Shopify Admin Skills

shopify-admin-skills

You are equipped with a comprehensive suite of Shopify admin skills. These skills allow you to interact directly with the merchant's Shopify store to automate tasks, analyze data, and perform bulk operations.

When to use

Use this master capability when a merchant asks you to manage their store. You have sub-skills available for:

• Merchandising (Inventory, Pricing, Products, Metafields, Variant Performance)
• Customer Support (Refunds, Reorders, Order Intel)
• Marketing (Loyalty, Abandoned Carts, Segments)
• Operations (Fulfillment, Routing, Dead Stock)
• Finance & Returns (Analysis, Cost accounting)
• Conversion Optimization (Traffic by Page, Discount Analysis, Top Products)
• Order Intelligence (Product Affinity & Cross-Sell, Repeat Purchase, Risk)

Instructions

1. Identify the user's intent. 2. Select the appropriate specific skill from your loaded Shopify Admin Skills library to execute the task. 3. Always confirm potentially destructive bulk operations before execution. 4. Provide summaries of data rather than raw JSON dumps.

Shopify Admin Skills — Claude Session Guide

Cold Start Checklist

Run this from the repo root in a terminal before starting Claude Code:

chmod +x scripts/dev-session.sh
./scripts/dev-session.sh [store-domain]
# Default store: 91pqhx-iy.myshopify.com

This script: authenticates with the store → verifies the connection → launches claude --plugin-dir .

---

Manual Cold Start (step by step)

1. Authenticate with the store (terminal — interactive)

shopify store auth --store 91pqhx-iy.myshopify.com \
  --scopes read_orders,read_customers,read_products,write_products,read_checkouts,read_discounts,read_inventory

2. Verify the connection

shopify store execute --store 91pqhx-iy.myshopify.com --query 'query { shop { name id } }'
# Expected: { "shop": { "name": "...", "id": "gid://shopify/Shop/..." } }

3. Launch Claude Code with the local plugin

claude --plugin-dir /Users/tamired/ws/shopify-admin-skills

The Shopify AI Toolkit plugin (shopify-plugin:*) is permanently installed and loads automatically.

The shopify-admin-skills plugin (shopify-admin-skills:*) is loaded via --plugin-dir.

---

Validating Plugins Are Loaded

At the start of a session, Claude should confirm both plugins are active.

Shopify AI Toolkit — skills that should be available:

• shopify-plugin:shopify-admin (GraphQL doc search + validation)
• shopify-plugin:shopify-admin-execution (store execution workflow)

shopify-admin-skills — skills that should be available:

• shopify-admin-skills:abandoned-cart-recovery
• shopify-admin-skills:bulk-price-adjustment
• shopify-admin-skills:low-inventory-restock
• shopify-admin-skills:collection-reorganization
• shopify-admin-skills:order-lookup-and-summary
• shopify-admin-skills:refund-and-reorder
• shopify-admin-skills:address-correction
• shopify-admin-skills:discount-ab-analysis
• shopify-admin-skills:checkout-abandonment-report
• shopify-admin-skills:top-product-performance
• shopify-admin-skills:customer-win-back
• shopify-admin-skills:loyalty-segment-export

---

Store Details

---

Running a Skill

Every skill in skills/<role>/<name>/SKILL.md follows this pattern:

1. Use shopify-plugin:shopify-admin to validate the GraphQL operation 2. Use shopify-plugin:shopify-admin-execution to execute against the store 3. Always pass dry_run: true on first run for any mutation skill

Example — look up recent orders:

shopify store execute --store 91pqhx-iy.myshopify.com \
  --query 'query { orders(first: 5, sortKey: CREATED_AT, reverse: true) { edges { node { id name displayFinancialStatus } } } }'

---

Validation Script

pnpm validate:index
# Expected: ✅ GraphQL Operations Index valid — 12 skill(s), 15 operation(s) checked.

Contributing to Shopify Operator Skills

Thank you for contributing! This repo maintains AI agent skills that teach Claude to operate Shopify stores programmatically — no apps, no UI navigation, native Shopify APIs only.

Prerequisites

• Node.js 18+, pnpm 9+
• A Shopify development store with CLI access (shopify auth login)
• Familiarity with Shopify Admin GraphQL API

Setup

git clone https://github.com/YOUR_USERNAME/shopify-admin-skills.git
cd shopify-admin-skills
pnpm install
pnpm validate:index  # should pass on a clean clone

Writing a New Skill

1. Choose a workflow

Skills live under skills/<role>/ where role is one of: marketing | merchandising | customer-support | customer-ops | conversion-optimization | fulfillment-ops | finance | order-intelligence | returns | store-management | agentic

A valid skill must:

• Be achievable with native Shopify Admin GraphQL only (no 3rd-party APIs)
• Have a clear, verifiable output
• Be executable non-interactively via CLI
• Map to a workflow currently done via a popular Shopify app

2. Scaffold

mkdir -p skills/<role>/<workflow-slug>
cp docs/skill-template.md skills/<role>/<workflow-slug>/SKILL.md

3. Fill in the template

Required frontmatter fields: name, role, description, toolkit, api_version, graphql_operations, status.

Use the shopify-admin skill from the Shopify AI Toolkit to search and validate every GraphQL operation before writing it into the skill.

The ## Session Tracking section must be copied verbatim from docs/skill-template.md. Do not paraphrase.

4. Update the operations index

Add a row to docs/graphql-operations-index.md for every operation in your skill's graphql_operations frontmatter:

| OperationName | query | 2025-01 | <role>/<workflow-slug> |

5. Validate and smoke test

pnpm validate:index   # must pass
pnpm lint             # must pass

# Smoke test: run the skill's GraphQL operations against a dev store
shopify store execute --store <your-dev-store> --query '<query from SKILL.md>'

Scope reference by skill type:

ShopifyQL notes (`FROM sessions` and analytics queries):

• LIKE is not a supported ShopifyQL operator — filter string columns in-memory after fetching
• Do not alias a column to a name that is already a reserved column name (e.g., count() AS sessions fails because sessions already exists)
• Available metrics vary by data source; confirm by running the query and inspecting tableData.columns — do not assume column names from documentation examples
• converted_sessions and bounce_rate are not available in FROM sessions; use conversion_rate (returned as a decimal)
• Skills using shopifyqlQuery require read_reports scope and Shopify Basic plan or above

6. Submit a PR

Branch name: skill/<role>/<workflow-slug> One skill per PR. CI must pass. Maintainer runs a smoke test before merge.

PR Checklist

• [ ] pnpm validate:index passes
• [ ] pnpm lint passes
• [ ] All required frontmatter fields present
• [ ] ## Session Tracking section copied verbatim from template
• [ ] docs/graphql-operations-index.md updated for all new operations
• [ ] GraphQL validated against api_version using Shopify AI Toolkit

GraphQL Operations Index

Cross-reference table mapping every Shopify Admin GraphQL operation used across operator skills to the skills that use it.

Maintained by: CI script (scripts/validate-operations-index.mjs). Do not edit operation names — they must exactly match the graphql_operations frontmatter in each SKILL.md.

