Agentic Commerce Catalog Audit

Purpose

Audits every product in a Shopify store against the Agentic Commerce Spec v1.0 — an open scoring standard for AI shopping agent visibility. Returns a 0-100 score per product, a per-dimension breakdown (title, description, images, metadata, taxonomy, variants, reviews, structured data), and a ranked list of issues by predicted impact on AI agent recommendation probability.

Use this skill when you want to know which of your products will get skipped by ChatGPT, Gemini, Claude, Mistral, and DeepSeek when they generate shopping recommendations, and which fixes would lift visibility fastest.

Prerequisites

• Authenticated Shopify CLI session: shopify auth login --store
• API scopes: read_products, read_product_listings

Parameters

| Parameter | Type | Required | Default | Description | |-----------|------|----------|---------|-------------| | store | string | yes | — | Store domain (e.g., mystore.myshopify.com) | | format | string | no | human | Output format: human or json | | limit | integer | no | 250 | Max products to audit in this run (paginated) | | minscore | integer | no | 0 | Only return products scoring below this threshold | | topissues | integer | no | 10 | How many top issues to surface in the summary |

Workflow Steps

1. OPERATION: products — query

Inputs: first: 50 per page, pagination via cursor until limit reached or no more products Expected output: Product nodes with title, descriptionHtml, productType, vendor, tags, images, media, variants, metafields, seo

1. For each product, compute the Agentic Commerce Spec v1.0 score using the reference scorer in @commerce-agentic/spec. The scorer is deterministic, runs locally, no network calls. Output per product:

``json { "productId": "gid://shopify/Product/12345", "title": "...", "totalScore": 67, "grade": "C", "dimensions": { "title": 14, "description": 9, "images": 7, "metadata": 10, "taxonomy": 6, "variants": 8, "reviews": 5, "structured_data": 8 }, "issues": ["Missing google_product_category metafield", "Alt text on 1/6 images"], "aiVisibilityIndex": 0.62 } ``

1. Aggregate across the catalog:

• Compute store-wide average score
• Bucket products: AI-Ready (>=80), Needs Work (50-79), Invisible ( Read-only operation. No mutations. Safe to run anytime, including on a production store.

GraphQL Operations

# products:query — validated against api_version 2025-01
query CatalogAudit($first: Int!, $after: String) {
  products(first: $first, after: $after) {
    edges {
      node {
        id
        title
        descriptionHtml
        productType
        vendor
        tags
        seo {
          title
          description
        }
        images(first: 10) {
          edges {
            node {
              altText
              url
            }
          }
        }
        variants(first: 50) {
          edges {
            node {
              sku
              barcode
              price
              inventoryQuantity
            }
          }
        }
        metafields(first: 25) {
          edges {
            node {
              namespace
              key
              value
            }
          }
        }
      }
    }
    pageInfo {
      hasNextPage
      endCursor
    }
  }
}

Reference

This skill implements the audit phase of the Agentic Commerce Spec v1.0. The scoring methodology is fully documented at METHODOLOGY.md. Cross-implementation test vectors are at test/vectors.json.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: commerce-agentic
• Source: commerce-agentic/agentic-commerce-skills
• License: MIT

Install and usage instructions live in the source repository linked above.