Shopify Admin Order Risk Report

Safety

ℹ️ Read-only skill — no mutations are executed. Safe to run at any time. Risk scores are generated by Shopify's fraud analysis and are advisory only — they do not block orders automatically unless you configure Shopify's fraud filters.

Workflow Steps

1. OPERATION: orders — query Inputs: query: "risk_level:<risk_level> created_at:>='<NOW - days_back days>'", first: 250, select riskLevel, riskFacts, totalPriceSet, customer, pagination cursor Expected output: Orders with risk data; paginate until hasNextPage: false

2. Sort by risk level (high first), then by order value descending

GraphQL Operations

# orders:query — validated against api_version 2025-01
query OrderRiskReport($query: String!, $after: String) {
  orders(first: 250, after: $after, query: $query) {
    edges {
      node {
        id
        name
        createdAt
        displayFinancialStatus
        displayFulfillmentStatus
        totalPriceSet {
          shopMoney {
            amount
            currencyCode
          }
        }
        riskLevel
        riskFacts {
          message
          sentiment
        }
        customer {
          id
          displayName
          defaultEmailAddress {
            emailAddress
          }
          numberOfOrders
        }
        shippingAddress {
          countryCode
          city
        }
        billingAddress {
          countryCode
          city
        }
      }
    }
    pageInfo {
      hasNextPage
      endCursor
    }
  }
}

Session Tracking

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

On start, emit:

╔══════════════════════════════════════════════╗
║  SKILL: Order Risk Report                    ║
║  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>

On completion, emit:

For format: human (default):

══════════════════════════════════════════════
ORDER RISK REPORT  (<days_back> days)
  Orders reviewed:  <n>
  High risk:        <n>
  Medium risk:      <n>
  Low risk:         <n>

  High Risk Orders:
    #<name>  $<amount>  <customer>
      Risks: <indicator>, <indicator>
  Output: risk_report_<date>.csv
══════════════════════════════════════════════

For format: json, emit:

{
  "skill": "order-risk-report",
  "store": "<domain>",
  "period_days": 7,
  "orders_reviewed": 0,
  "high_risk_count": 0,
  "medium_risk_count": 0,
  "output_file": "risk_report_<date>.csv"
}

Output Format

CSV file risk_report_<YYYY-MM-DD>.csv with columns: order_name, order_id, created_at, risk_level, total_price, currency, customer_name, customer_email, risk_indicators, financial_status, fulfillment_status

Error Handling

Best Practices

• Do not auto-cancel high-risk orders — use high-risk-order-tagger to flag them for manual hold first, then review before cancelling.
• An order from a repeat customer (numberOfOrders > 3) with a high risk score is usually a false positive — apply judgment before acting.
• min_order_value helps focus review effort on high-value fraud risk; $50+ is a reasonable floor for most stores.
• Run daily as part of a morning ops routine, especially during high-volume sale periods when fraud attempts increase.