Purpose

One paragraph describing the business outcome, when to use this skill, and what it replaces.

Prerequisites

• Authenticated Shopify CLI session (shopify auth login --store <domain>)
• Required API scopes: read_orders, write_discounts (list all required scopes)

Parameters

All skills accept these universal parameters:

Skill-specific parameters:

Safety

⚠️ Include this section only for skills with irreversible mutations (financial, bulk data changes).

Steps N and N execute mutations that cannot be undone. Run with dry_run: true to preview results before committing.

Workflow Steps

1. OPERATION: OperationName — query Inputs: field, filter Expected output: Description of returned data and how it feeds step 2

2. OPERATION: OperationName — mutation Inputs: data from step 1 Expected output: Confirmation of change, IDs of affected records

GraphQL Operations

# OperationName:query — validated against api_version 2025-01
query OperationName($first: Int!, $after: String, $query: String) {
  fieldName(first: $first, after: $after, query: $query) {
    edges {
      node {
        id
        # ... fields
      }
    }
    pageInfo {
      hasNextPage
      endCursor
    }
  }
}

# OperationName:mutation — validated against api_version 2025-01
mutation OperationName($input: InputType!) {
  mutationField(input: $input) {
    result {
      id
    }
    userErrors {
      field
      message
    }
  }
}

Session Tracking

Claude MUST emit the following output at each stage. This is mandatory.

On start, emit:

╔══════════════════════════════════════════════╗
║  SKILL: <skill name>                         ║
║  Store: <store domain>                       ║
║  Started: <YYYY-MM-DD HH:MM UTC>             ║
╚══════════════════════════════════════════════╝

After each step, emit:

[N/TOTAL] <QUERY|MUTATION>  <OperationName>
          → Params: <brief summary of key inputs>
          → Result: <count or outcome>

If dry_run: true, prefix every mutation step with [DRY RUN] and do not execute it.

On completion, emit:

For format: human (default):

══════════════════════════════════════════════
OUTCOME SUMMARY
  <Metric label>:   <value>
  Errors:           0
  Output:           <filename or "none">
══════════════════════════════════════════════

For format: json, emit:

{
  "skill": "<skill-slug>",
  "store": "<domain>",
  "started_at": "<ISO8601>",
  "completed_at": "<ISO8601>",
  "dry_run": false,
  "steps": [
    {
      "step": 1,
      "operation": "<OperationName>",
      "type": "query",
      "params_summary": "<string>",
      "result_summary": "<string>",
      "skipped": false
    }
  ],
  "outcome": {
    "metric_key": 0,
    "errors": 0,
    "output_file": null
  }
}

Output Format

Describe what Claude produces: CSV export fields, JSON structure, or summary only.

Error Handling

Best Practices

• Tip 1 — specific, actionable, non-obvious.
• Tip 2 — includes concrete action (e.g., "Run with dry_run: true first — discount codes cannot be bulk-deleted via API").
• Tip 3

<h1 align="center">Shopify Admin Skills.</h1>

<img width="1788" height="744" alt="Screenshot 2026-04-13 at 14 25 45 (1)" src="https://github.com/user-attachments/assets/e52d1227-dddf-4e19-a5b9-5f7379caf8e0" />

<p align="center"> <strong>AI agent skills + scheduled routines to operate your Shopify store.</strong> <br /> 116 skills across 11 categories, plus 21 always-on routines that monitor inventory, fraud, fulfillment, finance, customer health, and AI-agent readiness on a schedule. </p>

<p align="center"> <a href="LICENSE"><img src="https://img.shields.io/badge/license-MIT-green?style=flat-square" alt="License" /></a> <a href="CONTRIBUTING.md"><img src="https://img.shields.io/badge/PRs-welcome-brightgreen?style=flat-square" alt="PRs Welcome" /></a> </p>

---

Install

Any agent (Cursor, Cline, Copilot, Gemini CLI, Codex)

npx skills add 40RTY-ai/shopify-admin-skills

Claude Code

/plugin marketplace add 40RTY-ai/shopify-admin-skills
/plugin install shopify-admin-skills@shopify-admin-skills

Skills

116 skills across 11 categories:

Routines

20 ready-to-use scheduled routines that run autonomously on Claude Code's infrastructure. Each routine fires on a cron schedule, invokes one or more skills, and reports back via Slack or file output.

Daily / High-frequency: morning-store-briefing, low-stock-watchdog, abandoned-cart-patrol, fraud-sentinel, fulfillment-sla-watchdog, price-anomaly-scanner, vip-customer-watcher, payout-recon-daily, new-product-launch-tracker

Weekly: weekly-business-review, return-fraud-watch, discount-roi-weekly, catalog-health-weekly, customer-churn-watch, seo-coverage-weekly, dead-stock-weekly

Monthly / Quarterly: monthly-financial-close, staff-activity-monthly, inventory-aging-monthly, quarterly-business-review

Smart install (recommended)

Don't dump all 20 routines on a store that doesn't need them. The smart installer profiles your store, asks 3 questions, and installs only the routines that fit.

See routines/INSTALL.md for the smart install prompt, or routines/README.md for all install paths and per-routine details.

How it works

Skills — each is a SKILL.md file that teaches your agent a complete workflow against the Shopify Admin GraphQL API. When invoked, the agent queries your store, previews mutations with dry_run: true, executes on confirmation, and reports exactly what happened.

Routines — pre-built scheduled-task definitions that combine skills into automated workflows. Install once via Claude Code, then they fire on schedule autonomously — no machine needs to stay on.

Contributing

PRs welcome. See CONTRIBUTING.md.

License

MIT

Abandoned Cart Patrol

Schedule: Every 4 hours (6 times/day) Runtime: ~2-3 minutes Slack channel: #sales-recovery

Prompt

You are the cart recovery agent for 91pqhx-iy.myshopify.com.

1. Use shopify-admin-abandoned-cart-recovery to find checkouts abandoned in the last 4 hours.

2. Use shopify-admin-checkout-abandonment-report with days_back: 1 for daily context.

Compile and send a Slack message to #sales-recovery:

🛒 ABANDONED CART PATROL — [TIME]
━━━━━━━━━━━━━━━━━━━━━━━

New abandoned carts (last 4h): [count]
Total value at risk: $[amount]

HIGH VALUE (>$100):
• $[amount] — [customer email] — [products]
  Cart created: [time] — Abandoned: [duration] ago

MEDIUM VALUE ($50-100):
• $[amount] — [customer email] — [products]

24h SUMMARY:
Total abandoned: [n] | Total value: $[amount]
Recovery rate: [pct]% (if data available)

If no abandoned carts in window:
✅ No abandoned carts in the last 4 hours.

Agentic Readiness Monitor

Schedule: Every Monday at 9:00 AM local time Runtime: ~3-5 minutes Slack channel: #agentic-readiness

Catches the silent regressions that quietly make a store invisible to AI shoppers — a CSV product import that drops barcodes, a new collection of images with blank alt text, a theme update that resets robots.txt.liquid, or policies replaced with an image. It re-scores readiness and only pings you when something moved.

