AI agents call recommend_food to retrieve information from MCP-Demo without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves and returns information about local cuisine recommendations. It does not create, modify, delete, or execute any operations with external effects. It is purely informational, similar to a search or lookup function. The minimal blast radius if misused would be returning irrelevant or fabricated food recommendations.
From the tool's definition Tool name 'recommend_food' and description indicate it 'outputs local specialty foods based on input location name' — a query/retrieval operation with no side effects.
Attacks that exploit this kind of access
特色美食推荐。根据输入的地名，输出当地的特色美食。. It is categorised as a Read tool in the MCP-Demo MCP Server, which means it retrieves data without modifying state.
Register the MCP-Demo MCP server in PolicyLayer and add a rule for recommend_food: 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-Demo. Nothing to install.
recommend_food 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 recommend_food 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 recommend_food. 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.
recommend_food is provided by the MCP-Demo MCP server (tatocode/mcp-demo). 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 →