worktree-sync

Worktree Sync

Create and maintain git worktrees so each open PR branch has its own directory on disk. No stashing. No context switching. Open multiple editor windows on different branches simultaneously.

Step 1 — Determine Mode

If $ARGUMENTS contains --new <branch-name>: skip to Step 3 (create one new branch + worktree). Otherwise: run Step 2 (sync all open PRs).

Step 2 — Sync All Open PR Worktrees

# Verify GitHub CLI auth
gh auth status

# Create worktree base directory
mkdir -p ./tree

# Create a worktree for each open PR branch
gh pr list --json headRefName --jq '.[].headRefName' | while read branch; do
  safe_name="${branch//\//-}"
  branch_path="./tree/$safe_name"
  if [ ! -d "$branch_path" ]; then
    echo "Creating worktree: $branch → $branch_path"
    git fetch origin "$branch" 2>/dev/null || true
    git worktree add "$branch_path" "$branch"
  else
    echo "Already exists: $branch_path (skipping)"
  fi
done

git worktree list

After sync: open each ./tree/<branch>/ in a separate terminal or editor window. Each directory is a full independent working copy sharing the same .git object store.

Step 3 — Create New Branch + Worktree

Use when starting a new feature branch and wanting an isolated working directory from day one.

# branch_name comes from --new argument (e.g. --new feature/auth)
branch_name="<extract branch name from $ARGUMENTS after --new>"
repo_root=$(git rev-parse --show-toplevel)
parent_dir=$(dirname "$repo_root")
safe_branch="${branch_name//\//-}"
worktree_path="$parent_dir/$(basename "$repo_root")-$safe_branch"

git worktree add -b "$branch_name" "$worktree_path"
echo "Worktree created at: $worktree_path"
echo "Open this directory in a separate editor window to work in parallel."
git worktree list

Step 4 — Prune Stale Worktrees

If a worktree directory was deleted manually, clean up the dangling reference:

git worktree prune
git worktree list

$ARGUMENTS