Get started
Workloom is an org-level package manager for AI dev-env configs — define canonical packages of skills, agents, and prompts once, then render provider-specific projections for Claude, Codex, Copilot, and Gemini. Install the wl CLI to get started.
Install the CLI
The one-line installer detects your platform, downloads the prebuilt binary, and verifies its checksum. No Node required — the binary bundles its own runtime.
macOS / Linux
curl -fsSL https://raw.githubusercontent.com/catesandrew/work-loom-releases/main/packaging/install.sh | sh
On Windows, run the PowerShell installer instead:
Windows (PowerShell)
irm https://raw.githubusercontent.com/catesandrew/work-loom-releases/main/packaging/install.ps1 | iex
Prefer to inspect first? Download, read, then run:
curl -fsSL https://raw.githubusercontent.com/catesandrew/work-loom-releases/main/packaging/install.sh -o install.sh less install.sh sh install.sh
Have Node 20+? Install from npm instead:
npm
npm install -g @workloom/cli
Want a specific version? Grab a binary from the releases page. Verify it worked:
wl --version
Point at your registry & sign in
Tell the CLI where your Workloom API lives (once, global), then authenticate — this opens your browser for sign-in.
mkdir -p ~/.config/workloom printf 'registry_url = "https://warp-api.workloom.app"\nauth_mode = "supabase"\n' \ > ~/.config/workloom/config.toml wl login wl whoami
Bind a repo & render
Bind a repository to a workspace, preview what would be written, then materialize the rendered provider files.
cd my-repo wl init # interactive: workspace id + registry wl pull --dry-run # preview — writes nothing wl pull # render + write wl status # binding + state wl diff # any local drift?
Day-to-day: wl sync reconciles, wl sync --check fails CI on drift, and wl keep-local <path> protects a hand-edited file.
Publish a package (optional)
Push a local asset directory as an immutable, published version — like npm publish for AI configs.
wl push .claude/skills/security-review \ --package @acme/security-review \ --bump minor
Install a package directly
Skip the workspace bind — install one published package into your home AI dev-env à-la-carte, for whichever provider you use (requires wl login).