agentsclimarketplace

Ecommerce store

Skill Skryx-L-A/project-kit/skills/ecommerce-store

Stand up an e-commerce STORE / storefront in a project folder from the user's answers — Shopify-first, with a product catalog, collections, inventory, a working checkout, a discount, and analytics wired up. A project-kit sub-skill loaded by new-project's routing whenever the user wants to BUILD an online store / storefront / sell products / set up Shopify / a shop with checkout. Technical build only; the acquisition/marketing money strategy is build-business's job. Composes with website (custom storefront) and chains the Shopify MCP.From its SKILL.md

Install
npx -y skills add Skryx-L-A/project-kit --skill ecommerce-store

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • reads credentialsReads from 1 credential source: `.env`.
  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

7.1 KB, ~1.6k tokens by cl100k_base, as published. Nobody here has run it

ecommerce-store — get a real store live with a working checkout

What this sub-skill is for

Standing up an e-commerce store: a product catalog, collections, inventory, a discount, and a checkout that actually completes, plus analytics. Loaded by new-project when the project is "sell products online". Shopify-first (driven by the Shopify MCP); for a fully custom build it composes with website (storefront) + api-backend (cart/checkout API). Scope is the technical storefront — getting it live and transacting. The money/marketing side (offers, ads, acquisition) is build-business's job; chain it when the user wants customers, not just a store.

Mandatory grill-questions (fold into the Definition of Ready)

  • Platform: Shopify (default — via the Shopify MCP) or a custom stack (website storefront + api-backend)? Shopify unless there's a real reason to hand-roll checkout.
  • Catalog: what products, how many SKUs, variants (size/colour), pricing, digital vs. physical? Where do product data + images come from?
  • Payments, region & tax: which payment providers, selling into which countries, and what tax / VAT handling? (Shopify Payments availability varies by region — confirm.)
  • Shipping: rates, zones, free-shipping threshold, fulfillment (self / 3PL / dropship)?
  • Theme & brand: colours, logo, typography, voice — harvest a reference look if the user has one.
  • Channels: web only, or also social / marketplace? Domain to use?

Project sub-agents to generate (.claude/agents/)

  • catalog-builder — creates products, variants, and collections and sets inventory via the Shopify MCP (or the custom API), keeping titles/SKUs/prices consistent (delegate-by-default for catalog work). Confirms before publishing live products.
  • storefront-themer — sets up the theme / brand (Shopify theme settings, or the website storefront), responsive and on-brand; pulls tokens from a harvested reference.
  • listing-copywriter — writes product titles, descriptions, and metadata that are accurate and conversion-aware — benefits and real specs, no invented claims.
  • analytics-reporter — wires analytics and runs ShopifyQL / run-analytics-query to report sales, orders, and product performance; never fabricates numbers.

Tools / CLIs / MCP / skills needed

  • Shopify MCP (chain — primary): get-new-store-previews to spin up a store from a description (let the user browse previews and sign up through the link), then create-product, create-collection, add-to-collection, set-inventory, create-discount, run-analytics-query, get-shop-info, and graphql_query / graphql_mutation for anything without a built-in tool (metafields, pages, markets, etc.).
  • web-aeo (chain) — llms.txt + schema.org Product/Offer structured data so the store is legible to AI assistants and search.
  • design-harvest (chain) — pull palette/typography/layout tokens from a reference shop for the theme (patterns/tokens only, never copy assets/code).
  • Design chain (2026-07-29 addition): for a custom storefront or a hand-edited Shopify theme (real HTML/CSS/Liquid, not the theme editor's presets), design-bausteine structures the build brief and slop-detect is a mandatory ship gate before the theme goes live — same order as website's SKILL.md. A storefront is client-facing by definition, so also run design-critique before publishing.
  • Compose, don't duplicate: website for a custom storefront, api-backend for a custom cart/checkout, build-business for the acquisition/marketing engine.
  • Global skills/agents to chain: deep-research (payment/tax/region rules, platform trade-offs), verify (walk a real checkout end-to-end), code-review (custom-storefront code), market-researcher (agent) for facts on fees / payment providers / shipping.

File / asset nudges (on top of the base set)

  • STORE.md — platform, store URL/domain, payment + tax + shipping config, and the channel list. Single source of truth for how the shop is set up.
  • catalog/ — product source data (titles, descriptions, prices, SKUs, variants) and the collection structure, version-controlled before it's pushed live.
  • assets/products/ — product imagery (own or licensed — track the license).
  • POLICIES.md — returns / refunds / shipping / privacy text, drafted to the relevant consumer law, ready to paste into the store's policy pages.
  • .env.template — any API keys/tokens as empty keys; real .env git-ignored (never commit Shopify or payment credentials).

Stack defaults & done-bar

Default stack: Shopify via the Shopify MCP — store from get-new-store-previews, catalog

  • collections + inventory through the MCP tools, Shopify Payments + Shopify-managed checkout, a percentage discount code, analytics via ShopifyQL, schema.org from web-aeo. Swap to a website + api-backend custom build only when the grill calls for it. Done-bar (all must hold):
  1. The store is live with real products in at least one collection, inventory set.
  2. Checkout completes — a test order goes through end-to-end (verify), payment + shipping + tax resolving correctly for the target region.
  3. At least one discount code exists and applies correctly at checkout.
  4. Analytics is wired and returns real sales/traffic data (ShopifyQL or the chosen tool).
  5. Product listings are accurate; returns/shipping/privacy policies are present.

Guardrails

  • No fake reviews and no fake scarcity — no invented "12 people are viewing", fake countdowns, or planted testimonials. Social proof must be real.
  • Consumer-law / returns compliance — a clear, lawful returns + refund + shipping + privacy policy for the selling region; don't ship a store that can't legally take an order.
  • Product claims must be accurate — describe what the product actually is and does; no unverifiable health/performance/"best" claims. Flag anything the user can't substantiate.
  • Never expose payment or store secrets — API keys, Shopify Admin tokens, and payment credentials live in env / the platform vault, never in the repo, listings, or chat.
  • Confirm before going live / charging — publishing products, taking real payments, and creating discounts are outward-facing; confirm with the user first.
  • NO emojis in store UI, product copy, or notifications. Commits under the user's own name only (Skryx-L-A) — never add Claude as a co-author. Never commit secrets / API keys.

Keep looking

Skills are one crate of 325,949. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.