shiichan

Claude Code v2.1.271 Brings Fast Mode to Remote Sessions!

Hi everyone, it's Shii-chan, and I've got another one for you today! Let's take a look at the newest Claude Code release, v2.1.271.

Claude Code Release github.com

What was announced?

Claude Code v2.1.271 landed on the GitHub releases page. Along with new features, this release packs in a wide range of bug fixes.

Here's what's new:

  • Fast mode now works in Claude Code Remote sessions (both cloud and self-hosted runners): the host's fast-mode setting, or /fast typed in the session, applies wherever your organization allows it
  • Mouse support was added to the /config panel in fullscreen mode: the wheel scrolls the settings list, clicking a setting's value changes it, and the row under the pointer is highlighted
  • Per-command allowed_domains was added to Bash, PowerShell, and Monitor in auto mode with sandboxing: the hosts a command needs are reviewed and opened for that command alone, while every other host stays blocked
  • omitClaudeMd was added to agent frontmatter and the --agents JSON, letting custom and plugin subagents run without loading user, project, and local CLAUDE.md files (managed policy files still load)
  • An --accept-command flag (taking a SHA-256 hash) was added to claude plugin install and claude plugin update, letting you accept exactly the command a previous --json run displayed, instead of using -y
  • The modelPricing managed setting and the Claude apps gateway pricing block now support a multiplier above 1, up to 10, for marked-up internal chargeback rates

What changes

  • If you wanted Fast mode in Remote sessions, v2.1.271 finally gives you that
  • Anyone running Bash or PowerShell in a sandboxed environment can now allow only the hosts a specific command needs, keeping things flexible without giving up security
  • Plugin authors distributing subagents can use omitClaudeMd to get more predictable behavior that doesn't depend on whatever CLAUDE.md happens to be present
  • Organizations that charge back Claude Code usage internally get more control over billing with the new modelPricing multiplier

Dive Deep

The bug fixes cover a lot of ground, and several look like things people were quietly running into.

  • Fixed a cached organization policy being reused after switching accounts, organizations, or API keys
  • Fixed schema validation errors on subagent tool calls in cloud sessions after approving a workflow or agent
  • Fixed MCP OAuth client registrations racing across multiple registrations, removals, and writes
  • Fixed Code Review posting duplicate findings, and re-posting security findings that had already been resolved
  • Fixed Claude Tag's working context in a channel resetting roughly every hour

On the changes side, Monitor now always runs with a deadline — 30 minutes by default, 10 minutes for a single -p prompt — and the old persistent option is gone. The default dynamic workflow size guideline for Pro plans also dropped from 15 agents to 10.

Wrap-up

  • Claude Code Remote sessions now support Fast mode
  • Bash, PowerShell, and Monitor gained per-command allowed_domains for more flexible sandboxed operation
  • omitClaudeMd for subagents and --accept-command for plugin commands strengthen operational tooling
  • Fixed a cached organization policy bug, MCP OAuth registration races, duplicate Code Review findings, and more small-but-annoying issues

This one's especially worth it if your organization relies on Remote sessions, a Claude apps gateway, or Claude Tag!