Blog

Cloning all my GitHub repos faster than the default

A Bash + Git + jq script to fetch every repo I own with submodules, handling pagination and dry-run previews.

14 October 2025 · Git, Automation, Bash, GitHub API

Why #

I wanted a faster, zero-click way to back up every GitHub repo I own, including submodules, without relying on the GitHub UI. The goal: a single command that pulls everything down (or simulates it) using SSH.

What I built #

Bash script using curl + GitHub API + jq to list all repos (with pagination).
Clones via git clone --recurse-submodules to keep submodules in sync.
--dry-run flag to preview what would be cloned without touching disk.
Skips already-present repos to avoid redundant work.

How it works (high level) #

Requires GITHUB_TOKEN with repo metadata read access; uses SSH for cloning.
Fetches paginated repo lists via GitHub API, writes targets into a loop.
For each repo: clone with submodules if missing; otherwise skip.
Optional dry-run prints planned actions only.

Notes #

Set GITHUB_TOKEN in your environment and ensure ssh-agent is running with your key added.
Configure CLONE_DIR in the script to choose where backups land.
Designed for Linux/macOS with Bash + curl + jq + git available.

Source