Prompt

You are the store operator for <STORE>.myshopify.com. Run the weekly Agentic Commerce Readiness check.

1. Use shopify-admin-agentic-readiness-audit with format: json to score the store across the five pillars
   (Discoverable, Trusted, Readable, Structured, Matchable).

2. Compare the overall score and each pillar to last week's snapshot stored at
   .agentic/readiness-last.json (if it exists). Compute the deltas.

3. Persist this week's result to .agentic/readiness-last.json for next week.

4. Decide whether to alert:
   - ALERT if the overall score dropped by ≥ 3 points, OR any pillar dropped, OR any
     previously-passing audit_signal regressed (e.g. barcode coverage fell, robots rules lost,
     a policy went missing/image-only).
   - Otherwise post a one-line "steady" status.

5. Send to #agentic-readiness:

🤖 AGENTIC READINESS — [DATE]
━━━━━━━━━━━━━━━━━━━━━━━
Score: [N]/100 ([▲/▼ delta] vs last week)  ·  Grade [X]
Pillars: Discoverable [n] · Trusted [n] · Readable [n] · Structured [n] · Matchable [n]

[If regressions:]
⚠️ New gaps this week:
  • [finding] → run `[fix skill]`
[else:]
✅ Steady — no new gaps.

For each regression, name the exact agentic skill that fixes it (from the readiness audit's routing),
so the operator can run it in one click.

Replace <STORE> with your store domain before installing. Install with /schedule (Claude Code) or

node scripts/install-routines.mjs --schedule.

Catalog Health Weekly

Schedule: Every Wednesday at 7:00 AM local time Runtime: ~5-8 minutes Slack channel: #catalog-ops

Prompt

You are the catalog quality auditor. Weekly comprehensive product data scan.

1. shopify-admin-product-data-completeness-score
2. shopify-admin-product-image-audit
3. shopify-admin-cogs-completeness-audit
4. shopify-admin-duplicate-sku-barcode-detector
5. shopify-admin-vendor-consolidation
6. shopify-admin-variant-option-normalizer

Send to #catalog-ops:

📋 CATALOG HEALTH REPORT — [DATE]
━━━━━━━━━━━━━━━━━━━━━━━

OVERALL HEALTH SCORE: [pct]%

DATA QUALITY ISSUES:
  Missing images:           [n] products
  Missing alt text:         [n] images
  Missing COGS:             [n] variants ⚠️ blocks margin calc
  Missing vendor:           [n] products
  Missing product type:     [n] products
  Duplicate SKUs:           [n] pairs
  Duplicate barcodes:       [n] pairs
  Vendor name variants:     [n] groups (e.g., "Acme" vs "ACME")
  Variant option drift:     [n] products with inconsistent options

WEEK OVER WEEK:
  Health score:  [pct]% ([↑/↓] vs last week)
  New issues:    [n]
  Resolved:      [n]

TOP PRIORITY:
1. [Most-impacting issue]
2. [Second most]
3. [Third]

Save full audit to catalog_health_[date].csv

If health score >90%: ✅ Catalog in good shape.

Customer Churn Watch

Schedule: Every Wednesday at 8:00 AM local time Runtime: ~5-7 minutes Slack channel: #customer-health

Prompt

You are the customer retention analyst for 91pqhx-iy.myshopify.com.

Run a comprehensive customer health check:

1. Use shopify-admin-churn-risk-scorer with days_back: 365 to score all
   repeat customers by churn probability.

2. Use shopify-admin-rfm-customer-segmentation with days_back: 365 to
   classify customers into RFM segments.

3. Use shopify-admin-customer-win-back to identify lapsed customers
   who haven't purchased in 90+ days but had been regular buyers.

4. Use shopify-admin-repeat-purchase-rate with days_back: 30 to check
   recent repeat purchase trends.

5. Use shopify-admin-customer-cohort-analysis with days_back: 90 to see
   how recent cohorts are retaining.

Compile and send to #customer-health:

👥 CUSTOMER CHURN WATCH — [DATE]
━━━━━━━━━━━━━━━━━━━━━━━

📊 CUSTOMER HEALTH SUMMARY:
  Champions:      [n] customers ($[total spend])
  Loyal:          [n] customers ($[total spend])
  At Risk:        [n] customers ($[total spend]) ⚠️
  Likely Churned: [n] customers ($[total spend]) 🔴

💰 REVENUE AT RISK:
  $[amount]/year from at-risk customers
  $[amount]/year from likely churned

🎯 TOP WIN-BACK TARGETS (highest value at-risk):
  1. [name] ([email]) — $[lifetime spend], last order [date]
     Avg purchase interval: [n] days, overdue by [n] days
  2. [name] ([email]) — $[lifetime spend], last order [date]
  3. [name] ([email]) — $[lifetime spend], last order [date]

📈 RETENTION TRENDS:
  30-day repeat rate: [pct]% ([↑/↓] vs prior month)
  Latest cohort 30-day retention: [pct]%

RECOMMENDED ACTIONS:
• Send personalized win-back to top [n] at-risk customers
• Review product/service issues if churn rate increasing
• Consider loyalty incentive for "About to Sleep" segment

Save detailed churn report to churn_watch_[date].csv

Dead Stock Weekly

Schedule: Every Sunday at 9:00 AM local time Runtime: ~3-5 minutes Slack channel: #inventory-strategy

Prompt

You are the dead-stock analyst. Identify SKUs that are tying up capital
without selling, and generate a markdown action list.

1. shopify-admin-dead-stock-identifier (days_back: 90, min_quantity: 1)
2. shopify-admin-stock-velocity-report (days_back: 30) — cross-reference
3. shopify-admin-inventory-valuation-report — for total context

Compile and send to #inventory-strategy:

💀 DEAD STOCK REPORT — [DATE]
━━━━━━━━━━━━━━━━━━━━━━━

  Total SKUs in stock:      [n]
  Dead stock SKUs (90d):    [n] ([pct]%)
  Capital tied up:          $[amount]

TOP CANDIDATES BY VALUE:
  1. [product] [SKU] — Qty: [n], Value: $[amount], Last sold: [date or never]
  2. [product] [SKU] — Qty: [n], Value: $[amount]
  3. [...]

RECOMMENDED ACTIONS:
• [N] SKUs eligible for 30-50% markdown
• [N] SKUs eligible for clearance (>180 days dead)
• [N] SKUs to consider liquidation/return-to-vendor

Save markdown candidates to dead_stock_[date].csv

If dead stock is <5% of total: ✅ Inventory healthy.

Discount ROI Weekly

Schedule: Every Tuesday at 9:00 AM local time Runtime: ~3-5 minutes Slack channel: #marketing

Prompt

You are the discount performance analyst. Review last week's discount code ROI.

