Vector search: discover embedding tables (no query_text) or perform semantic search (with query_text). Configure BIGQUERY_EMBEDDING_MODEL env var for search mode.
AI agents call vector_search to retrieve information from BigQuery MCP Server without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
Vector search is a data retrieval operation that queries embeddings and returns results without modifying, deleting, or executing code. While it accesses BigQuery data, this falls squarely under Read operations. The low severity reflects minimal blast radius—misuse would result in unauthorized data access rather than data loss or system compromise.
From the tool's definition Tool description states it is used to 'discover embedding tables' or 'perform semantic search' with query_text. No modification, deletion, or execution of arbitrary code is implied. The server description reinforces this is a 'safe read-only' operation.
Documented attack patterns abuse exactly the kind of access vector_search gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and BigQuery MCP Server, and nothing reaches the server without passing your rules. This is the rule we recommend for vector_search:
{
"version": "1",
"default": "deny",
"tools": {
"vector_search": {}
}
}vector_search 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.
Vector search: discover embedding tables (no query_text) or perform semantic search (with query_text). Configure BIGQUERY_EMBEDDING_MODEL env var for search mode. It is categorised as a Read tool in the BigQuery MCP Server MCP Server, which means it retrieves data without modifying state.
Register the BigQuery MCP Server MCP server in PolicyLayer and add a rule for vector_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 BigQuery MCP Server. Nothing to install.
vector_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 vector_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 vector_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.
vector_search is provided by the BigQuery MCP Server MCP server (pvoo/bigquery-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from BigQuery MCP Server, add the rest of your stack, and see everything your agents can call. Then put policy on all of it.
Free to start. No card required.
5 BigQuery MCP Server tools catalogued and risk-classified — across an index of 43,000+ MCP servers.