claudekit / updates / claude-code-2-1-316
[ PATCH · ]

Claude Code 2.1.316

Adds `CLAUDE_CODE_DISABLE_ULTRATHINK`, which turns the `ultrathink` keyword off, `CLAUDE_CODE_SUBAGENT_EFFORT` for the effort level subagents run at, and pooled group budgets on the Claude apps gateway. Fast mode for Opus 4.8 is deprecated as of October 30, 2026 and will be removed on November 30, 2026. Over 50 changes, including a fix for shell commands being left running when Claude Code exits by itself.

Official announcement →

This article is a summary based on official documentation.

What changed

Claude Code 2.1.316 shipped on October 31, 2026. It adds environment variables for ultrathink and subagent effort, pooled budgets and layered policies on the Claude apps gateway, and deprecates fast mode for Opus 4.8. Over 50 changes span the Bash tool’s shell, model aliases on Vertex AI and Amazon Bedrock, gateway model checks, auto mode, the VS Code extension, cloud sessions and Claude Tag.

New features

  • CLAUDE_CODE_DISABLE_ULTRATHINK

    There was no way to stop the ultrathink keyword from raising a turn’s effort. This environment variable turns the ultrathink keyword off.

  • CLAUDE_CODE_SUBAGENT_EFFORT

    Subagents inherited the session’s effort level unless something named one. This variable sets the effort level subagents run at when nothing else names one, instead of inheriting the session’s.

  • CLAUDE_CODE_DISABLE_LOOPED_CROSS_CERT_DROP

    Claude Code drops system cross-certificates that certify each other in a loop. This variable keeps them, for servers whose private CA needs one.

  • Pooled group budgets on the Claude apps gateway

    Each member could only have their own limit. Now one monthly budget that a group’s members share can sit on top of each member’s own limit, set with pooled_budget on the spend limits admin API.

  • managed.policy_mode: layered on the Claude apps gateway

    Only the first policy that matched a developer applied. With layered, every policy that matches a developer applies, each merged over the ones listed below it.

  • guardrail.input_tagging on Bedrock upstreams

    An Amazon Bedrock guardrail’s prompt attacks filter needs prompts wrapped in input tags. On Claude apps gateway Bedrock upstreams, this setting wraps prompts in those tags, and all_messages_and_system also tags the system prompt.

  • [VS Code] Claude apps gateway sign-in from the chat panel

    Signing in to a Claude apps gateway needed a terminal. You can now sign in from the chat panel.

  • [Cloud sessions] Form card for MCP server requests

    When an MCP server asked for a form in a cloud session, the request was canceled without you seeing it. A form card now lets you fill in or decline the form.

  • [Cloud sessions] “Catching up…” line

    A “Catching up…” line now shows while a cloud or Remote Control session you switch back to is still bringing its messages up to date.

  • [Code Review] Findings-by-severity card

    The Code Review analytics page now has a card showing the period’s total findings, a daily chart stacked by severity, and each severity’s count.

Key improvements

  • Findings card for /code-review from Claude Desktop’s Code tab or the web

    When /code-review was started from the Code tab in Claude Desktop or on the web, cloud and claude remote-control sessions showed a plain-text list. They now show the findings card.

  • Faster /diff on large repositories

    Each refresh usually runs one fewer full-repository git command, and refreshes are spaced out when git is slow.

  • Messages when no GitHub remote is found

    When /ultrareview, a cloud session or a cloud agent finds no GitHub remote, the message now says to create a repository at github.com/new, and /ultrareview no longer suggests a gh command that gh rejects.

  • SSL certificate errors

    SSL certificate errors now name the likely causes, among them a server that leaves out an intermediate certificate, and the fix for each.

  • Server-managed settings for Team and Enterprise logins

    Settings fetched from Anthropic’s API needed an approval prompt before they applied. They now apply without one, and a stored copy that doesn’t verify is refused.

  • Auto mode keeps the scheduled-task prompt rule

    An autoMode.allow list without "$defaults" removed the built-in rule on scheduled-task prompts. It no longer does.

  • [VS Code] ultrathink as in the terminal

    ultrathink now runs its turn at xhigh effort and shows in a rainbow in the chat box, as in the terminal.

  • [Cloud sessions] Network secrets for Enterprise members

    Enterprise members could only view network secrets. They can now add and delete them on their own cloud environments.

  • [Cloud sessions] Changes pane badge

    The note about what the Changes list leaves out took a row above the diff. It is now a badge beside the branch name.

  • [Claude Tag] “How Claude participates”

    A Slack channel’s Respond automatically switch is now a “How Claude participates” choice on the channel’s Configure page and in Claude Tag admin settings.

