Skip to main content

Install

Install the latest Gitmoot beta with the installer:

curl -fsSL https://gitmoot.io/install.sh | sh
gitmoot version

Gitmoot depends on Git and GitHub CLI for repository and PR workflows:

git --version
gh --version
gh auth status

gh auth status must succeed before PR comments, issue comments, review publication, or merge workflows can work. Do not paste GitHub tokens into tracked files, issue comments, PR bodies, or logs.

Ask Your Agent To Install It

For Codex or Claude Code, the easiest path is to ask your active coding agent:

Install Gitmoot as a Codex or Claude skill/plugin in this repo, verify `gitmoot version`, run `gitmoot plugin doctor`, check `gh auth status`, and summarize the next Gitmoot workflow I can use.

The plugin or skill is the discovery and guidance surface. The gitmoot CLI and local daemon remain the execution path.

Install runtime plugin guidance when you want Codex or Claude Code to discover Gitmoot from its plugin system:

gitmoot plugin install codex
gitmoot plugin install claude
gitmoot plugin doctor

Direct Binary Fallback

If the installer is not appropriate, download the matching release artifact from GitHub Releases, verify it, make it executable, and move it onto PATH.

For Linux:

sha256sum gitmoot_linux_amd64
chmod +x gitmoot_linux_amd64
mkdir -p ~/.local/bin
mv gitmoot_linux_amd64 ~/.local/bin/gitmoot
command -v gitmoot
gitmoot version

For macOS, pick the Apple Silicon or Intel artifact when macOS artifacts are published:

shasum -a 256 gitmoot_darwin_arm64
chmod +x gitmoot_darwin_arm64
mkdir -p ~/.local/bin
mv gitmoot_darwin_arm64 ~/.local/bin/gitmoot
command -v gitmoot
gitmoot version

For Windows checksum verification:

certutil -hashfile gitmoot_windows_amd64.exe SHA256

Compare the printed SHA256 with the published checksum before running or moving the artifact. If the values differ, delete the download and fetch it again from the release page.

Update, Roll Back, Or Uninstall

Check for updates and restart the daemon after updating when it is running:

gitmoot update --check
gitmoot update --restart-daemon

For rollback, reinstall a specific versioned release artifact manually and verify it with gitmoot version.

To uninstall the CLI, remove the installed binary path returned by command -v gitmoot. Remove Codex or Claude plugin discovery only if you want the runtime to stop finding Gitmoot guidance. Do not delete ~/.gitmoot unless you intentionally want to remove local Gitmoot state, job history, and config.

Kimi Code Runtime

To run agents on the Kimi Code runtime (gitmoot agent start <name> --runtime kimi), install the external kimi CLI, then authenticate so background jobs can reuse the session:

kimi login
gitmoot update --restart-daemon

Restart the Gitmoot daemon after kimi login so it inherits the logged-in session. The Kimi runtime is a first-class adapter alongside Codex and Claude Code; it does not use the gitmoot plugin install discovery surface.

SkillOpt Optimizer

Gitmoot's SkillOpt train workflow invokes a separate Python optimizer only at optimizer handoff. Install and preflight it before running optimizer-backed training:

python3 -m pip install --user pipx
python3 -m pipx ensurepath
pipx install https://github.com/jerryfane/gitmoot-skillopt/releases/download/v0.4.2/gitmoot_skillopt-0.4.2-py3-none-any.whl
gitmoot-skillopt --version
gitmoot-skillopt optimize --help

If you install it in a virtualenv, pass the executable with gitmoot skillopt train continue --skillopt-bin /path/to/gitmoot-skillopt.