Claude Code v2.1.318

Remote sessions get new timing options for sending held output#

Remote sessions gain two settings, refreshed from the server, for when streamed output is sent and how long held output waits

New plumbing for remote and attached-project sessions#

Claude Code gains several internal additions for remote and attached-project sessions, including a set_attached_project control message

Unclear It is not clear whether any of these are used yet or how set_attached_project is reached.

Claude Code can send a handoff message into a remote session#

Claude Code can now send a handoff message into a remote session without asking Claude to reply

Unclear It is not clear whether anything in Claude Code sends this message yet.

Remote setting changes are stripped unless only the session owner can write#

Settings sent through apply_flag_settings are now removed unless the session service confirms only the owner's account can write to the session

Remote session workers can mark a request so its permission errors are counted apart#

Remote session workers can flag a request so a refused-permission error on it no longer adds to the count of repeated errors

Keep-alive can now tell when a session is only waiting on a permission prompt#

The keep-alive check now learns whether a session is only waiting for you to answer a permission card

Remote sessions report whether they already have an uploaded media file#

In remote sessions, image, audio and video uploads can now say whether the session already has a matching file, except in self_hosted setups

Turns that end with no reply can be recorded as an empty message#

Behind a server switch, a main-conversation turn where Claude says nothing now ends with an empty reply marked as finished

Server can set how long headless remote sessions wait on background shells#

In remote non-interactive sessions waiting on background shell commands, the long wait length can now come from the server instead of being fixed

Permission denials can mention a session limit in remote sessions#

The message shown when a permission is denied can now mention a session limit, and waiting requests record their folder

Unclear The wording a reader sees when a denial mentions the session limit is not known.

Some remote workers now ignore lines sent on standard input#

On some remote workers, a line arriving on standard input is dropped with a warning, because only the session stream may deliver input

Unclear It does not say what makes a worker one where only the session stream may deliver input.

Two new internal labels for remote sessions#

Claude Code adds two internal labels for remote sessions, one about handing over while draining and one about answers from subagent cards

Session resume records more details about how transcripts are read#

When loading a saved session, Claude Code now checks the key order of each line and records many more counts about the scan

Resuming a session judges interrupted turns differently when extra context follows#

On resume, Claude Code can look past trailing context rows when deciding whether your last turn was interrupted

Unclear How these cases change what resume offers to continue is not shown.

Remote sessions record agents that a person stopped#

In remote sessions, Claude Code now records agents that were no longer tracked because a person stopped them

Usage reporting records more about plugins, Remote Control, memory and permissions#

Claude Code's usage reporting gains fields about plugin steering, Remote Control devices, unanswered memory context and where permissions were asked

New internal reporting step that can be switched off remotely#

Claude Code has a new internal reporting step that is skipped when a remote switch is on and another condition holds

Usage data now records web address and directory details for claude.ai connectors#

Claude Code now records the web address and directory entry of claude.ai connector tools it uses, checking the address is https

Permission prompt usage data now records the folder it was asked in#

Usage data sent for permission prompts now includes the folder the prompt was asked in, and two new field names were added

Usage data on background shell commands stopped at their time limit on hosted workers#

Claude Code now reports when a background shell command is stopped at its time limit, restarted, and what work follows

Usage reports now record REPL mode and a few new events#

Usage reports can now include whether REPL mode is on, and three new event names are allowed through

Usage data can be sent before waiting on a sign-in refresh#

Under some condition, Claude Code now sends pending usage data first and waits for its sign-in refresh afterwards

More detail in usage data when cloud subagent transcripts fail to load#

Usage data about loading cloud subagent transcripts now records their size, the event count, and which limit was exceeded

Usage data now notes zsh commands that start with an equals sign#

Claude Code's usage data about shell commands now notes when a zsh command begins with '=', which is reported as an error

More diagnostic detail recorded when resuming sessions#

Claude Code records about 15 more measurements when it scans a transcript to resume a session, and more detail when it clears idle tool uses

Failed reads of shell command CPU time are logged and reported differently#

When Claude Code cannot read a shell command's CPU time from /proc, it now always logs it and reports unexpected errors

Usage report for asking an admin about rate limits no longer records a mode#

Choosing to ask an admin from the rate-limit menu no longer sends a mode field in its usage report; the menu itself is unchanged

Forked agents can be given a fixed effort level and close their own REPL sessions#

Forked agents can now be started with a pinned effort level, and their REPL session is closed when they finish if a check allows it

Unclear It is unclear when the REPL clean-up runs, so whether subagents actually use their own REPL sessions in this release is not settled.