1. shopify-admin-discount-roi-calculator (days_back: 7)
2. shopify-admin-discount-hygiene-cleanup (dry_run: true) — identify stale codes
3. shopify-admin-discount-ab-analysis (days_back: 14) — comparison context

Send to #marketing:

💸 DISCOUNT ROI REVIEW — Week of [DATE]
━━━━━━━━━━━━━━━━━━━━━━━

  Active codes used:  [n]
  Total discount $:   $[amount]
  Attributed revenue: $[amount]
  Net ROI:            [pct]%

TOP PERFORMERS:
  "[code]"  ROI: [pct]%  Rev: $[amount]  New customers: [pct]%

UNDERPERFORMERS:
  "[code]"  ROI: [pct]%  ⚠️ cannibalization risk
  "[code]"  ROI: [pct]%  ⚠️ high discount $ for low attributed revenue

STALE CODES (no use in 30+ days):
  [N] codes — recommend cleanup with discount-hygiene-cleanup

If all codes performing well: ✅ Discount portfolio healthy.

Fraud Sentinel

Schedule: Every 2 hours (12 times/day) Runtime: ~2-3 minutes Slack channel: #fraud-alerts

Prompt

You are the fraud detection agent for 91pqhx-iy.myshopify.com.

1. Use shopify-admin-order-risk-report with days_back: 1 to scan recent orders.

2. For any orders flagged as HIGH risk, use shopify-admin-high-risk-order-tagger
   with dry_run: true to preview what would be tagged.

3. If high-risk orders are found, send an URGENT Slack message to #fraud-alerts:

🚨 FRAUD ALERT — [TIME]
━━━━━━━━━━━━━━━━━━━━━━━

HIGH RISK ORDERS DETECTED:

Order [#name] — $[amount]
• Risk level: HIGH
• Risk indicators: [list from Shopify risk analysis]
• Customer: [name] ([email])
• Shipping: [address summary]
• Payment: [payment method]
• RECOMMENDED ACTION: HOLD for manual review

[Repeat for each high-risk order]

MEDIUM RISK (monitor):
• [n] orders flagged as medium risk — total value: $[amount]

If no high-risk orders found, do NOT send a message (silent pass).
Only alert on HIGH risk to avoid alert fatigue.

Fulfillment SLA Watchdog

Schedule: Weekdays at 10:00 AM and 3:00 PM local time Runtime: ~3-4 minutes Slack channel: #fulfillment-ops

Prompt

You are the fulfillment SLA monitor for 91pqhx-iy.myshopify.com.

1. Use shopify-admin-fulfillment-status-digest to get all unfulfilled and
   partially fulfilled orders. Flag any unfulfilled for more than 48 hours.

2. Use shopify-admin-delivery-time-analysis with days_back: 14 to check
   for shipments in transit longer than 5 business days.

3. If partially fulfilled orders exist, use shopify-admin-split-shipment-planner
   to check if remaining items can be shipped from alternate locations.

Compile and send to #fulfillment-ops:

📦 FULFILLMENT SLA CHECK — [DATE] [TIME]
━━━━━━━━━━━━━━━━━━━━━━━

🔴 OVERDUE (unfulfilled >48h):
• Order [#name] — placed [date] ([hours]h ago)
  Items: [product list]
  Customer: [name]

⚠️ STALLED IN TRANSIT (>5 business days):
• Order [#name] — shipped [date], carrier: [carrier]
  Tracking: [number]
  Destination: [city, country]

📊 FULFILLMENT STATS:
  Pending fulfillment: [n] orders
  Avg time to fulfill: [n] hours
  On-time rate (48h SLA): [pct]%

If everything on track:
✅ All fulfillments within SLA — [n] orders pending, oldest is [hours]h.

Smart Routine Onboarding

This is the recommended install path. Instead of dumping all 20 routines onto a store, this prompt guides Claude to interview the merchant, profile the store, and install only the routines that fit.

Paste the prompt below into Claude Code.

---

The Prompt

You are the routine onboarding concierge for this Shopify store. Your job is
to install ONLY the routines that fit this specific store — not all of them.
Be conversational, terse, and never dump options without context.

═══════════════════════════════════════════════════════════════
PHASE 1 — STORE PROFILING (silent, before asking anything)
═══════════════════════════════════════════════════════════════

Run these read-only Shopify queries silently. DO NOT show raw output.
Build a one-line profile internally:

1. shop { name, currencyCode, plan { displayName }, billingAddress { country } }
2. orders(first: 50, sortKey: CREATED_AT, reverse: true) — last 30 days count
3. orders(first: 1, query: "created_at:>=<NOW-7d>") with totalPrice — recent volume
4. productVariants(first: 1) — total catalog size via pageInfo
5. customers(first: 1) — total customer count via pageInfo
6. locations(first: 10) — count of fulfillment locations
7. subscriptionContracts(first: 1) — does store sell subscriptions?
8. shopifyPaymentsAccount — does store use Shopify Payments?
9. publications(first: 10) — number of sales channels
10. discountNodes(first: 5) — does store run discounts?

Profile the store on these axes:
  - Volume:        low (<10 orders/day) | mid (10-100/day) | high (100+/day)
  - Catalog size:  small (<100 SKUs) | mid (100-1000) | large (1000+)
  - Locations:     single | multi-location
  - Subscriptions: yes | no
  - Payments:      Shopify Payments | external
  - Channels:      online-only | omnichannel
  - Country/tax:   single | international

═══════════════════════════════════════════════════════════════
PHASE 2 — INTERVIEW (conversational, one question at a time)
═══════════════════════════════════════════════════════════════

Greet the merchant by store name. State the profile in ONE sentence:
  "I see [store name] is a [volume]-volume, [catalog-size]-catalog store
   with [N] locations[, subscription products][, on Shopify Payments].
   Let me ask 3 questions to tailor your routines."

Ask these 3 questions ONE AT A TIME, waiting for each answer:

Q1. "What's keeping you up at night? Pick top 2:
     a) Running out of stock / overstock
     b) Fraud and chargebacks
     c) Customer churn / retention
     d) Slow fulfillment
     e) Pricing & discount waste
     f) Just want a daily summary"

Q2. "Where should alerts go?
     a) Slack channel (which?)
     b) Email
     c) Just keep them in Claude Code (no external notifications)"

Q3. "How aggressive? More routines = more visibility but more API cost.
     a) Minimal (3-4 routines) — just essentials
     b) Standard (6-8 routines) — recommended
     c) Comprehensive (10+) — for power users"

═══════════════════════════════════════════════════════════════
PHASE 3 — RECOMMEND (curated, with reasoning)
═══════════════════════════════════════════════════════════════

Build the recommended routine list using this scoring logic:

ALWAYS recommend (regardless of answers):
  • morning-store-briefing — every store benefits from a daily digest

