claudekit / updates / claude-code-2-1-318
[ PATCH · ]

Claude Code 2.1.318

Adds the date a stored claude.ai sign-in ends with an advance warning, an `effort` option for a mod's `$.model.fork`, and `assume_role` on `mantle` upstreams of the Claude apps gateway. Over 80 changes, including fixes for config files and the stored login being lost or changes being dropped with no word on shared network home directories.

Official announcement →

This article is a summary based on official documentation.

What changed

Claude Code 2.1.318 shipped on November 4, 2026. It adds the date a stored claude.ai sign-in ends with a warning ahead of it, an effort option for mods, and assume_role on mantle upstreams. Fixes make up more than half of the release. The largest group protects ~/.claude.json and the stored login on shared network home directories, with others covering restarted sessions, claude -p, cloud sessions and MCP servers.

New features

  • Sign-in end date in claude auth status and claude -p

    The date a stored claude.ai sign-in ends now appears in claude auth status (loginExpiresAt) and in claude -p JSON results (credential_expires_at).

  • Sign-in warning in claude -p text output

    A one-line stderr warning now starts 10 days before a stored claude.ai sign-in ends.

  • effort option for a mod’s $.model.fork

    A quick side job can now run at low whatever effort the session uses. The result reports the effort sent.

  • Record of your answer on mods Claude Code writes

    When Claude Code asks to turn on the mods it writes in a session, your answer is now recorded.

  • assume_role on mantle upstreams of the Claude apps gateway

    Use it for a role in another AWS account and for per-developer AWS cost attribution.

  • Google access token warning on the Claude apps gateway

    The gateway now logs a warning when the Google access token from a federation (external_account) credentials file is about to expire.

  • CLAUDE_RUNNER_SPAWN_REASON on self-hosted runners

    The orchestrator’s spawn-runner hook environment now says why Anthropic sent the request, for example runner_lost or resumed. It is unset when none was sent.

  • --release-idle-session-min startup warning on self-hosted runners

    A warning now shows when the flag is the last argument with no value (still ignored) or has a value of only spaces (still read as 0).

  • [Cloud sessions] No repository option

    A cloud session on an Anthropic-hosted environment can now start without picking a repository.

  • [Cloud sessions] Usual Chrome names

    Tools other than Playwright can now find the built-in Chromium through google-chrome and CHROME_PATH.

  • [Cloud sessions] New session button on the deleted environment card

    The card shown when a cloud session’s environment was deleted hides the message box. It now has a New session button.

Key improvements

  • Saves on shared network home directories

    ~/.claude.json, settings files and the stored login are now less often deleted, briefly emptied or replaced by a near-empty or older copy.

  • Sessions on network file systems look again for the configuration file

    When the file is wrongly reported as missing, sessions now look again, so fewer start without your MCP servers and trusted folders.

  • Unattended runs stop rewriting ~/.claude.json

    claude -p and SDK apps rewrote the file on every run even when nothing in it changed. They no longer do.

  • Auto-update reliability

    A brief network failure while verifying a release is now retried right away instead of waiting for the next check.

  • MCP tool errors

    A call made before the tool is loaded now returns its input schema.

  • Git on macOS’s built-in /bin/bash

    Git commits and pull request creation are more reliable when the Bash tool runs macOS’s built-in /bin/bash.

  • Remote session reconnects

    Remote sessions reconnect faster after a pause or sleep.

  • “Ask your admin for more usage”

    The option in the usage limit menu now sends the request in one step.

  • You should know toasts

    They now say “try again” instead of “press 2”, which only works in the terminal.

  • No IDE extension install from an IDE terminal

    Claude Code no longer installs or updates the IDE extension when started in an IDE terminal. Use /ide from an outside terminal or your editor’s extensions view.

  • Subagents under “Ask each time”

    Subagents now stop instead of switching models when “Switch models when a message is flagged” is set to “Ask each time”.

  • Background subagents’ shell commands

    A shell command now moves to the background after 3 minutes on all providers, so the prompt cache doesn’t expire while it runs.

  • Subagent progress summaries

    On Bedrock, Vertex, Foundry, or with telemetry off, a summary on a cold prompt cache can now wait up to 5 minutes for the subagent’s pending response.

  • What a mod attaches at prompt.submit

    It now stays in the user’s turn rather than going out as a system message.

  • deniedMcpServers URL entries

    They now 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.

  • /permissions views

    The rule details, built-in rules and delete confirmation views now use the standard dialog. Pressing Ctrl+C twice there now goes back or cancels instead of quitting.

  • /claude-api managed-agents-onboard default permissions

    A new agent’s built-in tools and a quickstart template’s MCP tools now get the auto permission policy (“Judge risky actions”) by default, matching the Claude Console. Say “run without asking” to opt out.

  • Claude apps gateway log messages

    Hints about listen.trusted_proxies now say to list only a proxy that adds the connecting address to X-Forwarded-For itself. The warning when AWS CountTokens fails on Amazon Bedrock now says Claude Code, not the gateway, counts after the 501, and points at IAM only when AWS denies access.

  • Self-hosted runner orchestrator log lines

    ok session=… now prints at every log level on stdout (was debug only, stderr) with reason=… after jti=…. nack session=… ends its fields with reason=… after retryable=….

  • Remote Control fork titles

    A fork of a terminal fork, when the app doesn’t set a title, is now created as “Name (fork 2)”.

  • [VSCode] Sign-in screen

    It now shows what Claude Code printed when its sign-in status check fails.

  • [Cloud sessions] Create PR failure messages

    Most failures in cloud and Remote Control sessions now give the reason or a link to GitHub instead of saying Try again.