Messages handed back to other agents carry extra markers#

When one agent hands work back to another, the message now begins with markers carried over from earlier messages

Background subagent tool calls now use a fixed 180000 instead of a server value#

The number applied to background subagent tool calls is now fixed at 180000 rather than set by the server, and is 0 in more cases

Unclear It is not clear what the 180000 controls for background subagent calls.

Subagent time limits now depend on the message that started them#

The time limit for a subagent now takes the triggering message into account, and REPL mode is now kept per session

Agent task notices warn Claude about text relayed from Remote Control#

Agent task notices can carry a caller note and warn Claude to treat text relayed from a Remote Control session as data

Background tasks and subagents record more about where they started#

Background tasks now record what they were started for, and subagents can record which action launched them

Creating a cloud session can return move and self-hosted start details#

A newly created cloud session can now carry a flag saying it may run on a runner, plus a start notice for self-hosted runners

Unclear It is not clear what a user sees or can do with these new session details.

Cloud worker sessions get a new check on narrowing settings#

Cloud worker sessions add a check on narrowing settings to a list, applied when CLAUDE_CODE_ENTRYPOINT is unset or on an allowed list

Permission requests from cloud sessions can carry conversation and compaction IDs#

Permission requests can now include a conversation ID and a compaction ID, and Claude Code can read such requests coming from a cloud session

Unclear It is not clear whether reading permission requests from cloud sessions is switched on for anyone yet.

Labels added for moving a local session to the cloud#

Claude Code has three new internal labels about moving a local session to the cloud, including by copying it

Unclear It is not clear whether moving a session to the cloud is something you can do yet.

Cloud session creation tells a refusal apart from a failure#

When creating a cloud session, Claude Code now records a refused request separately from one that failed

Remote session transcripts change the order some entries are uploaded in#

Remote sessions may now hold back prompt snapshot entries until pending messages upload, changing transcript upload order

Unclear The holding rule comes from the server, so which sessions are affected is not known.

Workflow results warn Claude about text relayed from Remote Control#

When workflow results contain text marked as coming from a Remote Control session, Claude is told to treat it as data, not instructions

Placed files now report their checksum and size#

When a file is placed, the result now includes its sha256 checksum and its size

Audio capture and clipboard loading now consult a server setting before falling back#

Audio capture and clipboard support check a server setting before loading fallback module files, and may skip them

Unclear What the further check tests is not clear, so it is unknown when audio capture or clipboard support would stop working.

Resource monitor now keeps memory readings with a timestamp#

The part of Claude Code that samples process resources now stores its last memory reading along with when it was taken

File downloads now record the size the server says the file is#

When Claude Code downloads a file, it now notes the size the server reported, so a download that arrives short can be spotted

Some document attachments are no longer handled like images#

A check that used to treat every document attachment like an image now leaves some documents out

Unclear Which documents are now left out, and what that changes for them, is not known.

An unexplained experiment and beta entry are wired up#

Claude Code now checks a remote experiment setting and can send one more beta feature request, but what either one does is not known

Unclear What this experiment and the new beta entry control is not known.

New server-set number controls an unidentified check#

A new check compares a value against a number set from Anthropic's servers, which falls back to 0

Unclear What this check controls is not known.

A remote switch's fallback now depends on how Claude Code was started#

When the server sends nothing, a remote switch no longer always falls back to off and may be on depending on how Claude Code was started

Unclear It is not clear which ways of starting Claude Code get an on fallback, or what the switch controls.

Feature flag results are re-saved less often, and flags can be fetched while signed out#

Claude Code skips re-saving unchanged feature flag results in one more case, and has a new option for fetching flags while signed out

Streaming connections can be told to reconnect on a fresh connection#

A streaming connection can now ask to open a fresh network connection when it reconnects

Unclear It is not clear what marks a connection this way, so it is unknown which sessions get the fresh connection.

