Integrate Shopify's AI Toolkit MCP server with Claude Code for GraphQL validation, Liquid linting, and documentation search. Use when setting up Shopify MCP integration, validating GraphQL queries, or linting Liquid templates. Trigger with phrases like "shopify ai toolkit", "shopify mcp", "shopify claude integration", "shopify graphql validation", "shopify liquid lint".
Shopify's AI Toolkit provides an official MCP (Model Context Protocol) server for developer workflows. This skill configures it for Claude Code, enabling GraphQL schema validation, Liquid template linting, and real-time Shopify documentation search without leaving your editor.
read_products, read_themes)Create or update .mcp.json in your project root:
{
"mcpServers": {
"shopify": {
"command": "npx",
"args": [
"-y",
"@shopify/ai-toolkit-mcp"
],
"env": {
"SHOPIFY_ACCESS_TOKEN": "${SHOPIFY_ACCESS_TOKEN}",
"SHOPIFY_STORE_URL": "https://your-store.myshopify.com"
}
}
}
}
Note: The package name
@shopify/ai-toolkit-mcpmay change. Check shopify.dev/docs for the latest official MCP package name and configuration.
Set your environment variables:
export SHOPIFY_ACCESS_TOKEN="shpat_xxxxxxxxxxxxxxxxxxxx"
export SHOPIFY_STORE_URL="https://your-store.myshopify.com"
See references/mcp-config.md for advanced configuration and troubleshooting.
With the MCP server running, validate GraphQL queries against your store's live schema before executing them. The MCP introspects the schema and catches:
priceRange should be priceRangeV2)ProductInput split into ProductCreateInput/ProductUpdateInput in 2024-10)See references/graphql-validation.md for common validation errors and deprecated field mappings.
The AI Toolkit checks Liquid templates for syntax errors, deprecated filters, and performance anti-patterns:
| Issue | Bad | Fix |
|---|---|---|
| Deprecated filter | img_url: '800x' | image_url: width: 800 |
| Scope leak | {% include 'card' %} | {% render 'card', product: product %} |
| Unbounded loop | collections.all.products | Add limit: 50 or use paginate |
Query Shopify's official documentation through the MCP server for up-to-date API information:
# Example MCP queries you can make:
# - "What fields are available on the Product object?"
# - "Show me the companyCreate mutation input shape"
# - "What scopes do I need for B2B operations?"
# - "What changed in the 2024-10 API version?"
This is especially useful for checking API version changes and deprecated fields without leaving your editor.
See references/workflow-integration.md for combining MCP tools with other shopify-pack skills.
| Error | Cause | Solution |
|---|---|---|
MCP connection refused | MCP server failed to start | Check Node.js version (18+), run npx @shopify/ai-toolkit-mcp manually to see errors |
Invalid access token | Token expired or wrong format | Regenerate token in Shopify admin (Settings > Apps > Develop apps) |
Schema introspection failed | Insufficient API scopes | Add read_products and read_themes scopes to your app |
Rate limited on doc queries | Too many documentation requests | Space out queries; docs are cached after first fetch |
ENOENT: .mcp.json not found | Config file missing or wrong location | Must be in project root or ~/.claude/.mcp.json for global config |
Package not found | MCP package name changed | Check shopify.dev/docs for the current official package name |
You need to call productCreate but aren't sure which input fields changed in the latest API version. Use the MCP server to validate your query against the live schema.
See GraphQL Validation for common validation errors and deprecated field mappings.
Configure the Shopify MCP server at project, user, and global levels so the entire team gets schema validation automatically.
See MCP Config for advanced configuration and troubleshooting.
Chain MCP schema validation before running mutations from other shopify-pack skills like B2B or checkout extensions.
See Workflow Integration for the combined workflow patterns.