  • [Cloud sessions] Connector search results

    They now tell Claude that the search covers only part of the connector directory.

  • [Code Review] Analytics page

    The page now retries once after a server or network failure, then shows a Try again banner instead of empty cards.

Bug fixes

Config files and sign-in

  • Non-interactive launches replacing a ~/.claude.json that isn’t valid JSON with defaults — the start now fails, says how to restore the file, and leaves it as is. (#87551)
  • An empty ~/.claude.json with no backup being reported as corrupted — Claude Code now starts as it does when the file is missing.
  • The restore command for a damaged or missing ~/.claude.json naming an empty or cut-off backup, or failing on a read-only file.
  • Retention cleanup leaving a home that only runs claude -p with no ~/.claude.json backup — it now takes a fresh backup before deleting the last good one.
  • /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.
  • The stored login being replaced by an empty file when a network home directory has no room left for data but still takes new files.
  • Jobs exiting with Not logged in when another terminal or host signs in again to the same account and organization on the same home.
  • The sign-in failure when a stored claude.ai login reaches its time limit — it now says so, and in claude -p names claude auth login and claude setup-token.
  • Tokens from setup-token being refused at start on machines with forceLoginOrgUUID, when the token belongs to a permitted organization.
  • Sessions being refused at startup by an allowedProviders list that another login on the same machine had stored.
  • Every connection failing when the 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.

Sessions and restarts

  • --continue sometimes losing the earlier conversation after a malformed tool call’s retries ran out (regression in 2.1.314).
  • Idle compaction replacing a conversation with a reply that was not a summary.
  • A restarted session re-sending a message that failed with an API error, ahead of the newer message waiting behind it.
  • A restarted session dropping its interrupted turn when the restart came right after Claude read a large image or retried a tool call.
  • A deleted session transcript being re-created with its last prompt when an idle non-interactive session exited.
  • macOS sessions with many running Bash commands stalling while memory grew.

claude -p and Agent SDK

  • Sessions reporting success when an API error cut off a reply while a mod hooked turn.step.
  • A mod’s turn.step hook making a stopped reply finish as a success in headless runs.
  • Custom slash commands that fail, or that only Claude can invoke, printing nothing.
  • A message sent right after a model change still running after Stop in headless sessions — Stop now cancels or stops it as it would a queued message.

MCP and plugins

  • 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.
  • Stdio MCP servers that exit on the version probe being re-probed on every start — a server that dies in under 3s for its own reason is also pinned to the legacy handshake for 7 days.
  • claude plugin commands exiting before a slow token refresh was saved, which could sign you out on the next run.
  • claude plugin validate passing a custom-named marketplace file that claude plugin marketplace add rejects.

Cloud sessions and self-hosted runners

  • A shell command approved after a restart running in the session’s start folder, not the folder it was asked in. If that folder can’t be entered, it is not run.
  • Claude not being told when an approved connector tool was unavailable after a cloud session restart.
  • Cloud sessions being stopped in the middle of a turn when the service refuses uploads for several minutes.
  • Variables a Setup hook writes to CLAUDE_ENV_FILE not reaching the Bash tool in some cloud sessions.
  • The iPhone app showing a cloud session’s turn as finished while it kept running, after a response was cut off at a tool call or a refused image or document was removed.
  • /mcp in cloud sessions counting tools the model can’t use — a server’s row now shows how many of its tools are available.
  • [Cloud sessions] Reading GitHub Actions logs — gh run view --log and gh api log requests now get the log in most environments, not an unreachable redirect.

Remote Control and version cleanup

  • 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.
  • A running Claude Code process owned by another user being treated as dead by version cleanup, for example one left by sudo claude, which could delete its binary.
  • /remote-control showing a session as available after a failed connection.
  • A misleading “spawn error: ENOENT” when a session’s working folder is missing — the error now names the folder.
  • Side-by-side checkouts of one repository looking identical in the environment menu’s Remote Control rows — each now shows its own folder name.

Tools and commands

  • Permission prompts being answered by keys typed just after sending a message, or before a slow prompt appeared.
  • Workflows re-running finished agents when resumed after a usage limit.
  • The “Now using usage credits” notice reappearing about every 30 seconds while a background subagent is running.
  • /design revoke reporting success when the request failed with a 404.
  • A duplicate notification when Claude left the status field out of a PushNotification call.
  • The error for a too-large image in a request with more than 20 images always saying the limit was 2000px — it now shows the API’s limit.
  • The runtime in background shell details not updating while the dialog is open.

VS Code extension

  • A window whose sign-in the organization’s settings refuse starting a new Claude Code process in the background every minute.
  • A long word in the chat panel’s error banner pushing its close button out of view.

Claude Tag

  • Keyboard focus and screen reader announcements in Claude Tag admin setup — pressing Escape in the connect form no longer also closes the Add dialog.

Notes

  • If a script reads claude -p stderr, expect a one-line warning starting 10 days before a stored claude.ai sign-in ends.
  • A ~/.claude.json that isn’t valid JSON now fails non-interactive starts instead of being replaced with defaults. The error says how to restore the file.
  • Starting in an IDE terminal no longer installs or updates the IDE extension. Use /ide from an outside terminal or your editor’s extensions view.
  • If a script reads self-hosted runner logs, update anything that matches retryable=true: or retryable=false:.
  • deniedMcpServers URL entries can now also block Anthropic-provided servers and connectors in some environments.
  • New agents from /claude-api managed-agents-onboard default to the auto permission policy.
  • Resuming a conversation started on 2.1.286 or earlier can now drop the model’s earlier thinking. The 2.1.288 fix that kept it was removed.
  • This post covers the main items. See the GitHub release notes for the full list.