Getting started

Workloom is the package manager for AI agent configuration — install the wl CLI, bind a workspace, and render it into your provider directories (.claude/, .codex/, .cursor/, …).

Install the CLI

The install script detects your OS and architecture and drops a standalone binary in ~/.local/bin. No Node.js required.

macOS / Linux
curl -fsSL https://raw.githubusercontent.com/catesandrew/workloom-releases/main/scripts/install.sh | bash
Windows (PowerShell)
irm https://raw.githubusercontent.com/catesandrew/workloom-releases/main/scripts/install.ps1 | iex

Prefer npm? If you already have Node 20+:

npm
npm install -g @workloom/cli
one-off, no install
npx wl --help

Verify

wl --version

Authenticate

Sign in once so the CLI can reach your registry.

# authenticate once (or export WORKLOOM_DEV_TOKEN)
wl auth login

Bind a workspace and render it

Bind a project to a registry workspace, then render its packages into the provider directories your agents read.

# in a project directory
wl init                          # scaffold .loom/
wl bind --registry <workspaceId> # bind to a registry workspace
wl pull --dry-run                # preview what will be written
wl pull                          # render into .claude/, .codex/, .cursor/, …

Keep it in sync

Check for local drift and enforce a clean state in CI.

wl status                        # show owned files + drift
wl sync --ci                     # non-zero exit if drifted (gate CI)

Author and publish (optional)

Package your own agent configuration and share it through the registry.

wl review --local                # advisory AI check before publishing
wl push                          # upload a package as a draft
wl publish @scope/name@1.0.0     # promote draft to pending review