AI agents call catalog_search to retrieve information from DB-MCP without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool performs read-only metadata queries against a catalog. It returns schema information (measures, dimensions, segments) to help users understand available data, without executing data queries, modifying records, or triggering side effects. The governance layer (PII blocking, access limits) mentioned in the server description applies downstream at query execution, not at the catalog search stage.
From the tool's definition Tool name 'catalog_search' and description 'Fuzzy-search the data catalog for measures, dimensions, and segments' indicate a metadata discovery operation.
Attacks that exploit this kind of access
Fuzzy-search the data catalog for measures, dimensions, and segments. It is categorised as a Read tool in the DB-MCP MCP Server, which means it retrieves data without modifying state.
Register the DB- MCP server in PolicyLayer and add a rule for catalog_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 DB-MCP. Nothing to install.
catalog_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 catalog_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 catalog_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.
catalog_search is provided by the DB- MCP server (scrappymonkey/db-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 →