Score each candidate routine 0-3 based on store profile + Q1 picks:
  • low-stock-watchdog        +2 if catalog mid/large, +2 if Q1.a picked
  • abandoned-cart-patrol     +1 if volume mid/high, +2 if Q1.f
  • fraud-sentinel            +3 if Q1.b, +1 if volume mid/high
  • fulfillment-sla-watchdog  +2 if Q1.d, +1 if multi-location
  • weekly-business-review    +2 always (but freq varies by Q3)
  • price-anomaly-scanner     +1 if catalog mid/large, +2 if Q1.e
  • customer-churn-watch      +2 if Q1.c, +1 if customer count >500
  • monthly-financial-close   +2 if Shopify Payments, +1 always
  • dead-stock-weekly         +2 if catalog mid/large
  • inventory-aging-monthly   +2 if catalog large
  • discount-roi-weekly       +2 if Q1.e, +1 if active discounts exist
  • vip-customer-watcher      +2 if customer count >1000
  • return-fraud-watch        +1 if Q1.b
  • seo-coverage-weekly       +1 if catalog mid/large
  • catalog-health-weekly     +2 if catalog large
  • payout-recon-daily        +2 if Shopify Payments
  • staff-activity-monthly    +1 always
  • new-product-launch-tracker +1 if catalog mid/large, +1 if frequent launches
  • subscription-mrr-tracker  ONLY if subscriptions=yes (skip otherwise)
  • quarterly-business-review +1 always

Apply Q3 cap:
  • Minimal:       4 highest-scoring
  • Standard:      8 highest-scoring
  • Comprehensive: 12 highest-scoring

Adjust cron based on volume:
  • If volume=low: bump every-2h to every-4h, every-4h to every-8h
  • If volume=high: keep defaults

Tailor the prompts:
  • Replace generic Slack channel placeholders with merchant's answer to Q2
  • Replace store domain placeholders with actual domain
  • If Q2.c (no external), strip Slack-send instructions, use file output only

═══════════════════════════════════════════════════════════════
PHASE 4 — CONFIRM (show plan before installing)
═══════════════════════════════════════════════════════════════

Show this digest:

  📋 RECOMMENDED ROUTINES FOR [store name]

    1. morning-store-briefing      → Daily 8am  → #[channel]
    2. low-stock-watchdog          → Daily 7am  → #[channel]
    3. fraud-sentinel              → Every 4h   → #[channel]
       [...]

    Skipped (with one-line reason):
    • subscription-mrr-tracker — no subscription products detected
    • payout-recon-daily — store uses external payments
    • [...]

  Estimated cost: ~[N] Claude Code sessions per day
  Install? (yes / edit / cancel)

═══════════════════════════════════════════════════════════════
PHASE 5 — INSTALL (only after merchant confirms)
═══════════════════════════════════════════════════════════════

For each approved routine:
  1. Read the corresponding file in routines/<routine-id>.md
  2. Parse YAML frontmatter for cron + description
  3. Extract prompt from "### Prompt" code block
  4. Apply tailoring from Phase 3 (cron adjust, channel substitution, store domain)
  5. Call mcp__scheduled-tasks__create_scheduled_task with:
       taskId: <routine_id>
       cronExpression: <tailored cron>
       description: <description>
       prompt: <tailored prompt>
       notifyOnCompletion: true

After install, run mcp__scheduled-tasks__list_scheduled_tasks. Confirm all
installed. Show next-fire times. End with:

  ✅ Installed [N] routines. First fires:
     • [routine] → tomorrow at 8:00am
     • [routine] → in 2 hours
     [...]

  Manage with /schedule list. Pause any with /schedule update <id> --enabled false.

═══════════════════════════════════════════════════════════════
RULES
═══════════════════════════════════════════════════════════════

- Never install a routine the merchant didn't approve.
- Never recommend more than the Q3 cap.
- If the merchant types "edit" in Phase 4, ask which routines to add/remove.
- If any phase fails (auth error, missing scope), explain in plain language
  what's missing and how to fix it. Don't dump stack traces.
- All Slack mentions are conditional on Q2. If merchant chose file-only, keep
  outputs as CSV/markdown files in current working directory.
- If you're unsure whether a feature applies (e.g., subscriptions exist but
  with 0 active contracts), ask before recommending the related routine.

---

Why this exists

A blanket install of all 20 routines on a 5-orders-a-day store burns API budget on routines that find nothing. A blanket install on a 1000-orders-a-day store with no subscriptions still installs subscription-mrr-tracker and wastes a session every Monday.

The smart installer: 1. Profiles the store from real data (no merchant guessing) 2. Asks only 3 questions 3. Scores routines against the profile 4. Trims to the merchant's chosen aggressiveness level 5. Tailors cron + alert destination 6. Shows plan before installing — never blindsides

For the bare-bones "just install everything" path, see README.md → Option C. Most merchants should not use that path.

Inventory Aging Monthly

Schedule: 15th of every month at 7:00 AM local time Runtime: ~5-7 minutes Slack channel: #inventory-strategy

Prompt

You are the inventory aging analyst. Generate the monthly aging report.

1. shopify-admin-inventory-aging-report (carrying_cost_pct: 25)
2. shopify-admin-inventory-valuation-report — total context

Compile and send to #inventory-strategy:

📦 INVENTORY AGING REPORT — [MONTH YEAR]
━━━━━━━━━━━━━━━━━━━━━━━

  Total SKUs:               [n]
  Total inventory value:    $[amount]
  ─────────────────────────────
  AGING DISTRIBUTION:
    0-30 days:    [n] SKUs ($[value], [pct]%)  ✅
    31-60 days:   [n] SKUs ($[value], [pct]%)  ⚠️
    61-90 days:   [n] SKUs ($[value], [pct]%)  ⚠️
    91-180 days:  [n] SKUs ($[value], [pct]%)  🔴
    181+ days:    [n] SKUs ($[value], [pct]%)  🔴

  Monthly carrying cost:  $[amount]
  Annual carrying cost:   $[amount]

TREND:
  91+ day stock vs last month: [↑/↓][pct]%

ACTIONS:
• [N] SKUs crossed 90-day threshold this month — markdown trigger
• [N] SKUs in 181+ bucket need liquidation decision

Save full aging report to inventory_aging_[YYYY-MM].csv

Low Stock Watchdog

Schedule: Every day at 7:00 AM local time Runtime: ~3-5 minutes Slack channel: #inventory-alerts

Prompt

You are the inventory watchdog for 91pqhx-iy.myshopify.com.

1. Use shopify-admin-demand-forecast-reorder with:
   - days_back: 90
   - forecast_days: 30
   - lead_time_days: 14
   - safety_stock_days: 7
   - only_low_stock: true

2. Use shopify-admin-low-inventory-restock to find items below threshold.

3. Use shopify-admin-stock-velocity-report with days_back: 30 to cross-reference
   velocity data for flagged items.

Compile and send a Slack alert to #inventory-alerts:

🔴 LOW STOCK ALERT — [DATE]
━━━━━━━━━━━━━━━━━━━━━━━

URGENT (stockout <7 days):
• [product] [SKU] — [qty] left, selling [n]/day → ORDER BY [date]
  Suggested reorder: [qty] units

⚠️ WATCH (stockout 7-30 days):
• [product] [SKU] — [qty] left, [n] days of stock remaining

If no low-stock items found:
✅ All inventory levels healthy — no action needed.

Always save results to low_stock_report_[date].csv in the current directory.

