AI agents call query_rag to retrieve information from Verbatim without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves or queries academic papers and related metadata from document collections, consistent with the server's stated purpose of searching and citing research. No description provided, which slightly lowers confidence, but the context from sibling tools and server purpose strongly indicates a read-only search operation with no side effects.
From the tool's definition Tool named 'query_rag' is part of a server that 'Search academic papers and ask research questions with verbatim, cited answers from document collections.' Sibling tools include search_papers, get_paper, get_citation, list_* operations — all read-only…
Attacks that exploit this kind of access
query_rag. It is categorised as a Read tool in the Verbatim MCP Server, which means it retrieves data without modifying state.
Register the Verbatim MCP server in PolicyLayer and add a rule for query_rag: 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 Verbatim. Nothing to install.
query_rag 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 query_rag 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 query_rag. 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.
query_rag is provided by the Verbatim MCP server (krlabsorg/verbatim-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 →