Claude Code 2.1.304
76changes- Add /auto-mode-setup, a guided setup that teaches auto mode about your project and environment.
- Let Code Review owners choose Quick, Standard, or Deep review levels for the organization or each repository.
- Stop model switches, messages, and interrupts from waiting behind a context usage check in SDK and desktop sessions.
- Fix claude --worktree removing a worktree and its branch without asking when HEAD was detached, which could lose commits.
- Fix the rm -r safety prompt missing for mounted drives and top-level mount folders such as /Volumes and /media.
New features
10Add /auto-mode-setup guided setup
Added /auto-mode-setup: a guided setup that reads your project and recent sessions to teach auto mode about your environment, now also available on Bedrock, Vertex, Foundry, Claude apps gateway and telemetry-disabled sessions
▤ In the docs[Code Review] Add Quick, Standard, Deep levels
Added review levels to Code Review: Owners choose Quick, Standard (the default), or Deep for the organization, and can give individual repositories their own level
▤ In the docs[Code Review] Add org-wide repository coverage
Added an option in Code Review admin settings to review every repository in a GitHub organization, including ones added later, without selecting them one by one
▤ In the docsAdd /claude-api managed-agents-secure check
Added /claude-api managed-agents-secure, which compares a Managed Agents agent's settings with Anthropic's published security guide, without changing anything
More changes (6) Fewer changes
Show when a plugin skips a mod's hook
Added a transcript line when a plugin that runs first skips one of your mod's hooks, naming the event and the plugin
Add person-only decider for tool.check
Added decider: "person" to a mod's tool.check answer: only the permission dialog or a host's callback can approve the ask, and it is refused with the mod's reason where nobody can be asked
Add e.isReadOnly to tool.check event
Added e.isReadOnly to a mod's tool.check event, set when the tool call is read-only
Show prior tool names in compaction summary
Added the names of the tools called earlier to the compaction summary in the transcript view (Ctrl+O)
Warn before resuming near usage limit
Added a warning on Pro and Max plans when resuming an inactive conversation would use a noticeable share of your 5-hour usage limit: resume anyway, or start a new conversation and Claude can look up the previous one if you refer to it
[VSCode] Add mod drawing on more platforms
Added mod drawing in the chat panel for sessions on Amazon Bedrock, Google Vertex AI and Microsoft Foundry, behind an LLM gateway, or with telemetry turned off
Improvements
21Speed up resuming large compacted sessions
Improved resume of large compacted sessions in SDK and headless mode: the first request after resuming is sent sooner
Enable tool search by default behind gateways
Changed tool search to be on by default behind a custom ANTHROPIC_BASE_URL; a 400 or 422 from the gateway is retried without it
More changes (19) Fewer changes
Show when the usage rate limit ends
Improved /usage to say when a rate limit on usage data ends instead of "try again in a moment"
Improve session mod load status lines
Improved the lines shown when a mod written in the session loads, reloads or fails to load
List each artifact view's link
Improved artifacts with several views: when you ask for links to them, Claude now writes out each view's link and explains that the address bar stays the same as you switch views
Speed up cloud session first turns
Improved cloud sessions: a new conversation's first turn no longer waits for the local (stdio) MCP servers of plugins synced from claude.ai, unless a server sets alwaysLoad: true
▤ In the docsImprove remote session status recovery
Improved how a remote session recovers its status when the server refuses a status update that carried other session data
Shrink auto mode session files
Improved session file size in auto mode: the safety classifier's context is no longer stored with every request
Improve claude -p invalid-settings warnings
Improved the claude -p invalid-settings warning: a file that did not load lists only what stopped it, and long lists are capped per file
Improve throughput with many subagents
Improved responsiveness and throughput when many subagents are running at the same time
Reduce memory use sending long conversations
Improved memory use when sending requests in long or resumed conversations
Name the repo in self-hosted runner errors
Self-hosted runner: Improved the error shown when the git host answers "not found" for a session's repository: it now names the repository and whom to ask
Speed up git startup on Windows
