Work on multiple branches of one repository at the same time using separate linked working trees via git worktree instead of separate clones

Documented steps

1. Create a new worktree with a brand-new branch: `git worktree add ../hotfix -b hotfix` — creates branch `hotfix` starting at HEAD and checks it out at path `../hotfix`.
2. Create a worktree for an already-existing branch: `git worktree add ../release release-1.0` — checks out `release-1.0` there; Git refuses if that branch is already checked out in another worktree (use `-f`/`--force` to override, with care).
3. Create a throwaway worktree with no branch (detached HEAD) for quick experiments: `git worktree add --detach ../scratch`.
4. Omit the branch name for a convenience default: `git worktree add ../hotfix` alone infers a branch named `hotfix` (the path's basename) — creating it from HEAD if it doesn't exist, or checking it out if it does and isn't in use elsewhere.
5. List all worktrees with their branch/HEAD/lock/prunable status: `git worktree list`, or `git worktree list --porcelain` (add `-z` for NUL-terminated output) for scripting.
6. Lock a worktree that lives on removable/network storage so it can't be auto-pruned, moved, or deleted: `git worktree lock --reason "on USB drive" ../hotfix`; reverse with `git worktree unlock ../hotfix`.
7. Remove a finished worktree: `git worktree remove ../hotfix` — only succeeds if the worktree is clean (no untracked or modified tracked files); pass `-f`/`--force` to override uncleanliness, or `--force` twice to remove a locked worktree.
8. Clean up stale metadata after a worktree directory was deleted by hand (e.g. `rm -rf`) instead of via `remove`: `git worktree prune -n` (dry run) then `git worktree prune -v`.