Shopify Admin High Risk Order Tagger

Purpose

Queries recent high-risk orders and takes two protective actions: tags the order for staff visibility and optionally places a fulfillment hold to prevent the order from shipping until reviewed. Complements order-risk-report (which only reads) with write actions that create a reviewable queue.

Prerequisites

• Authenticated Shopify CLI session: shopify store auth --store <domain> --scopes read_orders,write_orders,write_fulfillments
• API scopes: read_orders, write_orders, write_fulfillments

Parameters

Safety

⚠️ fulfillmentOrderHold prevents orders from being fulfilled until the hold is explicitly released. Customers will experience a shipping delay while on hold. Use hold_fulfillment: false if you only want to tag without blocking fulfillment. Run with dry_run: true to confirm the order list before committing. Release holds with the order-hold-and-release skill after review.

Workflow Steps

1. OPERATION: orders — query Inputs: query: "risk_level:high created_at:>='<NOW - days_back days>'", first: 250, select riskLevel, fulfillmentOrders, totalPriceSet Expected output: High-risk orders in window

2. OPERATION: tagsAdd — mutation Inputs: Order id, tags: [<tag>] Expected output: Updated order tags; userErrors

3. OPERATION: fulfillmentOrderHold — mutation (if hold_fulfillment: true) Inputs: fulfillmentOrderId, reason: <hold_reason>, reasonNotes: "High-risk order — awaiting fraud review" Expected output: heldFulfillmentOrder { id, status }, userErrors

GraphQL Operations

# orders:query — validated against api_version 2025-01
query HighRiskOrders($query: String!, $after: String) {
  orders(first: 250, after: $after, query: $query) {
    edges {
      node {
        id
        name
        riskLevel
        totalPriceSet {
          shopMoney {
            amount
            currencyCode
          }
        }
        tags
        fulfillmentOrders(first: 5) {
          edges {
            node {
              id
              status
            }
          }
        }
        customer {
          id
          displayName
          numberOfOrders
        }
      }
    }
    pageInfo {
      hasNextPage
      endCursor
    }
  }
}

# tagsAdd:mutation — validated against api_version 2025-01
mutation TagsAdd($id: ID!, $tags: [String!]!) {
  tagsAdd(id: $id, tags: $tags) {
    node {
      id
    }
    userErrors {
      field
      message
    }
  }
}

# fulfillmentOrderHold:mutation — validated against api_version 2025-01
mutation FulfillmentOrderHold($id: ID!, $fulfillmentHold: FulfillmentOrderHoldInput!) {
  fulfillmentOrderHold(id: $id, fulfillmentHold: $fulfillmentHold) {
    fulfillmentOrder {
      id
      status
    }
    userErrors {
      field
      message
    }
  }
}

Session Tracking

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

On start, emit:

╔══════════════════════════════════════════════╗
║  SKILL: High Risk Order Tagger               ║
║  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
  High-risk orders found:  <n>
  Orders tagged:           <n>
  Fulfillment holds placed: <n>
  Errors:                  <n>
  Output:                  risk_tagging_<date>.csv
══════════════════════════════════════════════

For format: json, emit:

{
  "skill": "high-risk-order-tagger",
  "store": "<domain>",
  "started_at": "<ISO8601>",
  "dry_run": true,
  "outcome": {
    "orders_found": 0,
    "tagged": 0,
    "holds_placed": 0,
    "errors": 0,
    "output_file": "risk_tagging_<date>.csv"
  }
}

Output Format

CSV file risk_tagging_<YYYY-MM-DD>.csv with columns: order_name, order_id, risk_level, total_price, currency, tag_applied, hold_placed, customer_name

Error Handling

Best Practices

• Run within 1–2 hours of order placement — most fraud orders are placed and expected to ship same-day.
• After a hold is placed, use order-risk-report to review the risk indicators in detail before deciding to cancel or release.
• Release legitimate orders with the order-hold-and-release skill to minimize shipping delay.
• Orders from repeat customers (numberOfOrders > 3) are unlikely to be fraudulent — consider filtering them out with min_order_value or a separate query.