npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details

    pkg:[package-name]

  • User packages

    @[username]

Sponsor

Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.

About

Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2026 – Pkg Stats / Ryan Hefner

@liaisonio/cli

v0.2.13

Published

Liaison Cloud CLI — manage connectors, entries, and applications from the command line

Readme

Liaison Cloud CLI

Official command-line interface for liaison.cloud, designed to be scripted and agent-friendly.

  • One-shot bootstrapliaison quickstart creates a connector + application + public entry in a single call
  • 5 Agent Skills — drop-in Skill files for AI agents (Claude/Cursor/etc.)
  • JSON output by default — pipe into jq or parse from any LLM agent
  • --output table for humans, --output yaml when you prefer it
  • Credentials from env var (LIAISON_TOKEN), config file, or explicit --token flag
  • Browser-based PAT login (or --no-browser for headless / SSH)
  • Every command has -h / --help with examples
  • Non-interactive by default — destructive operations require --yes

Install

Install the CLI and Agent Skills together — the CLI is the tool, the skills teach your AI agent how to use it.

# 1) Install the CLI (pick one)
npm i -g @liaisonio/cli                               # npm (recommended)
npx -y @liaisonio/cli@latest whoami                   # or run without installing
curl -fsSL https://liaison.cloud/install-cli.sh | sh  # or curl one-liner

# 2) Install Agent Skills (for Claude / Cursor / Continue / etc.)
npx skills add liaisonio/cli -y -g

That's it. Your AI agent now has both the binary and the knowledge to drive it.

npm i -g @liaisonio/cli and the curl installer also drop the bundled skill files at ~/.claude/skills automatically as a safety net — set LIAISON_CLI_SKIP_SKILLS=1 to opt out.

If npx skills add can't reach GitHub

On networks that can't clone GitHub (e.g. some CN environments), npx skills add may time out. The skill files are embedded in the CLI binary, so you can install them offline instead — and across every AI agent detected on the machine (Claude Code, Codex, Cursor, Pi, Trae, OpenClaw), same coverage as npx skills add liaisonio/cli -a '*':

liaison skills agents                  # which agents are detected here
liaison skills install                 # fan out to every detected agent
liaison skills install --agent claude  # just Claude Code
liaison skills install -p              # ./.claude/skills (per-repo)
liaison skills install --force         # overwrite existing copies
liaison skills uninstall               # remove liaison-* from every agent

No network needed — the SKILL.md files are read straight out of the liaison binary you just installed. liaison skills uninstall also fills a gap where npx skills remove doesn't touch agents like Pi.

Alternative install methods

go install github.com/liaisonio/cli/cmd/liaison@latest

Requires Go 1.22+. Best for Go developers who already have $GOPATH/bin in their PATH.

git clone https://github.com/liaisonio/cli
cd cli
make build           # ./bin/liaison           (current platform)
make release         # ./dist/liaison-*        (all 5 platforms + SHA256SUMS)

Verify

liaison version

Agent Skills

The CLI ships 5 Skill files so AI agents know how to use it without a learning curve. Each skill is a self-contained Markdown spec with frontmatter.

| Skill | Purpose | |-------|---------| | liaison-shared | Auth, install, token precedence, error handling, output format (auto-loaded by other skills) | | liaison-quickstart | One-shot bootstrap: connector + application + entry in a single call | | liaison-connector | Connector lifecycle: create / list / inspect / enable+disable / delete | | liaison-application | Backend service metadata: register / list / update / delete | | liaison-entry | Public exposure: HTTP domains, TCP ports, enable+disable, delete |

After installing the skills, point your agent at liaison.cloud and ask it things like:

  • "Set up a public SSH endpoint for my home server"
  • "List all my connectors and tell me which ones are offline"
  • "Disable connector 100017 — I'm doing maintenance"
  • "Expose the local Postgres on 5432 via Liaison"

Authenticate

The CLI uses long-lived Personal Access Tokens (PATs) — liaison_pat_xxx... issued by the Liaison dashboard. Three ways to provide one:

# 1) Browser flow (recommended for humans)
liaison login
# Opens https://liaison.cloud/dashboard/cli-auth in your default browser,
# you click "Authorize", a fresh PAT is minted and persisted to ~/.liaison/config.yaml.

# 2) SSH / headless / no browser
liaison login --no-browser
# Prints the URL — open it on any device that has a browser, click Authorize,
# the CLI receives the token via a localhost callback.

# 3) Already have a token (CI, agent secrets store)
LIAISON_TOKEN=liaison_pat_a1b2c3... liaison whoami
# Or: liaison login --token liaison_pat_a1b2c3...

Precedence (highest wins): --token flag > LIAISON_TOKEN env > ~/.liaison/config.yaml > no token.

