In your coding agent · team skills

Run a store's day in one command.

Three team skills ship with the datavessel CLI. Your AI — Claude Code or ChatGPT — is the lead: it reads the live roster, picks the specialists that fit the goal, dispatches them, and hands you one report. Every write still waits for your sign-off.

Three teams, one pattern

Operator, Marketing, Builder.

Each is a one-command skill over the same workflow library the operations center runs — one client store at a time, ledger receipts included.

01

Ecommerce Operator Team

Runs a client's store day to day — orders moving, customers answered, revenue protected — and records what happened in the ledger.

  • Platform operationsthe Order Desk on Shopify, WooCommerce, or Shopware
  • Supportticket resolution & CSAT rescue
  • Revenue protectioncheckout recovery, dispute evidence
  • Reportingtraffic health, revenue alerts, the client report

The Operator team in full

02

Ecommerce Marketing Team

Grows qualified demand for a client's store — organic, AI answers, content, paid hygiene, lifecycle — and compounds learning every cycle.

  • SEOgrowth autopilot, audits, execution
  • AEOAI-answer visibility monitoring & fix-it
  • Content & lifecycleproduct copy refresh, winback email
  • Paid searchwaste/winners hygiene; campaign creation ships nextadvisory

The Marketing team in full

03

Store Expansion Builder

Builds new surface for a client store that already trades: landing pages, campaign pages, deeper catalogue — grounded in demand data.

  • Page buildingSEO landing & CMS pages on Shopware, draft-by-default
  • Catalogue depthproduct content refresh on your platform
  • SEO plumbingclean URLs, meta, internal links
  • Shopify / Woo expansioncopy-ready build plans while their page tools landadvisory

The Builder team in full

How a cycle runs

Your AI leads. Specialists execute. You sign off.

Each team is your model leading specialists from the workflow library — over the same 200+ typed tools the operations center, MCP and the CLI use.

01

Recall

Your AI lead reads that store's memory first: prior cycles, outcomes, and the standing rules you've set for this client.

02

Plan

It turns your goal — or picks the highest-leverage one itself — into a few scoped assignments.

03

Dispatch

Specialist agents run in parallel, each on its own lane, each with only the tools its job needs.

04

Approve

Any write that needs sign-off pauses that specialist and puts the item in your inbox — web or mobile. The rest keep working.

05

Report

Every action lands in the client's ledger with who approved it; one synthesized summary — done, pending, recommended — sits on top.

Reads run free. Writes ask first — every specialist carries the same approval gate, and your AI lead can surface an approval but never grant one. Nothing touches a client's store without a row in the ledger and, unless you've allowed autopilot for that step, your sign-off.

Setup, once

The teams remember your stores.

One interview saves a setup profile — stores, properties, brand voice — to per-store memory. Every team lead recalls it before it plans, so you never re-explain your client.

terminal

$ npm i -g datavessel-cli

$ datavessel setup

✔ Signed in as ops@agency.example

✔ Catalog synced: 213 tools

Not connected yet: google_analytics, shopify

Connect them here: app.datavessel.io/settings

✔ Connected: google_analytics, shopify, slack

✔ GA4 property: Web — Northwind (348211906)

✔ Saved your setup profile to team memory.

Setup complete. Your teams will recall this automatically.

One command from here.

Install the CLI plugin, run setup once, then ask for the operator, marketing, or builder team by name.