using-git-worktrees

To isolate feature branch work, you can use a git worktree to create a separate workspace. This prevents accidental changes to your active checkout while allowing you to run implementation steps safely without contamination.

The best way to use git worktrees for feature development is to use native worktree tooling first, falling back to safe git worktree add commands. This ensures proper isolation, ignore verification, and avoids harness-invisible phantom state.

After creating a git worktree, you set up dependencies by running project-appropriate install commands like npm, cargo, pip, poetry, or go. Then, you verify a clean test baseline before proceeding with implementation.

Git worktree creation includes submodule guarding to prevent false positives. It detects whether you are already in a linked worktree and guards against submodule issues before attempting to create a new isolated workspace.

If native git worktree tools are unavailable, a safe sandbox fallback uses the standard git worktree add command. This fallback follows a clear directory selection priority and mandates ignore verification for project-local directories.

You should use an isolated git workspace when starting feature work or before executing an implementation plan that should not risk current changes. This prevents accidental modifications to your current branch during development.