Skip to content
Back to skills

Shopify Admin Email Deliverability Audit

ASecurity

Read-only: scans the customer database for malformed emails, role accounts, disposable domains, and bounce-suspect patterns to protect sender reputation.

  • 181 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added May 28, 2026
businessnodeapidatabase

Works with

  • claude code
  • cursor
  • cli
  • api

Security analysis

A100/100

Scanned May 28, 2026

npx -y skills add 40RTY-ai/shopify-admin-skills --skill shopify-admin-email-deliverability-audit --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Shopify Admin Email Deliverability Audit?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Shopify Admin Email Deliverability Audit
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/40rty-ai-shopify-admin-email-deliverability-audit/badge)](https://www.skillsdirectory.com/skills/40rty-ai-shopify-admin-email-deliverability-audit)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: shopify-admin-email-deliverability-audit
role: customer-ops
description: "Read-only: scans the customer database for malformed emails, role accounts, disposable domains, and bounce-suspect patterns to protect sender reputation."
toolkit: shopify-admin, shopify-admin-execution
api_version: "2025-01"
graphql_operations:
  - customers:query
status: stable
compatibility: Claude Code, Cursor, Codex, Gemini CLI
---

## Purpose
Scans the entire customer list and flags addresses that will hurt email deliverability if included in marketing sends: syntactically invalid addresses, role accounts (info@, admin@, sales@), known disposable / temporary domains, and suspected hard-bounce patterns. Output is a suppression list ready to import into your email platform. Read-only — no mutations.

## Prerequisites
- Authenticated Shopify CLI session: `shopify store auth --store <domain> --scopes read_customers`
- API scopes: `read_customers`

## Parameters

| Parameter | Type | Required | Default | Description |
|-----------|------|----------|---------|-------------|
| store | string | yes | — | Store domain (e.g., mystore.myshopify.com) |
| marketing_consent_only | bool | no | true | Only scan customers who currently accept marketing — those are the ones at risk of being mailed |
| disposable_domains | array | no | built-in list | Override built-in disposable-domain list |
| role_localparts | array | no | `["info","admin","sales","support","contact","noreply","help","webmaster","postmaster"]` | Local-part prefixes to flag as role accounts |
| format | string | no | human | Output format: `human` or `json` |

## Safety

> ℹ️ Read-only skill — no mutations are executed. Safe to run at any time. No emails are sent, no marketing consent is changed.

## Detection Rules

For each customer email, run these checks in order and assign one or more flags:

1. **`invalid_syntax`** — fails RFC 5322 local-part / domain validation, missing `@`, contains whitespace, double dots, leading/trailing dot
2. **`role_account`** — local part exactly matches a `role_localparts` entry (case-insensitive)
3. **`disposable_domain`** — domain matches the disposable-domain list (mailinator, guerrillamail, tempmail-style domains, etc.)
4. **`plus_alias`** — contains `+` in local part (informational; not a deliverability problem on its own, but useful for de-duplication)
5. **`bounce_suspect`** — heuristics: numeric-only local part, length > 64 chars, ALL-CAPS, repeated characters (`aaaaa`), keyboard rolls (`asdfghjk`)
6. **`duplicate`** — same normalized email already seen in the customer set

A customer can carry multiple flags; the most severe (`invalid_syntax` > `bounce_suspect` > `disposable_domain` > `role_account` > `plus_alias`) drives the recommended action.

## Workflow Steps

1. **OPERATION:** `customers` — query
   **Inputs:** `first: 250`, select `id`, `defaultEmailAddress { emailAddress, marketingState }`, `numberOfOrders`, `tags`, pagination cursor. If `marketing_consent_only: true`, filter `query: "email_marketing_state:subscribed"`
   **Expected output:** All targeted customers with email; paginate until `hasNextPage: false`

2. Run detection rules over each email; collect flags

3. Aggregate counts per flag, build suppression list

## GraphQL Operations

```graphql
# customers:query — validated against api_version 2025-01
query DeliverabilityAudit($query: String, $after: String) {
  customers(first: 250, after: $after, query: $query) {
    edges {
      node {
        id
        displayName
        defaultEmailAddress {
          emailAddress
          marketingState
        }
        numberOfOrders
        amountSpent {
          amount
          currencyCode
        }
        tags
        createdAt
      }
    }
    pageInfo {
      hasNextPage
      endCursor
    }
  }
}
```

## Session Tracking

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

**On start**, emit:
```
╔══════════════════════════════════════════════╗
║  SKILL: Email Deliverability Audit           ║
║  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):
```
══════════════════════════════════════════════
EMAIL DELIVERABILITY AUDIT
  Customers scanned:       <n>
  Subscribed customers:    <n>
  ─────────────────────────────
  Invalid syntax:          <n> (<pct>%)   🔴 suppress
  Role accounts:           <n> (<pct>%)   ⚠️ suppress
  Disposable domains:      <n> (<pct>%)   ⚠️ suppress
  Bounce-suspect patterns: <n> (<pct>%)   ⚠️ review
  Plus aliases:            <n> (<pct>%)   ℹ️ informational
  Duplicates:              <n> (<pct>%)   ℹ️ informational

  Recommended suppression: <n>  (<pct>% of subscribed)
  Output: deliverability_audit_<date>.csv
══════════════════════════════════════════════
```

For `format: json`, emit:
```json
{
  "skill": "email-deliverability-audit",
  "store": "<domain>",
  "customers_scanned": 0,
  "flags": {
    "invalid_syntax": 0,
    "role_account": 0,
    "disposable_domain": 0,
    "bounce_suspect": 0,
    "plus_alias": 0,
    "duplicate": 0
  },
  "recommended_suppressions": 0,
  "output_file": "deliverability_audit_<date>.csv"
}
```

## Output Format
CSV file `deliverability_audit_<YYYY-MM-DD>.csv` with columns:
`customer_id`, `email`, `flags`, `recommended_action`, `marketing_state`, `order_count`, `total_spent`, `created_at`

`recommended_action` is one of: `suppress`, `review`, `keep`.

## Error Handling
| Error | Cause | Recovery |
|-------|-------|----------|
| `THROTTLED` | API rate limit exceeded | Wait 2 seconds, retry up to 3 times |
| Customer with no email | Phone-only / POS account | Skip, do not include in suppression list |
| Disposable list out of date | New temp-mail domain not in built-in list | Caller can override via `disposable_domains` parameter |
| Unicode local parts | Internationalized email addresses | Normalize via NFC; do not flag valid IDN domains |

## Best Practices
- Run before every large promotional send — high invalid / bounce rates above 2% put your sending domain reputation at risk.
- Treat `role_account` flags as soft-suppression: those addresses rarely consent to marketing meaningfully and frequently mark mail as spam.
- Re-audit quarterly even if the customer list is static — domain reputation lists change, and disposable-email providers add new domains.
- Pair with `marketing-consent-report` to confirm consent state aligns with what your email service provider has on file.
- Hand the resulting suppression CSV to your email service provider's import-suppression-list feature; do not silently delete customers from Shopify based on this audit.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.