Windows: Improved git command start-up with a standard Git for Windows install by running its git.exe directly instead of through its cmd launcher
Let mods enforce settings as a ceiling
Changed mods' settings.read hooks to receive each settings layer and rewrite the merged settings, so a managed mod can make an organization's setting a ceiling, including the cleanupPeriodDays the retention sweep uses
Add numbered choices to artifact delete
Changed the /artifacts delete confirmation to show numbered choices: 1 or Enter keeps the artifact, 2 deletes it, and y, n and Esc still work
Simplify the in-session mod activation prompt
Changed the question that turns on the mods Claude writes in a session to plain text with no link or session id; it stays up when a new turn starts, and Claude tells you what loaded when you turn them on
Resend skill list after compaction
Changed compaction to send Claude the skill list again afterwards; /reload-skills now sends only new or changed skills
Cap skill hooks per source
Changed skill hooks: the skills of one source (a repo, your own folder, or a plugin) can register at most 1,024 hooks and 16,384 matcher characters per session
Restrict allowed-tools for blocked skill hooks
Changed skills under a policy that allows hooks from plugins only: a skill used in the main conversation whose hooks are kept out no longer gets its allowed-tools
Use ant CLI for Claude on AWS
Updated the /claude-api skill to use the ant CLI on Claude Platform on AWS instead of falling back to the SDK or curl
[VSCode] Collapse subagent steps to one line
Changed background sub-agents to show as one status line under their Agent card instead of streaming every step into the chat; the full transcript stays in /tasks
Bug fixes
43Fix delays from context usage check
Fixed model switches, messages and interrupts waiting behind a context usage check in SDK and desktop sessions
Fix worktree removal without confirmation
Fixed claude --worktree removing the worktree and its branch without asking at exit when HEAD was detached or on another branch, which could lose commits
Fix missing safety prompt for mounted drives
Fixed rm -r of a mounted drive or a folder at its top level (/Volumes, /media, WSL's /mnt/c) not getting the safety prompt that rm -rf /Applications gets
More changes (40) Fewer changes
Fix delayed policy refresh after API key change
Fixed the organization policy not refreshing until the hourly check after approving an ANTHROPIC_API_KEY at startup or switching "Use custom API key" in /config
Fix sandbox glob brackets on Linux
Fixed sandboxed commands failing to start on Linux when a sandbox.filesystem.denyRead glob contains brackets that are not a character class, such as a [user-id] route folder
Fix transcript GC message loss
Fixed CLAUDE_CODE_TRANSCRIPT_LOCAL_GC dropping messages from the saved transcript when the disk filled up while the file was being trimmed
Fix plugins bypassing headless run limits
Fixed a plugin keeping a headless run going past --max-turns, Stop, or ended input by spawning subagents or re-running a finished one
Fix failed mod reload removing tools
Fixed a failed mod reload removing the tools and commands of the version that stays loaded, and its messages being wrong or silent about whether the previous version is still loaded
Fix lost clicks on unchanged mod buttons
Fixed a click on a mod's button being lost when buttons with the same label and no key were redrawn unchanged
Fix mod sharing stuck at secret scan
Fixed mod sharing in the built-in plugin-authoring skill stopping at the secret scan on machines with managed settings and on Team and Enterprise plans
Fix repeated 5-minute waits in scheduled Cowork
Fixed scheduled Cowork cloud sessions waiting five minutes again for each subagent whose WebFetch needed a URL approval that nobody was there to give
Fix outside-directory paths shown as dots
Fixed Write results and rejected edits showing a file outside the working directory as ../../...; they now show ~/... or the full path
Fix Directory plugin skills in Cowork menu
Fixed a claude.ai Directory plugin's skills not running as commands when picked from the "/" menu in Cowork cloud sessions
Fix Remote Control withholding chat history
Fixed Remote Control permanently withholding a desktop conversation's history when a different account was signed in to the terminal
Fix Remote Control start with telemetry disabled
Fixed Remote Control failing to start from Claude Desktop when DISABLE_TELEMETRY or DO_NOT_TRACK is set on a paid plan
Fix context_management errors behind gateways
Fixed context_management: Extra inputs are not permitted errors behind LLM gateways; Claude Code now resends the request without that field
Fix ignored OpenTelemetry settings after sign-in
