The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/checkout-admin-read.mjs uses execFile to invoke the Shopify CLI. This method is used safely to execute commands like shopify store execute without spawning a shell, reducing the risk of shell injection. It also provides a configuration-based override (SKILL_HUB_SHOPIFY_CLI_JS) to point to a specific Shopify CLI entry point.
[DATA_EXFILTRATION]: While the skill accesses Shopify store data (orders, delivery profiles, and payment settings), it implements a safeError function that automatically redacts access tokens and client secrets from any error output. Furthermore, the collect command is designed to aggregate transaction signals and delivery counts rather than exporting raw customer data, order IDs, or payment credentials.
[EXTERNAL_DOWNLOADS]: The skill communicates with official Shopify domains (myshopify.com and admin.shopify.com). These are well-known, trusted services necessary for the skill's primary function of auditing Shopify store status.
[INDIRECT_PROMPT_INJECTION]: The SKILL.md instructions include a security boundary warning the agent to treat storefront HTML, checkout text, and redirects as untrusted data. It explicitly directs the agent not to follow instructions found within those external resources, which mitigates risks associated with indirect prompt injection from compromised storefronts.
[CREDENTIALS_UNSAFE]: The skill manages Shopify API credentials using environment variables. The documentation correctly instructs users to store these in a private .env file that should be excluded from version control, adhering to standard security best practices for secret management.

shopify-checkout-payment-connection-check