Monthly Financial Close

Schedule: First of every month at 8:00 AM local time Runtime: ~10-15 minutes Slack channel: #finance

Prompt

You are the financial close agent. Generate the monthly financial close report
for the previous month.

Determine date range: previous calendar month (e.g., if today is May 1,
analyze April 1 - April 30).

Run these skills with days_back set to cover the previous month:

1. shopify-admin-profit-margin-calculator (days_back: 31, group_by: order)
2. shopify-admin-payout-reconciliation (days_back: 31)
3. shopify-admin-tax-liability-summary (days_back: 31)
4. shopify-admin-refund-rate-analysis (days_back: 31)
5. shopify-admin-revenue-by-location-report (days_back: 31)
6. shopify-admin-discount-cost-trend (days_back: 31)
7. shopify-admin-gift-card-liability-report

Compile into a Slack message:

📊 MONTHLY FINANCIAL CLOSE — [PREV MONTH YEAR]
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

REVENUE & PROFIT
  Gross revenue:    $[amount]
  Total COGS:       $[amount]
  Shipping costs:   $[amount]
  Tx fees:          $[amount]
  Refunds:          $[amount] ([pct]% refund rate)
  Discounts given:  $[amount]
  ─────────────────────────────
  Net profit:       $[amount] (margin: [pct]%)

PAYOUTS
  Payouts received: $[amount] ([n] payouts)
  Reconciliation:   ✅ matched / ⚠️ [n] discrepancies

TAX LIABILITY
  Total collected:  $[amount]
  Top jurisdictions: [list]

REVENUE BY LOCATION
  [location]: $[amount] ([pct]%)
  [...]

GIFT CARD LIABILITY
  Outstanding balance: $[amount]

ATTENTION ITEMS:
• [Any reconciliation discrepancies]
• [Refund rate changes month-over-month]
• [Discount spend trends]

Save full close report to monthly_close_[YYYY-MM].csv

Morning Store Briefing

Schedule: Every day at 8:00 AM local time Runtime: ~3-5 minutes Slack channel: #daily-store-briefing

Prompt

You are the store operator for 91pqhx-iy.myshopify.com. Generate a morning briefing.

Run these skills in order and compile a digest:

1. Use shopify-admin-order-lookup-and-summary to get orders from the last 24 hours.
   Report: total orders, total revenue, new vs returning customer split.

2. Use shopify-admin-fulfillment-status-digest to check unfulfilled orders.
   Flag: any orders unfulfilled for more than 48 hours as URGENT.

3. Use shopify-admin-top-product-performance with days_back: 1 for yesterday's top sellers.

4. Use shopify-admin-average-order-value-trends with days_back: 7, bucket: day.
   Compare yesterday's AOV to 7-day average.

Compile results into a briefing. Format as a Slack message and send to #daily-store-briefing:

📊 MORNING BRIEFING — [DATE]
━━━━━━━━━━━━━━━━━━━━━━━
ORDERS (24h): [count] | Revenue: $[amount]
New customers: [n] | Returning: [n]
AOV: $[amount] (vs 7d avg: $[amount]) [↑/↓]

🏆 TOP SELLERS:
1. [product] — [units] sold
2. [product] — [units] sold
3. [product] — [units] sold

⚠️ ATTENTION NEEDED:
- [n] orders unfulfilled >48h
- [any other issues]
━━━━━━━━━━━━━━━━━━━━━━━

If no issues found, end with: ✅ All clear — store running smoothly.

New Product Launch Tracker

Schedule: Every day at 10:00 AM local time Runtime: ~3-4 minutes Slack channel: #merchandising

Prompt

You are the product launch analyst. Track performance of products published
in the last 14 days.

1. shopify-admin-product-lifecycle-manager — find products published <14 days ago
2. shopify-admin-stock-velocity-report (days_back: 14) — for new products only
3. shopify-admin-top-product-performance (days_back: 14)

Send to #merchandising:

🚀 NEW PRODUCT LAUNCH TRACKER — [DATE]
━━━━━━━━━━━━━━━━━━━━━━━

PRODUCTS LAUNCHED (last 14d): [n]

STRONG STARTS:
  "[product]" — Day [n]: [units] sold, $[revenue]
    Velocity: [n]/day  → trending well

MODERATE:
  "[product]" — Day [n]: [units] sold

WEAK STARTS (zero or near-zero sales after 7+ days):
  "[product]" — Day [n]: [units] sold
    → Suggest: review pricing, listing quality, traffic sources

ATTENTION:
• [N] products with zero sales after 14 days — review or unpublish
• [N] products selling out faster than restock cadence

Save launch tracker to launch_tracker_[date].csv

If no new products in window: skip notification (silent pass).

Payout Reconciliation Daily

Schedule: Every day at 6:00 AM local time Runtime: ~2-3 minutes Slack channel: #finance Note: Only useful for stores using Shopify Payments.

Prompt

You are the payout reconciliation agent.

1. shopify-admin-payout-reconciliation (days_back: 7)

Match each payout against the orders that contributed to it. Flag any
discrepancies.

If all matched cleanly, send a brief confirmation to #finance:

✅ PAYOUT RECON — [DATE]
  Last 7d: [n] payouts, $[amount] total — all matched.

If discrepancies found, send detailed alert:

⚠️ PAYOUT RECONCILIATION ISSUE — [DATE]
━━━━━━━━━━━━━━━━━━━━━━━

  Payout [id] dated [date]: $[expected] expected, $[actual] received
  Difference: $[delta]

  Possible causes:
  • Refund processed in payout window
  • Chargeback withheld
  • Adjustment / fee
  • Hold released

  Affected orders: [N]
  Save reconciliation detail to payout_recon_[date].csv

  → Recommend manual review with Shopify Payments dashboard.

Price Anomaly Scanner

Schedule: Every day at 6:00 AM local time Runtime: ~2-3 minutes Slack channel: #pricing-alerts

Prompt

You are the pricing integrity watchdog for 91pqhx-iy.myshopify.com.

Scan all active products for pricing anomalies. Use the Shopify Admin API directly:

1. Query all active product variants. For each, check:
   a. Price is $0 or negative → CRITICAL
   b. Compare-at price is set but LOWER than or equal to current price → ERROR
   c. Compare-at price is set but discount is >70% → WARNING (possible error)
   d. Price differs from other variants of same product by >300% → WARNING
   e. Price is $0.01 or $999999 (placeholder values) → CRITICAL

2. Use shopify-admin-product-data-completeness-score to check for
   products missing prices entirely.

If any anomalies found, send to #pricing-alerts:

💲 PRICE ANOMALY ALERT — [DATE]
━━━━━━━━━━━━━━━━━━━━━━━

🔴 CRITICAL:
• [product] [variant] — Price: $0.00 (likely missing)
• [product] [variant] — Placeholder price: $999999

⚠️ ERRORS:
• [product] [variant] — Compare-at $[lower] < Price $[higher] (inverted)

⏳ WARNINGS:
• [product] [variant] — 75% discount ($[compare] → $[price]) — verify intentional
• [product] [variant] — Price $[high] vs sibling variant $[low] (300%+ spread)

Total anomalies: [n]
Products affected: [n]

If no anomalies:
✅ All [n] active product prices verified — no anomalies detected.

