file-recovery claude -p names claude auth login and claude setup-token
Probably warning-when-the-oauth-sign-in-is-near-its-end, stop-message-uuid-description-extended-to-claude-p, expired-login-messages-login-reached-its-time-limit ~/.claude.json, settings files and the stored login survive saves on shared network home directories: they are less often deleted, briefly emptied or replaced by a near-empty or older copy
Probably config-lock-only-saves-and-lock-wait-env-var, config-file-backup-and-empty-file-recovery /bin/bash
Probably bash-heredoc-on-stdin-teaching-macos-binbash-only-flag-d claude -p, SDK apps) to stop rewriting ~/.claude.json on every run when nothing in it changed
Probably config-lock-only-saves-and-lock-wait-env-var, stop-message-uuid-description-extended-to-claude-p, config-file-backup-and-empty-file-recovery /ide from an outside terminal or your editor's extensions view
Probably awayidle-jetbrains-plugin-not-found-message-now-tells-user prompt.submit to stay in the user's turn rather than go out as a system message
No entry names this deniedMcpServers URL entries to also block Anthropic-provided servers whose upstream URL matches, and connectors in cloud sessions and on self-hosted runners, where the session is given the connector's own URL; otherwise deny a connector by name
Probably mcp-deniedmcpservers-matches-mcp-url-parameter-addresses /permissions rule details, built-in rules and delete confirmation views to use the standard dialog; pressing Ctrl+C twice there now goes back or cancels instead of quitting
No entry names this /claude-api managed-agents-onboard to give a new agent's built-in tools, and a quickstart template's MCP tools, the auto permission policy ("Judge risky actions") by default, matching the Claude Console; say "run without asking" to opt out
No entry names this external_account) credentials file is about to expire
No entry names this listen.trusted_proxies: they now say to list only a proxy that adds the connecting address to X-Forwarded-For itself
Probably gateway-guidance-on-x-forwarded-for-and-trusted-proxies-rewo CLAUDE_RUNNER_SPAWN_REASON to the orchestrator's spawn-runner hook environment: why Anthropic sent the request (e.g. runner_lost, resumed); unset when none was sent
Probably cloud-sessions-and-spawn-reason-in-the-runner --release-idle-session-min is the last argument with no value (still ignored) or has a value of only spaces (still read as 0)
Probably self-hosted-runner-server-sent-idle-session-release, self-hosted-runner-idle-limit-from-anthropics-control-plan ok session=… now prints at every log level on stdout (was debug only, stderr) with reason=… after jti=…, and nack session=… ends its fields with reason=… after retryable=…; update scripts that match retryable=true: or retryable=false:
No entry names this gh run view --log and gh api log requests now get the log in most environments, not an unreachable redirect
No entry names this A model matched these bullets to the GitHub issues they fix, so a link can be wrong.
No change to the system prompt since v2.1.317.
Claude Code, interactive mode
No prompt capture for v2.1.318, so this release's prompt surface is unknown.
482 documentation changes were recorded within 24 hours either side of this release, nearest first. The closest 12 are below. They're here because they happened near this release in time. That's not a claim that this release caused the edit, or that the page documents anything in it.
The 62 literal strings found in the bundle, with the number of entries that name each one. Picking one searches for it. A name is here because this build's code mentions it, which is not the same as it working or being finished.
What's wrong with this entry?