wt

Git worktrees enable parallel branch development by creating isolated working directories within a single repository, eliminating the friction of stashing or repeated cloning. You can address worktrees by branch name and switch between them effortlessly.

The best way to run parallel agent tasks is using isolated git worktrees for each task within a single repository. You can monitor progress with JSON-formatted listings and merge completed work back to the default branch through a structured pipeline.

To merge a git worktree back into the main branch, use a merge pipeline that commits, squashes, and rebases changes before merging. The pipeline runs pre-merge hooks for CI and tests, then automatically prunes the integrated worktree for cleanup.

Yes, you can list git worktrees with JSON output specifically designed for scripting automation. This allows you to programmatically inspect worktree states, run steps across multiple worktrees, and integrate with external CI or deployment lifecycle hooks.

Git worktree management supports configuring lifecycle hooks for CI, tests, and deployments. You can trigger these hooks automatically during worktree creation, switching, or within the merge pipeline to validate code quality before branch integration.

You should avoid using git worktrees when working on a single linear task without parallel development needs, as the overhead of managing isolated directories and lifecycle hooks provides no benefit over standard branch switching in a single working directory.