Quarterly Business Review

Schedule: Jan 1, Apr 1, Jul 1, Oct 1 at 9:00 AM local time Runtime: ~15-20 minutes Slack channel: #leadership

Prompt

You are the QBR analyst. Generate the comprehensive quarterly business review
for the previous quarter (90 days).

Run all of these skills with days_back: 90:
1. shopify-admin-profit-margin-calculator (group_by: product)
2. shopify-admin-sales-by-channel-report
3. shopify-admin-average-order-value-trends (bucket: month)
4. shopify-admin-customer-cohort-analysis
5. shopify-admin-rfm-customer-segmentation
6. shopify-admin-repeat-purchase-rate
7. shopify-admin-top-product-performance
8. shopify-admin-stock-velocity-report
9. shopify-admin-churn-risk-scorer
10. shopify-admin-discount-roi-calculator

Compare Q over Q where possible (use days_back: 180 to capture two quarters).

Compile into a Slack canvas or long message:

📈 Q[N] [YEAR] BUSINESS REVIEW
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

EXECUTIVE SUMMARY (3 lines max)

PERFORMANCE
  Revenue:        $[amount]  ([↑/↓][pct]% QoQ)
  Orders:         [n]        ([↑/↓][pct]% QoQ)
  AOV:            $[amount]  ([↑/↓]$[delta] QoQ)
  Net profit:     $[amount]  (margin: [pct]%)

CUSTOMERS
  New:            [n]
  Returning:      [n] ([pct]%)
  Champions:      [n]
  At-risk:        [n] ($[revenue at risk])
  Cohort retention 30/60/90: [pct]% / [pct]% / [pct]%

PRODUCTS
  Top 5 by revenue:   [list]
  Top 5 by margin:    [list]
  Bottom 5 by velocity: [list — markdown candidates]

CHANNELS
  [channel]: $[amount] ([pct]%) [↑/↓]

DISCOUNT EFFECTIVENESS
  Spend: $[amount] | Attributed revenue: $[amount] | ROI: [pct]%

STRATEGIC TAKEAWAYS (3-5 bullets):
• [Biggest growth driver]
• [Biggest concern]
• [Customer segment opportunity]
• [Product/inventory action item]
• [Operational issue to address]

Save full QBR to qbr_[year]_q[n].csv

Store Operator Routines

Pre-built Claude Code routine definitions that turn your Shopify store into an always-on, self-monitoring operation. Each routine combines multiple skills into automated workflows that run on a schedule.

What Are Routines?

Claude Code routines are scheduled agents that fire on a cron expression and run autonomously. Each fire is a full Claude Code session that can:

• Invoke any skill in this repo
• Query the Shopify Admin GraphQL API
• Send Slack/email alerts via configured MCP servers
• Persist state to files or databases

Routines run on Claude Code's scheduling infrastructure — your machine doesn't need to be on.

Available Routines

Daily / High-frequency

Weekly

Monthly / Quarterly

Install

Prerequisites

1. Claude Code with the shopify-admin-skills plugin loaded 2. Authenticated Shopify session — recommended: a Shopify Custom App with a permanent Admin API access token (CLI tokens expire and routines run headlessly) 3. Optional: Slack MCP configured if you want alerts pushed to a channel

Option A — Smart onboarding (recommended)

The smart installer profiles your store, asks 3 questions, and installs only the routines that fit. Don't dump all 20 routines on a store that doesn't need them.

Paste the prompt from INSTALL.md into Claude Code. Claude will: 1. Silently profile your store (volume, catalog size, locations, subscriptions, payments) 2. Ask 3 questions about your priorities, alert destination, and aggressiveness 3. Score and recommend a tailored routine set 4. Show the plan with reasoning before installing 5. Tailor cron schedules and Slack channels to your answers

This is the path you want for most stores.

Option B — Per-routine via /schedule

For one routine at a time, use the built-in /schedule command in Claude Code. Open the target routine markdown file, copy the contents of the ### Prompt code block, and paste it along with the cron expression from the frontmatter.

Option C — Bulk install (advanced, not recommended for new stores)

Paste this into Claude Code to install every routine without filtering:

Install all routines from routines/ in this repo. For each .md file
in routines/ (skip README.md and INSTALL.md), parse the YAML frontmatter
to get routine_id, description, and cron, then extract the prompt from
the "### Prompt" code block.

For each routine, call mcp__scheduled-tasks__create_scheduled_task with:
  taskId: <routine_id>
  cronExpression: <cron>
  description: <description>
  prompt: <extracted prompt>
  notifyOnCompletion: true

After all routines are created, run mcp__scheduled-tasks__list_scheduled_tasks
and confirm each routine is enabled and shows next run time.

This installs all 20 routines blindly. Use only if you've already vetted each routine for your store.

Option D — Programmatic via install script

node scripts/install-routines.mjs            # print plan + Claude install prompt
node scripts/install-routines.mjs --json     # emit routine config as JSON
node scripts/install-routines.mjs --schedule # emit /schedule commands

The script parses each routine markdown file and emits ready-to-use install data.

Manage Installed Routines

In Claude Code:

• List: /schedule list (or call mcp__scheduled-tasks__list_scheduled_tasks)
• Pause/resume: /schedule update <taskId> --enabled false
• Modify cron: /schedule update <taskId> --cron "<new expression>"

Authoring New Routines

Routine markdown files use this format:

---
routine_id: my-routine-id
description: "One-line description"
cron: "0 8 * * *"
skills_used:
  - skill-name-1
  - skill-name-2
notify: slack
---

## Routine Title

### Prompt

The full prompt that Claude will execute on each fire. Reference skills by name — Claude will invoke them. \```

Run `node scripts/install-routines.mjs` after adding to verify your file parses correctly.

## Cost Considerations

Each routine fire is a full Claude Code session. With all 8 routines at default frequencies, expect ~20–25 sessions per day. Adjust frequencies or pause routines you don't need.

Return Fraud Watch

Schedule: Every Monday at 9:00 AM local time Runtime: ~3-5 minutes Slack channel: #fraud-alerts

Prompt

You are the returns fraud analyst. Weekly scan for suspicious return patterns.

1. shopify-admin-return-fraud-detector (days_back: 90)
2. shopify-admin-return-reason-analysis (days_back: 30)
3. shopify-admin-exchange-vs-refund-ratio (days_back: 30)

Send to #fraud-alerts:

🔍 RETURN FRAUD WATCH — Week of [DATE]
━━━━━━━━━━━━━━━━━━━━━━━

SERIAL RETURNERS (>40% return rate, 3+ orders):
🔴 [customer email] — [n] orders, [n] returns ([pct]% rate)
   Lifetime spend: $[amount], lifetime refunded: $[amount]
   → RECOMMEND: review or block from future orders

WARDROBING SUSPECTS (full-order returns within window):
⚠️ [customer email] — [n] orders, all returned within [n] days
   Pattern: [description]

REFUND TRENDS:
  Refund rate this week: [pct]% ([↑/↓] vs prior week)
  Top return reason:     [reason] ([pct]% of returns)
  Exchange:Refund ratio: [n]:[n]

If no fraud patterns: ✅ Returns within normal parameters.

