new-session

Creating isolated git worktrees with unique dev server ports involves provisioning a dedicated worktree, copying environment files from the main repository, and assigning a unique port via a lock-based system to prevent cross-branch conflicts.

Preventing dev server port conflicts when running multiple worktrees requires managing per-worktree port assignments using lock files stored under .claude/ports, which automatically cleans stale locks and avoids default port 3000 for concurrent sessions.

Yes, you need to copy env files when setting up a new git worktree to recreate the correct runtime context. Mirroring environment files like .env.local from the main repository prevents environment drift across isolated development workspaces.

Yes, you can symlink node_modules from the main repo into a new worktree to ensure consistent dependencies and achieve faster startup times without needing to reinstall packages for each isolated development workspace.

The best way to manage environment drift in git worktree sessions is to automatically mirror main repository environment files into the new worktree and maintain a lock-based port assignment system to isolate runtime contexts across branches.

Your dev server might use port 3000 because a per-worktree port lock was not established. Managing port assignments under .claude/ports explicitly avoids port 3000 and assigns unique, conflict-free ports to support multiple concurrent sessions.