Chrome auto-enable and two other switches are unchanged#

The remote switches for Chrome auto-enable, trimming trailing tags from MCP output and one other were only renamed internally

A warning appears when MCP servers are off in a REPL-mode session#

Claude Code can now show a 20-second warning when MCP servers are switched off in a REPL-mode session

Session records can note REPL mode#

When a code mode is set, Claude Code writes a "repl-mode" entry into the session's saved record

Claude can be told to investigate in fewer, fuller REPL scripts#

New guidance tells Claude to do a turn's investigating in as few REPL calls as possible, in certain conditions

Unclear When this guidance is given to Claude is not known.

Claude gets REPL-mode wording when told to clear queued notifications#

In REPL mode, Claude is now told to run a script that collects queued notifications until none remain, instead of calling the tool directly

Ending a conversation from a nested call now ends only the issuing turn#

When the end_conversation tool is called from inside another call, Claude Code can now check which message sent it and end that turn

Unclear It is unclear what switches the new behaviour on, so whether it applies to a given user is not known.

Waiting for MCP servers can describe tools as available inside the REPL#

In one mode, the tool that waits for MCP servers says their tools become callable inside the REPL instead of joining the main tool list

Unclear It is not known what switches this REPL mode on, so readers cannot tell whether they will see it.

Preview tunnel sends raw bytes instead of base64 text#

The preview tunnel now sends content as raw bytes instead of base64-encoded text

Remote-control device report is kept when a session is stopped#

Claude Code now remembers whether a remote-control device report existed when a session was killed, so that state carries over

The preview tunnel now sends data as raw bytes instead of text#

The preview tunnel now accepts requests as text or binary, passes connection data as raw bytes instead of base64 text, and drops malformed requests

Reporting on deleted attachment copies sent over the bridge is more detailed#

Reports on deleted attachment copies over the bridge now name seven failure reasons and use per-type size thresholds

Unclear It is not clear whether these checks only produce reports or can also stop an attachment from reaching a remote session.

Preview tunnels can now use a faster data format#

Preview tunnel connections can now ask the relay for binary data, falling back to JSON when that keeps failing

One kind of relayed message can now carry a tool result#

A kind of relayed message that used to produce nothing can now pass on a tool result when a condition holds, and a text mix-up was fixed

Unclear It is not clear what condition makes this kind of message carry a tool result.

Some shell commands get a second permission check that can overturn an allow#

Certain shell commands are checked twice for permission when a remote switch is on, and a deny from the second check wins

Unclear Which commands match the pattern, and what the second check does differently, is not clear.

Shell commands with process substitution skip special end-of-command handling#

Claude Code no longer gives special end-of-command handling to shell commands that contain <( or >(

Changes to how shell commands are checked before running#

Claude Code's shell command check skips its quick cached path in more cases and decides "too complex" commands with a single shared check

Unclear It is not clear whether the set of commands that trigger a permission prompt actually changes.

The shell setup line that turns off extended globbing now has extra text added#

The setup line Claude Code runs before shell commands to switch off zsh extended globbing now ends with a variable part instead of fixed text

Unclear It is unclear what text is now added to the end of the setup line, so its effect on commands is not known.

Bash commands report more detail after they finish#

After a shell command runs, Claude Code now passes on the kinds of command and shell use, and logs more about it

Notification tool no longer requires a status value#

A notification tool's status field, which had to be "proactive", can now be left out

Queued remote notifications can be limited to your own#

Claude Code can now pass on only queued remote notifications you wrote, and only nudges Claude about your own pending ones

Internal change to how a status or queue display is prepared#

Claude Code changed how an unidentified status or queue display is set up, with no visible effect known

Unclear Which part of the screen this affects is not known.

The theme picker is now marked as not titled by the surrounding app#

The theme selection screen is now drawn with a setting saying its title does not come from the app hosting it

Unclear It is not clear whether the theme picker looks any different as a result.

Prompt previews in list rows come from a ready-made summary#

A list row that previewed prompts no longer builds the first two plus a "+N more" count itself and shows a prepared preview instead

Unclear Which list this is, and whether the preview it shows now looks different, is not known.