Bug fixes

Shell and Bash tool

  • Shell commands being left running on macOS and Linux when Claude Code exits by itself, for example when credentials are refused or a cloud session is archived.
  • Bash aliases with line breaks in their value (except in Git Bash), and functions whose names start with a dash, being garbled or lost in the Bash tool’s shell.
  • Bash scripts started from the Bash tool stopping after their first command when SHELLOPTS is exported.
  • PowerShell hooks not blocking on exit code 2 when the command begins with a type cast, param, using or an attribute.
  • [Windows] The Bash tool losing your aliases and functions in Git Bash when the rc file sets errexit and pipefail and only winpty aliases, or none, are defined.
  • Grep and Glob never returning when the search process is killed at the timeout but does not exit, seen on Windows.

Models and networking

  • sonnet, opus and haiku on Vertex AI and Amazon Bedrock showing an error mid-session instead of switching to an older model when the newest is not yet available to the account or region.
  • Sub-agents that inherited a parent sub-agent’s model moving to the main model when a background job or message woke them.
  • /model <id> not warning when the model isn’t recognized by this version and auto-compact will assume a 200K context window.
  • Some network requests hanging when a compressed response was cut off or corrupted.

Claude apps gateway and security

  • The Claude apps gateway not checking every model a request names against availableModels, deniedModels and models — a request that names a blocked model now gets a 400 naming the field.
  • claude auth login refusing to run where managed settings require a Claude apps gateway — in a terminal it now shows the gateway sign-in.
  • --restricted sessions uploading files from outside the working directories when a Read allow rule covered them.
  • Plugin MCP headersHelper scripts receiving a corrupted CLAUDE_PLUGIN_ROOT and server URL when a credential-named environment variable was set to a switch value such as 1.

Auto mode and headless

  • A managed autoMode.allow list without "$defaults" getting the built-in exceptions back from a host app’s --settings, a first rule added in /permissions, or /auto-mode-setup.
  • Cloud sessions and cloud scheduled tasks not being marked as failed when auto mode stopped a turn because the server returned no safety verdict.
  • claude -p text output sometimes not saying why auto mode stopped a turn when a background agent was still running.
  • CLAUDE_CODE_AUTO_MODE_SERVER=1 being ignored in sessions started by a host application (such as Claude Desktop) behind a proxy or LLM gateway with experimental betas turned off.

SDK and tools

  • The SDK register_repo_root control request not loading the CLAUDE.md of a directory outside the working directory (with CLAUDE_CODE_ADDITIONAL_DIRECTORIES_CLAUDE_MD=1).
  • The Artifact tool showing ? for artifact database access rules at the self level, and not quoting those rules back when it refuses a republish that would drop them.

Display and input

  • Prompts containing invisible characters needing a second Enter to send — plain prompts now send cleaned on the first Enter, and commands still wait for review.
  • Resuming a conversation not restoring its plan in a rare case, when /clear or /resume had been run while part of that conversation was still being saved.
  • The “hold space to speak” hint using up its three showings in one session, and showing an empty key when push-to-talk is unbound.
  • First-run setup dropping words from the security notes and the terminal-setup explanation in narrow terminals.
  • The Enterprise survey card taking a typed 1 or 0 after a short pause — a typed answer now needs Enter.

Self-hosted runner

  • A message sent while Claude was working being answered after a Stop, once the session’s process had been replaced.
  • An MCP server’s on/off switch sometimes being lost when a session moved to a new machine.

VS Code extension

  • The Effort slider, Ultracode switch and model name missing when the model setting has the [1m] tag, such as opus[1m].
  • Saved /btw side-question history being deleted too early while a settings file has errors.
  • A conversation coming back in Auto mode after a reload or restart when its plan was approved with “Yes, and manually approve edits”.

Cloud sessions

  • File search in a cloud session’s Files pane showing folders as ../.. paths after Claude changed directory — results now show their path from the project’s root.

Claude Tag

  • Claude refusing to edit or append to a canvas it made in a Slack channel, from that channel, once the canvas was also shared to another channel or a DM.
  • Untagged replies going unanswered in a Slack Connect thread when both organizations have installed Claude and Slack delivered the reply to the other organization’s app.
  • A cloud scheduled task’s live status checklist in Slack staying half-done — a run can now edit the top-level messages it posted during that run.

Notes

  • If you use fast mode on Opus 4.8, migrate to Opus 5.5 before November 30, 2026. Fast mode for Opus 4.8 is deprecated as of October 30, 2026 and will be removed on November 30, 2026. After removal, fast mode requests on Opus 4.8 return an error and do not fall back to standard Opus 4.8.
  • To run subagents at a different effort than the session, set CLAUDE_CODE_SUBAGENT_EFFORT. It applies when nothing else names an effort level for the subagent.
  • For Team and Enterprise logins, server-managed settings fetched from Anthropic’s API now apply without an approval prompt.
  • If you manage an autoMode.allow list without "$defaults", the built-in rule on scheduled-task prompts now stays in place.
  • This post covers the main items. See the GitHub release notes for the full list.