From shopify-admin-skills
Read-only: tracks cancellation rate over time and breaks down cancelled orders by cancelReason to surface fraud, inventory, customer, and declined-payment patterns.
How this skill is triggered — by the user, by Claude, or both
Slash command
/shopify-admin-skills:shopify-admin-order-cancellation-analysisThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Computes cancellation rate (cancelled orders / total orders) over a configurable window, broken down by `cancelReason` (`CUSTOMER`, `FRAUD`, `INVENTORY`, `DECLINED`, `OTHER`, `STAFF`). Surfaces shifts in cancellation patterns — for example, a spike in `INVENTORY` cancellations suggests a stock data integrity problem, while a spike in `FRAUD` suggests a coordinated attack. Read-only — no mutations.
Computes cancellation rate (cancelled orders / total orders) over a configurable window, broken down by cancelReason (CUSTOMER, FRAUD, INVENTORY, DECLINED, OTHER, STAFF). Surfaces shifts in cancellation patterns — for example, a spike in INVENTORY cancellations suggests a stock data integrity problem, while a spike in FRAUD suggests a coordinated attack. Read-only — no mutations.
shopify store auth --store <domain> --scopes read_ordersread_orders| Parameter | Type | Required | Default | Description |
|---|---|---|---|---|
| store | string | yes | — | Store domain (e.g., mystore.myshopify.com) |
| days_back | integer | no | 30 | Lookback window for orders included in the analysis |
| bucket | string | no | day | Time bucket: day, week, or month |
| min_value | float | no | 0 | Only include orders above this total value |
| reason_filter | string | no | — | Optional filter to a single cancelReason |
| format | string | no | human | Output format: human or json |
ℹ️ Read-only skill — no mutations are executed. Safe to run at any time. The analysis uses
cancelReasonas recorded by Shopify or staff at cancellation time — accuracy depends on staff selecting the correct reason.
OPERATION: orders — query
Inputs: query: "created_at:>='<NOW - days_back days>'", first: 250, select cancelledAt, cancelReason, displayFinancialStatus, totalPriceSet, pagination cursor
Expected output: All orders created in the window (cancelled and non-cancelled) for rate calculation; paginate until hasNextPage: false
Partition orders into cancelled (cancelledAt != null) and not cancelled. Compute overall rate = cancelled / total.
For cancelled orders, group by cancelReason and by time bucket. Compute rate per bucket and per reason.
Identify time buckets where any single reason exceeds 2x its trailing 7-bucket average — flag as anomalies.
# orders:query — validated against api_version 2025-01
query OrdersForCancellationAnalysis($query: String!, $after: String) {
orders(first: 250, after: $after, query: $query) {
edges {
node {
id
name
createdAt
cancelledAt
cancelReason
displayFinancialStatus
displayFulfillmentStatus
totalPriceSet {
shopMoney {
amount
currencyCode
}
}
customer {
id
numberOfOrders
}
staffMember {
id
name
}
}
}
pageInfo {
hasNextPage
endCursor
}
}
}
Claude MUST emit the following output at each stage. This is mandatory.
On start, emit:
╔══════════════════════════════════════════════╗
║ SKILL: Order Cancellation Analysis ║
║ 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):
══════════════════════════════════════════════
CANCELLATION ANALYSIS (<days_back> days, by <bucket>)
Total orders: <n>
Cancelled orders: <n> (<pct>%)
Lost revenue: $<amount>
By reason:
CUSTOMER <n> (<pct>%)
FRAUD <n> (<pct>%)
INVENTORY <n> (<pct>%)
DECLINED <n> (<pct>%)
OTHER <n> (<pct>%)
Anomaly buckets (>2x trailing avg):
<bucket-key> reason=<reason> rate=<pct>%
Output: cancellation_analysis_<date>.csv
══════════════════════════════════════════════
For format: json, emit:
{
"skill": "order-cancellation-analysis",
"store": "<domain>",
"period_days": 30,
"bucket": "day",
"total_orders": 0,
"cancelled_orders": 0,
"cancellation_rate": 0,
"lost_revenue": 0,
"currency": "USD",
"by_reason": {
"CUSTOMER": 0, "FRAUD": 0, "INVENTORY": 0, "DECLINED": 0, "OTHER": 0
},
"anomalies": [],
"output_file": "cancellation_analysis_<date>.csv"
}
CSV file cancellation_analysis_<YYYY-MM-DD>.csv with columns:
bucket_start, bucket_end, total_orders, cancelled_orders, rate_pct, reason_customer, reason_fraud, reason_inventory, reason_declined, reason_other, lost_revenue, currency
| Error | Cause | Recovery |
|---|---|---|
THROTTLED | API rate limit exceeded | Wait 2 seconds, retry up to 3 times |
cancelReason is null on cancelled order | Older order pre-dating reason field | Bucket into OTHER, log count |
| No orders in window | Empty store or test domain | Exit with summary: 0 orders, 0% rate |
| Cancelled order created outside window | Cancellation happened in window but order older | Excluded by design — analyses creation cohort |
INVENTORY cancellations indicate a sync issue between storefront stock and warehouse — pair this skill with multi-location-inventory-audit.FRAUD cancellations indicate either improving fraud filters (good) or a coordinated attack (bad) — cross-reference with order-risk-report.DECLINED rates often correlate with checkout friction or expired payment methods — investigate alongside checkout abandonment data.npx claudepluginhub 40rty-ai/shopify-admin-skills --plugin shopify-admin-skillsCalculates refund rates by product, vendor, or time period from Shopify orders. Identifies quality and listing issues by surfacing products with high refund activity.
Triages WooCommerce refunds by analyzing refund metrics, top products, and countries to produce operational diagnostics without exposing PII.
Creates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.