Introduction

Use one command to work with every Git repository in a workspace.

Git Wrangler is a command-line tool for people who manage several related Git repositories. Run it from a workspace folder and it finds the repositories below that folder automatically.

Quick start

# Check that Git Wrangler and its optional tools are ready
git-wrangler doctor

# See the state of every repository in this workspace
git-wrangler status

How targeting works

Run a command from a folder that contains your repositories to apply it across the workspace:

cd ~/code/my-workspace
git-wrangler status

Use --repo <path> when you want to work with exactly one repository:

git-wrangler status --repo ./api

Git Wrangler shows results in a stable order and finishes with a summary. If one repository fails, it continues with the others and reports the failure at the end.

Helpful common options

  • --guided asks simple questions before running a command.
  • --repo <path> targets one repository instead of the whole workspace.
  • --no-fetch skips refreshing information from origin; results may be out of date.
  • --yes or -y skips a confirmation. It never supplies required values.
  • --json provides machine-readable output on supported reporting commands.

Choose what you want to do

Safety basics

Commands that delete files, reset work, force-push, or rewrite history show a preview and ask once before making the change. Read that preview carefully. Use --yes only in intentional automated runs.

AI commands explain what they will send before contacting the configured service. Sensitive file contents and common secret values are hidden, and coauthor identities added with commit --coauthor stay local.

Use rollback-rewrites to undo supported Git Wrangler history rewrites. Secret removal is intentionally excluded from rollback so removed secrets cannot be restored accidentally.

Next steps