Fixed OpenTelemetry settings delivered through server-managed settings being ignored for the whole session after signing in during first-run setup or the first start after /logout
Fix missing shell env vars in Code tab
Fixed unsandboxed Bash commands in Claude Desktop Code tab sessions on macOS and Linux not seeing environment variables exported in shell startup files such as ~/.zshrc
Fix skipped Consumer Terms notice
Fixed the Consumer Terms notice being skipped for a day or longer after a server error
Fix inapplicable choices in /privacy-settings
Fixed /privacy-settings offering choices that may not apply to you after a network or server error
Fix context errors after stopping compaction
Fixed a "Context limit reached" or "Prompt is too long" error after stopping a turn while Claude Code was compacting a conversation the API had rejected as too long
Fix subagents forgetting recent work
Fixed subagents sometimes forgetting their most recent work when they were messaged or resumed right after finishing
Fix inaccurate claude --help text
Fixed claude --help text for --tools, --resume, --debug, --json-schema, --disable-slash-commands, --system-prompt-snapshot and claude update to match actual behavior
Fix redundant skill list resends at startup
Fixed resumed and cloud sessions sending Claude the whole skill list again each time plugins reloaded at startup
Fix interrupts not stopping stuck Read calls
Fixed an interrupt not ending a Read that waits on an idle pipe or device named like an image
Fix guessed effort level display
Fixed /effort status, /effort ultracode and the prompt's effort indicator showing a guessed effort level when CLAUDE_CODE_EFFORT_LEVEL is auto or unset and no level is sent
Fix deny reason lost to callback timeout
Fixed a PreToolUse hook's deny reason being replaced by a timeout error when an Agent SDK callback hook on the same tool call timed out
Fix deleted output files on copy failure
Fixed Bash and PowerShell deleting a large command's output file when it could not be copied to the session's tool-results folder, unless the disk is full
Fix truncated notes in narrow terminals
Fixed notes typed into a multiple-choice question with previews being cut off with "..." in terminals under about 100 columns, hiding what was typed (anthropics/claude-code#82768)
Fix /loop wakeups missing the prompt cache
Fixed self-paced /loop wakeups scheduled at the 1-hour maximum firing just after the 1-hour prompt cache expired, so those wakeups no longer re-read the conversation uncached
Fix shutdown approval exiting whole CLI
Fixed agent teams: approving a shutdown request from the main session no longer exits the whole CLI
Fix once hooks removing other skills' hooks
Fixed a skill's once hook also removing another skill's identical hook after it ran
Fix { and } jumping to the wrong row
Fixed transcript mode's { and } landing the prompt on an inconsistent row, or off screen, instead of the transcript's first row
Fix AUTO_COMPACT_WINDOW silently using 100K
Fixed CLAUDE_CODE_AUTO_COMPACT_WINDOW=200k silently setting a 100K window: the variable now accepts k and M suffixes like --autocompact, and warns when a value is out of range or ignored
Fix Ctrl+C clearing uncopied selections
Fixed Ctrl+C in the agents view clearing a text selection without copying it when pressed before the mouse button was released
[VSCode] Fix startup timeout from SessionStart
Fixed one cause of sessions failing to start with "Subprocess initialization did not complete within 60000ms": time a SessionStart command hook spends running no longer counts toward the limit, up to 10 minutes in total
[VSCode] Fix lingering process on early close
Fixed a session closed while it was starting keeping its Claude Code process running until the start ended
[VSCode] Fix mod drawings vanishing in panels
Fixed a mod's drawings disappearing from a long-open panel until the next turn on machines that also run an older Claude Code
[VSCode] Fix hidden sessions with process wrapper
Fixed Past conversations hiding sessions when claudeCode.claudeProcessWrapper is set
[Cloud sessions] Fix unnamed folders in @ list
Fixed folders in the composer's @ list showing no name in cloud sessions, where picking one added only a space instead of a mention
[Claude Tag] Fix silence on shared channels
Fixed Claude going silent in a thread when its channel becomes shared across Enterprise Grid workspaces mid-task; it now posts a one-time note to @-mention it in a new thread
[Claude Tag] Fix silent !fork outside threads
Fixed @Claude !fork typed outside a thread getting no visible reply; Claude now shows you a private note saying to run it from inside a thread
[Code Review] Fix PRs skipped right after enabling
Fixed pull requests opened within a minute or two of an admin turning on Code Review being silently skipped; they are now retried and reviewed