The use-case ID attached to requests can now be left off under a new condition#

Claude Code can now skip sending the program_use_case_id value on a request, depending on an extra check made when the request is built

A background request stops labelling its purpose after that label is rejected#

If a side request that labels the turn's use case gets a not-found error, Claude Code stops sending that label for that model for the turn

Turning off kept request bodies now also clears those kept for subagents#

When keeping request bodies for debugging is switched off, Claude Code now also clears the copies kept for subagents

Request reports now record a gzip blocks mode#

A report about request bodies now also records a gzip blocks mode, and stream blocks are now marked when they finish

Artifact contents now reach Claude with a size limit and authorship details#

When Claude reads or publishes an artifact, the content now comes with a size limit and details of who wrote it

Artifact files are now tagged with a mode when read and written#

Claude Code now works out a framing mode for artifact files and applies it when it writes them, cleaning up if a write fails

Unclear It is not clear what the framing value does for the person opening the artifact.

Artifact sharing checks now separate an outside audience from an outside organisation#

The check on publishing an artifact now tracks whether the audience is outside your organisation, alongside the existing outside-org flag

Saving artifact files now cleans up after a failed write#

Fetched artifact files and saved assets are now tracked before and after writing to disk, with cleanup if the write fails

Unclear What the steps run before and after each write record, or whether they change anything a reader can see, is not known.

Claude Code can look up the organization a sign-in token belongs to#

Claude Code can now ask Anthropic's servers which organization a sign-in token belongs to

Claude Code can tell when a login ended because its time ran out#

Claude Code can now check its stored credentials to see whether a login ended because it reached its time limit

Login restore after switching accounts now depends on an extra check#

After you switch accounts or sign out, Claude Code now waits on one more step before reporting that your previous login was restored

Unclear What the extra step checks is not known, so it is unclear which situations now stop counting as a restored login.

Expired sign-in errors now record whether a time limit was reached#

The error shown when your sign-in can no longer be refreshed now also records whether it ran out of time

Sessions in code mode get the full tool list#

When a session is in code mode, Claude Code no longer trims its tool list to the usual allowed set

Unclear It is not stated how a session enters code mode or whether readers' builds can reach it.

The code-mode Poll tool works for sub-agents and handles hooks more strictly#

The Poll tool in code mode can now be called by sub-agents, drops repeated hook messages, and refuses tool calls when a PreToolUse hook hangs

Unclear It is not clear which condition makes this code mode and its Poll tool available.

A step now depends on whether the code mode experiment is on#

A step in Claude Code now runs only when a code mode experiment is switched on, and is skipped otherwise

Unclear It is not clear what turns the code mode experiment on or what this step changes when it does.

JavaScript code sent to tools can be repaired before it runs#

When certain internal checks pass, JavaScript in a tool's code field is repaired before the input is used and checked

Unclear It is not clear what the two checks test, so it is unknown whether any user gets this.

Waking a stopped teammate with a message now records where the message came from#

When a message to a stopped teammate restarts it, Claude Code now records the message's origin before restarting

Failed task claims no longer list which tasks are in the way#

When a task cannot be claimed, Claude Code now gives only the reason and the task, no longer the lists of blocking or busy tasks

Blocked task claims no longer list the blocking tasks#

When claiming a task fails because it is blocked, the result no longer includes the list of tasks blocking it

Clearing old tool results now counts tool calls made after you return from idle#

When old tool results are trimmed from context, the thresholds now add tool calls made since you came back from being idle

Unclear It is not clear what decides whether the after-idle path is used.

Unsent images and documents in tool results can be tracked and replaced#

Claude Code can record images and documents in tool results that have not been sent yet, and swap them for a placeholder

Unclear It is not clear whether this runs for anyone yet, or what triggers the replacement.

Context clearing reports more detail about cleared tool results and token counts#

After old tool results are cleared, Claude Code now records how many tool uses went, the token count, and whether that count was measured

New internal labels for adding an MCP server from the web#

Claude Code's list of start-up labels gains one for adding an MCP server from the web, and another list gains a recorded value

Refused MCP skill resource reads now log a separate reason for unloaded skills#

When Claude Code refuses to read an unlisted MCP skill resource, a resource outside the loaded skills now gets its own log message

MCP server import dialog gets a setting for whether cancelling is active#

The dialog for importing MCP servers now has a setting that controls whether cancelling is active, and the Chrome setup screen was only renamed internally

Unclear It is unclear what, if anything, changes for you when you cancel the MCP import dialog.

Internal checks for resuming interrupted turns and reading the queue#

Claude Code gained internal checks for resuming an interrupted turn and for listing queued commands, with no visible effect described