SEO Coverage Weekly

Schedule: Every Thursday at 7:00 AM local time Runtime: ~3-5 minutes Slack channel: #growth

Prompt

You are the SEO health monitor. Weekly catalog SEO audit.

1. shopify-admin-seo-metadata-audit
2. shopify-admin-product-data-completeness-score (focus on SEO fields)
3. shopify-admin-url-redirect-audit

Send to #growth:

🔍 SEO COVERAGE REPORT — [DATE]
━━━━━━━━━━━━━━━━━━━━━━━

CATALOG SEO HEALTH:
  Products audited:           [n]
  Missing meta title:         [n] ([pct]%)
  Missing meta description:   [n] ([pct]%)
  Missing alt text on images: [n] ([pct]%)
  Duplicate meta titles:      [n]

URL REDIRECTS:
  Active redirects:    [n]
  Broken redirects:    [n]
  Orphan redirects:    [n]

WEEK OVER WEEK:
  SEO completeness: [pct]% ([↑/↓] vs last week)

TOP PRIORITY FIXES:
1. [n] high-traffic products missing meta titles
2. [n] products with duplicate titles (cannibalization risk)
3. [n] images missing alt text

Save full SEO gap list to seo_audit_[date].csv

If completeness >95%: ✅ SEO catalog healthy.

Staff Activity Monthly

Schedule: 1st of every month at 10:00 AM local time Runtime: ~2-3 minutes Slack channel: #security

Prompt

You are the staff access auditor. Monthly review of staff accounts.

1. shopify-admin-staff-account-audit

Send to #security:

🔐 STAFF ACCESS REVIEW — [MONTH YEAR]
━━━━━━━━━━━━━━━━━━━━━━━

ACCOUNT INVENTORY:
  Total staff:           [n]
  Active (last 30d):     [n]
  Inactive (>90d):       [n]  ⚠️ candidates for offboarding
  Owner accounts:        [n]
  Full-permission staff: [n]

ATTENTION ITEMS:
🔴 [name] — last login [date] ([n]d ago) — recommend offboarding
⚠️ [name] — full permissions but role doesn't require it
⚠️ [n] accounts without 2FA enabled (if data available)

PERMISSIONS DISTRIBUTION:
  Limited:        [n]
  Custom roles:   [n]
  Full access:    [n]

If all accounts active and appropriately permissioned:
✅ Staff access in good standing.

VIP Customer Watcher

Schedule: Every day at 9:00 AM local time Runtime: ~3-5 minutes Slack channel: #vip-care

Prompt

You are the VIP customer concierge. Daily check on VIP customer activity.

1. shopify-admin-vip-customer-identifier (top 5%, dry_run: true) — get current VIPs
2. shopify-admin-order-lookup-and-summary (days_back: 1) — find any VIP orders
3. shopify-admin-churn-risk-scorer (days_back: 365) — filter to VIPs

Cross-reference: any VIPs in the at-risk segment? Any VIPs with orders that
have issues (held, refunded, complained)?

Send to #vip-care:

👑 VIP CUSTOMER WATCH — [DATE]
━━━━━━━━━━━━━━━━━━━━━━━

NEW VIP ORDERS (last 24h):
• [name] ($[lifetime spend]) → Order [#name] $[amount]
  Status: [fulfillment status]
  → Action: ensure expedited handling

VIP ATTENTION REQUIRED:
🔴 [name] — last order disputed/refunded — reach out
⚠️ [name] — churn risk score [n] — overdue by [n] days

VIP HEALTH:
  Total VIPs:         [n]
  Active (last 30d):  [n]
  At risk:            [n]
  Lifetime VIP rev:   $[amount]

If no VIP issues: ✅ All VIPs orders flowing smoothly.

Weekly Business Review

Schedule: Every Monday at 8:00 AM local time Runtime: ~8-12 minutes (runs 10 skills) Slack channel: #weekly-review

Prompt

You are the business analyst for 91pqhx-iy.myshopify.com. Generate the Weekly Business Review.

Run each skill with days_back: 7 (comparing to prior week where possible) and compile results:

1. REVENUE & ORDERS
   - shopify-admin-sales-by-channel-report (days_back: 7)
   - shopify-admin-average-order-value-trends (days_back: 14, bucket: week) — compare this week vs last
   - shopify-admin-profit-margin-calculator (days_back: 7, group_by: product)

2. PRODUCTS
   - shopify-admin-top-product-performance (days_back: 7)
   - shopify-admin-stock-velocity-report (days_back: 7)

3. CUSTOMERS
   - shopify-admin-repeat-purchase-rate (days_back: 7)
   - shopify-admin-customer-cohort-analysis (days_back: 7)
   - shopify-admin-churn-risk-scorer (days_back: 365, report only summary counts)

4. CONVERSION
   - shopify-admin-checkout-abandonment-report (days_back: 7)
   - shopify-admin-refund-rate-analysis (days_back: 7)

Format as a comprehensive Slack message to #weekly-review:

📈 WEEKLY BUSINESS REVIEW — Week of [DATE]
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

💰 REVENUE
  Total revenue:    $[amount] ([↑/↓][pct]% vs last week)
  Orders:           [n] ([↑/↓][pct]%)
  AOV:              $[amount] ([↑/↓]$[delta])
  Est. net profit:  $[amount] (margin: [pct]%)

  By channel:
  • Online Store: $[amount] ([pct]%)
  • [Other channels...]

🏆 TOP PRODUCTS
  1. [product] — [units] sold, $[revenue]
  2. [product] — [units] sold, $[revenue]
  3. [product] — [units] sold, $[revenue]

👥 CUSTOMERS
  New customers:     [n]
  Repeat rate:       [pct]%
  At-risk customers: [n] (total revenue at risk: $[amount])

🔄 CONVERSION HEALTH
  Cart abandonment: [pct]%
  Refund rate:      [pct]% ($[amount] refunded)

📦 INVENTORY
  Fast movers: [top 3 by velocity]
  Slow movers: [bottom 3 by velocity]

KEY TAKEAWAYS:
• [Biggest positive trend]
• [Biggest concern or action item]
• [Recommendation]
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Astro Starter Kit: Minimal

pnpm create astro@latest -- --template minimal

🧑‍🚀 Seasoned astronaut? Delete this file. Have fun!

🚀 Project Structure

Inside of your Astro project, you'll see the following folders and files:

/
├── public/
├── src/
│   └── pages/
│       └── index.astro
└── package.json

Astro looks for .astro or .md files in the src/pages/ directory. Each page is exposed as a route based on its file name.

There's nothing special about src/components/, but that's where we like to put any Astro/React/Vue/Svelte/Preact components.

Any static assets, like images, can be placed in the public/ directory.

🧞 Commands

All commands are run from the root of the project, from a terminal:

👀 Want to learn more?

Feel free to check our documentation or jump into our Discord server.