MCP HubMCP Hub
SKILL·D96C21

vertical-retail

avelikiy
Actualizado 19 days ago
5 vistas
53
11
53
Ver en GitHub
Metaaidesign

Acerca de

Esta habilidad proporciona conocimiento esencial del dominio minorista/comercio electrónico para desarrolladores que construyen productos de escaparates para pequeñas y medianas empresas. Codifica conceptos clave, errores comunes y modelos de datos necesarios como Producto-Variantes e inventario sensible al canal para evitar implementaciones ingenuas. Úsela al especificar funciones que involucren catálogos, inventario, precios o recuperación de carritos.

Instalación rápida

Claude Code

Recomendado
Principal
npx skills add avelikiy/great_cto -a claude-code
Comando PluginAlternativo
/plugin add https://github.com/avelikiy/great_cto
Git CloneAlternativo
git clone https://github.com/avelikiy/great_cto.git ~/.claude/skills/vertical-retail

Copia y pega este comando en Claude Code para instalar esta habilidad

Documentación

Retail & e-commerce — spec it like someone who's run a store

The SMB retail buyer already pays Shopify ($39–399/mo + 2.9%), BigCommerce, Wix, Ecwid, or WooCommerce. They are not naive — so the spec can't be either. A storefront that "has products and a cart" is table stakes; the value is in the parts those platforms do badly. Read this before writing the catalog/inventory/pricing/cart sections of any retail ARCH or PLAN doc.

1. Domain vocabulary (use these exact words)

  • SKU vs variant — a variant is one buyable configuration (Red / Large); its SKU is the unique code that variant ships and is counted under. A "product" is the parent; you stock, price, and sell variants, not products.
  • Multichannel / omnichannel — selling across several channels (own storefront, Amazon, eBay, in-store POS, Instagram). Omnichannel additionally means one inventory pool behind all of them. Channel-awareness is the whole game for SMB inventory.
  • Reorder point — stock level that triggers a purchase order = (avg daily demand × lead time in days) + safety stock. Lead time = supplier days from order to receipt. Safety stock = buffer for demand/lead-time variance. Reordering without all three is wrong.
  • COGS (cost of goods sold) and margin = (price − COGS) / price. Landed cost = unit cost + freight + duties + handling; margin must use landed cost, not invoice cost.
  • ATS / available to sell = on-hand − allocated (reserved by open orders) − safety stock. Customers buy against ATS, never raw on-hand.
  • Backorder vs preorder — backorder = out of stock now, will refill (sell against incoming PO). Preorder = not released yet, future availability date. Different fulfillment promises.
  • Cart abandonment rate = 1 − (completed checkouts / carts created); industry ~70%.
  • AOV (average order value) and conversion rate = orders / sessions. The two levers pricing/promotions move.
  • Fulfillment — pick/pack/ship. Dropship = supplier ships direct, seller never holds stock (so "stock" is the supplier's ATS feed, not yours).
  • MAP (minimum advertised price) — supplier-imposed price floor; a pricing rule must respect it or the seller loses the brand.

2. Non-obvious domain rules

  • Shopify owns the storefront — don't fight it head-on. A me-too checkout loses. The wedge is the platforms' weak spots: multichannel inventory + reorder, and cart recovery. Spec the storefront as competent-and-owned, and put the differentiation in the other three.
  • Variants explode combinatorially. options (Size × Color × Material) multiply: 5×8×3 = 120 variants per product. The data model, UI, and import flow must assume hundreds of variants per product, each with its own SKU / price / stock — not a flat product list.
  • Inventory must be channel-aware. The same SKU is sold on storefront + Amazon + POS; stock must decrement across all and sync back, or you oversell. Single-channel inventory is the most common naive failure and the strongest wedge.
  • Pricing rules interact with floors. A promotion or demand-based rule must clamp to a margin floor and MAP. A rule that can price below landed-cost margin is a bug, not a discount.

3. What a naive build gets wrong

  • Products without a variant model — a flat product { price, stock } table. Breaks the instant the seller stocks two sizes. Variants are core, not an add-on.
  • Single-channel inventory — stock that lives only in the storefront, no sync across Shopify / Amazon / POS. Guarantees overselling for any real SMB.
  • Reorder without lead-time / safety-stock — "reorder when stock < 10" stocks out during the supplier lead time. Must use reorder-point math.
  • Cart recovery that ignores suppression / consent — emailing/SMSing without consent, or after unsubscribe/purchase, is illegal (CAN-SPAM / TCPA / GDPR) and burns deliverability. Honor suppression + quiet hours.
  • Pricing that ignores the margin floor — a promo engine that can sell below cost, or below MAP.

4. Must-model entities

EntityKey fields
Productid, title, option axes (e.g. Size, Color) — the parent
Variantproduct_id, option values (Red/L), SKU, price, COGS/landed cost — one per option combo
InventoryLevelvariant_id, channel/location, on_hand, allocated, safety_stock → derive ATS
ReorderRulevariant_id, reorder_point, reorder_qty, lead_time_days, supplier
PricingRulescope (variant/collection), trigger (demand/margin/schedule), action, margin_floor, MAP
AbandonedCartcart_id, customer, line items, value, abandoned_at, recovery state, consent/suppression

The Variant option matrix and the channel-keyed InventoryLevel are the two that naive specs collapse — keep them explicit.