Tokens can be revoked any time at liaison.cloud > Settings > API Tokens, or by running:

liaison logout

Quick Start

The fastest way to expose a local service — one command does everything:

# 1) authenticate once
liaison login

# 2) create connector + install it locally + register app + expose publicly
#    (needs sudo for the install step)
liaison quickstart --name mybox \
  --app-name web --app-ip 127.0.0.1 --app-port 8080 --app-protocol http \
  --expose --install --wait-online 2m

This single command:

  1. Creates a connector on liaison.cloud
  2. Downloads and installs the connector agent on this machine (--install, needs sudo)
  3. Waits up to 2 minutes for the connector to come online (--wait-online 2m)
  4. Registers a backend application (--app-* flags)
  5. Creates a public entry so the service is reachable from the internet (--expose)

The output JSON includes everything you need:

{
  "connector": { "id": 100042, "name": "mybox", "access_key": "...", "secret_key": "..." },
  "install_command": "curl ... | bash -s -- ...",
  "installed": true,
  "online_achieved": true,
  "application": { "id": 1, "name": "web", "ip": "127.0.0.1", "port": 8080 },
  "entry": { "id": 10, "name": "web", "port": 34567, "domain": "web-username.liaison.cloud" }
}

If you're installing on a different machine (not the one running the CLI), omit --install and run the install_command from the output on the target host manually:

liaison quickstart --name mybox \
  --app-name web --app-ip 127.0.0.1 --app-port 8080 --app-protocol http \
  --expose --wait-online 2m

# Then on the target host:
# curl -k -sSL https://liaison.cloud/install.sh | bash -s -- --access-key=... --secret-key=...

See liaison quickstart --help for the full flag list.

Step-by-Step Guide

If you need more control than quickstart, you can create resources one at a time.

1. Create a connector

liaison edge create --name my-server --description "home lab"

Output includes access_key, secret_key and a one-line install command:

{
  "access_key": "MTc3...",
  "secret_key": "20S...",
  "command": "curl -k -sSL https://liaison.cloud/install.sh | bash -s -- --access-key=... --secret-key=..."
}

2. Install the connector agent

Run the command from step 1 on the target machine (needs curl + bash + sudo):

curl -k -sSL https://liaison.cloud/install.sh | bash -s -- \
  --access-key=MTc3... --secret-key=20S... \
  --server-http-addr=liaison.cloud --server-edge-addr=liaison.cloud:30012

The connector agent will start automatically and connect to liaison.cloud.

3. Register a backend application

Once the connector is online, register the service running behind it:

# SSH service
liaison application create \
  --name my-ssh --protocol ssh \
  --ip 127.0.0.1 --port 22 \
  --edge-id 100017

# HTTP web app
liaison application create \
  --name my-web --protocol http \
  --ip 127.0.0.1 --port 8080 \
  --edge-id 100017

Supported protocols: tcp, http, ssh, rdp, mysql, postgresql, redis, mongodb.

4. Expose via a public entry

# SSH entry — gets an auto-allocated public port
liaison proxy create --name my-ssh-entry --protocol ssh --application-id 100015
# => access at: ssh -p <port> [email protected]

# HTTP entry — gets a subdomain like myapp-username.liaison.cloud
liaison proxy create --name my-web-entry --protocol http --application-id 100038
# => access at: https://my-web-entry-username.liaison.cloud

All-in-one (equivalent of the 4 steps above)

# On the same machine — install + wait + app + expose in one shot
liaison quickstart --name mybox \
  --app-name web --app-ip 127.0.0.1 --app-port 8080 --app-protocol http \
  --expose --install --wait-online 2m

# On a different machine — omit --install, run install_command manually
liaison quickstart --name mybox \
  --app-name web --app-ip 127.0.0.1 --app-port 8080 --app-protocol http \
  --expose --wait-online 2m

Command Reference

liaison whoami                                    # who am I logged in as?

# Connectors (edges)
liaison edge list
liaison edge list --online 1                      # only online connectors
liaison edge list --output table
liaison edge get 100017
liaison edge create --name lab-server --description "office lab"
liaison edge update 100017 --status stopped       # disable + kick
liaison edge update 100017 --status running       # re-enable
liaison edge delete 100017 --yes

# Backend applications (IP:port exposed by a connector)
liaison application list
liaison application create --name my-ssh --protocol ssh --ip 192.168.1.10 --port 22 --edge-id 100017
liaison application update 123 --port 2222
liaison application delete 123 --yes

# Entries (public proxies)
liaison proxy list
liaison proxy create --name my-ssh-entry --protocol ssh --application-id 123
liaison proxy update 456 --status stopped
liaison proxy share 456                              # temp share link (http entries, ~1h)
liaison proxy share 456 --redirect /admin            # guest lands on /admin
liaison proxy delete 456 --yes

# Devices
liaison device list
liaison device get 789

Global flags

| Flag | Env | Description | |---------------|------------------|-------------------------------------------------------| | --server | LIAISON_SERVER | Liaison base URL (default https://liaison.cloud) | | --token | LIAISON_TOKEN | JWT bearer token | | --config | | Config file path (default ~/.liaison/config.yaml) | | --output,-o | | json (default), yaml, or table | | --insecure | | Skip TLS verification (self-signed testing only) | | --verbose,-v| | Print each HTTP request to stderr |

Output formats

liaison edge list                    # pretty JSON (default)
liaison edge list -o yaml            # YAML
liaison edge list -o table           # aligned text table
liaison edge get 100017 | jq .name   # pipe into jq

Exit codes

  • 0 — success
  • 1 — any error (auth, network, API error, invalid args)

Error messages go to stderr; output goes to stdout — safe to redirect.

License

Apache 2.0