AI agents call search_google to retrieve information from MCP Web Research Server without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool queries Google's search API and returns results. It has no capability to modify data, execute code, delete information, or commit financial transactions. The operation is purely informational and read-only, making it the lowest risk category. Severity is low because even if misused by an agent, the worst outcome would be performing unwanted searches, which causes no data loss or system compromise.
From the tool's definition Tool name 'search_google' combined with description 'Search Google for a query' indicates a read-only operation that retrieves search results without modifying any data or triggering external side effects beyond data retrieval.
Documented attack patterns abuse exactly the kind of access search_google gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and MCP Web Research Server, and nothing reaches the server without passing your rules. This is the rule we recommend for search_google:
{
"version": "1",
"default": "deny",
"tools": {
"search_google": {}
}
}search_google is read-only, so it stays allowed — but everything else on the server is denied unless you say otherwise.
Free to start. No card required.
Search Google for a query. It is categorised as a Read tool in the MCP Web Research Server MCP Server, which means it retrieves data without modifying state.
Register the MCP Web Research Server MCP server in PolicyLayer and add a rule for search_google: 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 Web Research Server. Nothing to install.
search_google 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 search_google 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 search_google. 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.
search_google is provided by the MCP Web Research Server MCP server (qpd-v/mcp-deepwebresearch). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Deterministic rules across all 5 MCP Web Research Server tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.
Free to start. No card required.
5 MCP Web Research Server tools catalogued and risk-classified — across an index of 42,500+ MCP servers.