shop

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill requires installing/running a remote MCP tool via npx (command shown: npx -y @henrylabs/mcp@latest), which fetches and executes remote code that provides the runtime "execute" tool (TypeScript sandbox) the skill depends on, so this is a runtime external dependency that executes remote code.

MEDIUM W009: Direct money access capability detected (payment gateways, crypto, banking).

• Direct money access detected (high risk: 1.00). The skill explicitly integrates with an e‑commerce/payment API (Henry MCP) and calls endpoints that create carts and hosted checkouts: search_products/retrieve_product_details/add_cart_items/create_cart_checkout (or henry.cart.create / henry.cart.item.add and henry.orders.list). Creating a checkout session is a payment gateway operation (it initiates the payment flow and can result in charges once the hosted checkout is completed). Although the skill disclaims "hosted-checkout only" and won't collect card data in chat, it is specifically designed to perform cart and checkout actions rather than being a generic browser or HTTP tool. Therefore it grants direct financial execution capability.