Search carts by criteria
AI agents call cart_search to retrieve information from MCP Dynamics 365 Commerce Server without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
The tool performs a search/query operation to retrieve cart information based on specified criteria. This is a non-destructive read operation with no side effects. Even in a commerce system context, searching carts only retrieves existing data without creating, modifying, or deleting anything. Blast radius is minimal—worst case is information disclosure of cart data.
From the tool's definition Tool name 'cart_search' and description 'Search carts by criteria' indicate a query operation that retrieves cart data without modifying state.
Attacks that exploit this kind of access
Search carts by criteria. It is categorised as a Read tool in the MCP Dynamics 365 Commerce Server MCP Server, which means it retrieves data without modifying state.
Register the MCP Dynamics 365 Commerce Server MCP server in PolicyLayer and add a rule for cart_search: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches MCP Dynamics 365 Commerce Server. Nothing to install.
cart_search is a Read tool with low risk. Read-only tools are generally safe to allow by default.
Yes. Add a rate_limit block to the cart_search rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for cart_search. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
cart_search is provided by the MCP Dynamics 365 Commerce Server MCP server (jiantmo/mcp-commerce). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Every MCP server has a record like this.
Type a name, get the same breakdown: verified identity, auth posture, risk grade, capabilities, recommended policy.
Teams ship this data inside their own products. See what a licence covers →