5. Per-product notes (wedge + the one domain thing)

  • storefront (content) — catalog, checkout, themes; a store the seller owns. Wedge: owned channel + SEO (it must rank — see [[local-seo]]). The one thing: the Product→Variant model and clean indexable URLs. Don't out-engineer Shopify's checkout; match it and move on.
  • inventory (crud) — track stock across channels, auto-reorder before stockout. This is the underserved-by-Shopify wedge. The one thing: channel-aware InventoryLevel + reorder-point math (lead time + safety stock). Get this right and the product justifies itself.
  • pricing (dashboard) — rules-based pricing + promotions reacting to demand/margin. Wedge: margin-aware automation SMBs do by hand. The one thing: every rule clamps to margin floor + MAP.
  • cart-recovery (crm) — win back abandoned carts via timed email/SMS. Wedge: recovering the ~70% that abandon. The one thing: consent + suppression + timing — defer the messaging mechanics to [[lifecycle-messaging]].

6. Compliance (light — defer the heavy parts)

  • Sales tax nexus — economic nexus thresholds vary by US state (post-Wayfair); the seller may owe tax in states they've never shipped to. Note it in the spec; defer the actual calc/filing to billing. Don't hand-roll tax.
  • Email / SMS consent — cart recovery needs prior consent (CAN-SPAM / TCPA / GDPR), honored unsubscribe, and quiet-hours/suppression. Defer the delivery + consent machinery to [[lifecycle-messaging]]; the spec just states the requirement.
  • PCI — checkout uses Stripe-hosted elements so card data never touches our servers (SAQ-A scope). State that intent; defer the scope proof to pci-reviewer.

Output

When applied, contribute a Retail domain section to the ARCH/PLAN/DESIGN doc:

## Retail domain
- model: Product→Variant (option matrix, per-variant SKU/price/stock) · channel-aware InventoryLevel (ATS = on_hand − allocated − safety_stock)
- reorder: reorder_point = avg_demand × lead_time + safety_stock (not "< N")
- pricing: every rule clamps to margin_floor + MAP (margin on landed cost)
- cart-recovery: consent + suppression + timing → [[lifecycle-messaging]]
- wedge: multichannel inventory + reorder, cart recovery (don't fight Shopify's storefront/checkout)
- compliance: tax nexus → billing · consent → [[lifecycle-messaging]] · PCI Stripe-hosted (SAQ-A) → pci-reviewer
- migration: catalog/variant/stock import path → [[migration-ready-schema]]

Repositorio GitHub

avelikiy/great_cto
Ruta: skills/vertical-retail
0
agentic-codingclaude-code-pluginclaude-code-skillsclaude-code-subagentscode-reviewcto
FAQ

Frequently asked questions

What is the vertical-retail skill?

vertical-retail is a Claude Skill by avelikiy. Skills package instructions and resources that Claude loads on demand, so Claude can perform vertical-retail-related tasks without extra prompting.

How do I install vertical-retail?

Use the install commands on this page: add vertical-retail to Claude Code as a plugin, or clone its repository into your skills directory, then restart Claude so it picks up the skill.

What category does vertical-retail belong to?

vertical-retail is in the Meta category, tagged ai and design.

Is vertical-retail free to use?

Yes. vertical-retail is listed on AIMCP and free to install. It runs inside Claude, so no separate service account is required to use the skill itself.

Habilidades relacionadas

content-collections
Meta

Esta habilidad proporciona una configuración probada en producción para Content Collections, una herramienta centrada en TypeScript que transforma archivos Markdown/MDX en colecciones de datos con tipado seguro mediante validación Zod. Úsala al construir blogs, sitios de documentación o aplicaciones Vite + React con mucho contenido para garantizar seguridad de tipos y validación automática de contenido. Abarca todo, desde la configuración del plugin de Vite y compilación MDX hasta la optimización de despliegue y validación de esquemas.

Ver habilidad
polymarket
Meta

Esta habilidad permite a los desarrolladores crear aplicaciones con la plataforma de mercados de predicción Polymarket, incluyendo la integración de API para operaciones y datos de mercado. También proporciona transmisión de datos en tiempo real a través de WebSocket para monitorear operaciones en vivo y actividad del mercado. Úsela para implementar estrategias de trading o crear herramientas que procesen actualizaciones de mercado en tiempo real.

Ver habilidad
creating-opencode-plugins
Meta

Esta habilidad ayuda a los desarrolladores a crear complementos de OpenCode que se conectan a más de 25 tipos de eventos, como comandos, archivos y operaciones LSP. Proporciona la estructura del complemento, las especificaciones de la API de eventos y los patrones de implementación para módulos en JavaScript/TypeScript. Úsala cuando necesites interceptar, monitorear o extender el ciclo de vida del asistente de IA de OpenCode con lógica personalizada basada en eventos.

Ver habilidad
sglang
Meta

SGLang es un framework de alto rendimiento para el servicio de LLM que se especializa en generación rápida y estructurada para JSON, expresiones regulares y flujos de trabajo de agentes utilizando su caché de prefijos RadixAttention. Ofrece una inferencia significativamente más rápida, especialmente para tareas con prefijos repetidos, lo que lo hace ideal para salidas complejas y estructuradas, y conversaciones multiturno. Elige SGLang sobre alternativas como vLLM cuando necesites decodificación restringida o estés construyendo aplicaciones con uso extensivo de prefijos compartidos.

Ver habilidad