AI agents call local.descriptions to retrieve information from Earch MCP without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves business location descriptions from the Brave Search API. It performs a read-only query operation with no side effects, no code execution, and no data modification. Even with 20 locations queried at once, the blast radius of misuse is minimal—an agent would only retrieve publicly available business information.
From the tool's definition Tool fetches descriptions for location IDs without modifying, deleting, or executing operations. The verb 'fetch' and the absence of write/delete/execute capability in the description indicate pure data retrieval.
Attacks that exploit this kind of access
Brave Local Search AI descriptions: fetch descriptions for up to 20 location ids. It is categorised as a Read tool in the Earch MCP MCP Server, which means it retrieves data without modifying state.
Register the Earch MCP server in PolicyLayer and add a rule for local.descriptions: 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 Earch MCP. Nothing to install.
local.descriptions 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 local.descriptions 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 local.descriptions. 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.
local.descriptions is provided by the Earch MCP server (nanameru/search-mcp). 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 →