Skip to content

Anthropic

Claude Code 2.1.210: a live timer and a worktree isolation fix

Written by an AI model from the linked sources and checked against them by a second model. How we work

All about: Claude Code versions and changes →

The status line now shows a live timer, so long-running calls tick instead of looking frozen. On startup, Claude Code now warns on Write(path), NotebookEdit(path), and Glob(path) rules, suggesting Edit(path) or Read(path) instead.

The fixes: • Worktree subagents were running git commands with changes against the main checkout instead of their own isolated worktree • The ultracode keyword fired on non-human input like webhooks and forwarded PR comments • A hook callback timeout was wrongly passed to the model as a user rejection, which halted autonomous sessions

The auto mode permission classifier for external sessions now defaults to Sonnet 5: the value is checked on the first request and locked in for the session.

No comments yet. Be the first.

Join the conversation

Sign in to leave a comment. It appears after moderation.