Claude Code v2.1.233 Major Updates - Todo Tool Disabled by Default

This is Ishikawa from the Cloud Business Division. Claude Code v2.1.233 (released 2026-08-14) has been released. Since there is a breaking change involving the default disabling of the Todo tool for certain models, we recommend checking the impact on your environment.

The previous update article is here.

https://dev.classmethod.jp/articles/20260814-cc-updates-v2-1-232/

Update Summary

v2.1.233 includes 20 changes. The breakdown is 4 new features, 6 improvements and developer experience updates, 8 bug fixes, 1 security fix, and 1 breaking change. Notable bug fixes specific to Windows and Linux stand out, and it also includes regression fixes and reverts of changes introduced in 2.1.232.

Notable Updates

GitLab merge request URLs are now supported

The --worktree flag and claude agents view now support GitLab merge request URLs. In the claude agents view, merge requests are displayed in the !N format.

For organizations using GitLab, workflows where a worktree is checked out per merge request for parallel work can now be used as-is, and I feel the gap from GitHub-centric workflows has narrowed.

A memory limit can now be set for Bash execution on Linux

An opt-in memory cgroup limit setting CLAUDE_CODE_TOOL_MEMORY_LIMIT has been added for Bash tool execution on Linux. This prevents runaway builds from halting sessions.

For those running Claude Code on CI, self-hosted runners, or shared development servers, I feel this is a setting worth being aware of.

The bug causing CPU to be pegged during idle on Linux has been resolved

A bug where idle sessions with sandboxing enabled on Linux would continuously use 100% of one CPU core has been fixed.

For those who keep sessions open with sandboxing enabled on Linux, if you've noticed fans spinning continuously even while idle, I feel it's worth updating.

The bug causing auto mode to repeatedly request approval on Windows has been resolved

A bug (regression from 2.1.232) where auto mode would repeatedly request manual approval for normal Bash commands like cd <dir> && <command> > file on Windows has been fixed.

For those who regularly use auto mode on Windows and felt that approval requests increased after 2.1.232, I believe this fix should resolve that.

The issue allowing Windows path validation to be bypassed has been fixed

An issue where Windows paths written with the NT device prefix \??\ could bypass UNC path validation has been fixed, closing a pathway that could lead to NTLM credential leakage.

For those who have Claude Code handle externally obtained repositories or files on Windows, I feel this is something to prioritize when considering an update.

Update Details

New Features

• The --worktree flag and claude agents view now support GitLab merge request URLs (MRs are displayed in !N format in the claude agents view)
• An opt-in setting forward_user_identity has been added to the apps gateway for Anthropic upstream, which sends the signed-in user's identity as a header. Proxies under the gateway can allocate usage on a per-user basis
• An opt-in memory cgroup limit setting CLAUDE_CODE_TOOL_MEMORY_LIMIT has been added for Bash tool execution on Linux
• An environment variable CLAUDE_CODE_WEBFETCH_CACHE_TTL_MS has been added to configure the session URL cache TTL for WebFetch (default value remains 15 minutes)

Improvements

• claude self-hosted-runner session startup is now faster. Session branches are created without rewriting the working tree, and 2 server round trips no longer block agent startup
• Error forwarding for the apps gateway has been improved, and 400/413 errors returned by Vertex, Foundry, and Claude Platform on AWS upstreams now retain the upstream's own messages. An auto-compact bug through the apps gateway has also been fixed
• claude plugin validate now inspects plain .claude/skills directories as well and reports SKILL.md files that fail frontmatter parsing
• Screen reader mode has been improved: the /effort selector is now displayed as a numbered list with a number input prompt, and hint and dialog text no longer gets cut off
• Print mode diagnostics have been improved: when a request is sent with a model ID that Claude Code does not recognize, a [claude-code:unrecognized_model] line is now output to stderr. This can be suppressed by mapping it with modelOverrides
• For repositories where the origin remote is gitlab.com or bitbucket.org, the GitHub app setup prompt is no longer displayed. The enterprise marketplace prompt covers non-GitHub in-house git hosts as well

Bug Fixes

• Fixed CPU pegging during idle on Linux: A bug where idle sessions with sandboxing enabled would continuously use 100% of one CPU core has been fixed
• Fixed unnecessary approval requests in auto mode on Windows: A bug (regression from 2.1.232) where auto mode would repeatedly request manual approval for normal Bash commands like cd <dir> && <command> > file has been fixed
• Fixed repeated stream reopening in MCP v2: A bug where the subscriptions/listen stream would be endlessly reopened against servers that disconnect long-held streams with a fixed timeout (such as serverless hosts) has been fixed
• Fixed incorrect "lost" status for cloud sessions: A bug where cloud sessions could be incorrectly marked as lost when the environment shuts down while waiting for a permission confirmation prompt has been fixed
• Fixed Unknown command for bundled skill aliases: A bug where aliases like /checkup and /review would result in "Unknown command" in -p mode or when loading plugins/MCP, when user or project skills were shadowing bundled skills, has been fixed
• Fixed argument expansion in skills/commands: Argument values are no longer re-expanded as template markers
• Partially reverted 2.1.232 changes around Bash permissions: Changes related to Windows Cygwin-format symbolic links and input redirection (< file) have been reverted. A more narrowly scoped version is planned for reintroduction in a later release
• A subtly welcome fix: A bug where Notification hooks would not fire on permission confirmation prompts when running on Claude Desktop or VS Code has been fixed. For those using hooks to send notifications at work breakpoints, I feel this is a fix that will quietly make a difference

Security

• An issue where Windows paths written with the NT device prefix \??\ could bypass UNC path validation has been fixed, closing a pathway that could lead to NTLM credential leakage

Breaking Changes / Deprecations

• Todo and task tracking tools (TaskCreate/Get/Update/List, TodoWrite) are no longer available on Opus 4.8, Sonnet 5, Fable 5, Mythos 5, and later models. Setting CLAUDE_CODE_ENABLE_TODO_TOOLS=1 will re-enable them.

  # Not available by default for affected models. Set the environment variable to use as before
  CLAUDE_CODE_ENABLE_TODO_TOOLS=1 claude
  

  1. Claude will not create task lists

     Since TaskCreate / TaskGet / TaskList / TaskUpdate cannot be called, no progress checklist is presented even for lengthy tasks. Instead, the approach becomes explaining the plan in prose and proceeding. For cases where you want to track "how far along we are" in a list during multi-step work, the experience will change.

  2. Reducing round trips and cost

     In actual measurements, the same task went from 7 turns → 5 turns, output 2,286 → 1,373 tokens, $0.405 → $0.323 (approximately 25% reduction). This is because 10 tool calls spent on task registration and updates are eliminated.

  3. Sub-agent related functionality is not affected

     Task (sub-agent launch), TaskOutput, and TaskStop remain available by default. Despite the similar names, these are different things, and /workflows and background agent operations remain unchanged.

In Closing

v2.1.233 has one breaking change, but overall I feel it is a release centered on bug fixes and infrastructure improvements. If you don't have workflows that depend on the Todo tools, it should be straightforward to upgrade.

I feel that those using Claude Code on Windows or Linux, and those doing team development with GitLab, will particularly want to review the changes. Why not update and give it a try?

References

https://github.com/anthropics/claude-code/blob/main/CHANGELOG.md

https://code.claude.com/docs/en/changelog

https://dev.classmethod.jp/articles/20260814-cc-updates-v2-1-232/