Git in a Minute: Worktrees

In one sentence

A Git worktree is an additional working directory linked to the same repository, usually checked out on a different branch.

Why it matters

git checkout switches one working tree. That forces serial work. Agents that edit files concurrently will collide unless each agent has its own directory and branch.

How it works

A repository has one main working tree plus zero or more linked worktrees. Each worktree:

• has its own directory on disk;
• checks out one branch (a branch cannot be checked out in two worktrees at once);
• shares the same .git object database.

Create and remove worktrees with git worktree commands. List them when you lose track of agent directories.

Example

git worktree add ../app-feature-a -b feature/a
git worktree add ../app-feature-b -b feature/b
git worktree list

Point each coding agent at its own directory. When finished:

git worktree remove ../app-feature-a

Agentic coding use

Worktrees are the standard pattern for parallel agents: same repo history, separate dirty states, separate branches, merge via pull requests. This is infrastructure for agentic development, not a niche Git trick.

Watch out

Worktrees still share remotes and credentials configuration. Isolation of files is not isolation of secrets or deploy keys. Also clean up stale worktrees or disk and branch clutter will grow quietly.