AI agents call grimoire_search to retrieve information from Grimoire without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves and queries data from a local security knowledge base without creating, modifying, deleting, or executing any operations. It is a pure read operation analogous to a search or get function. The low severity reflects that misuse would at worst return irrelevant security information, with no destructive, financial, or code execution consequences.
From the tool's definition Tool name is 'grimoire_search' and description states it 'Search[es] the Grimoire security knowledge base' with support for 'keyword (FTS5), semantic (embedding), and hybrid search modes.' The verb 'search' and explicit mention of query modes confirm…
Attacks that exploit this kind of access
Search the Grimoire security knowledge base. Supports keyword (FTS5), semantic (embedding), and hybrid search modes. It is categorised as a Read tool in the Grimoire MCP Server, which means it retrieves data without modifying state.
Register the Grimoire MCP server in PolicyLayer and add a rule for grimoire_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 Grimoire. Nothing to install.
grimoire_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 grimoire_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 grimoire_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.
grimoire_search is provided by the Grimoire MCP server (tannernicol/grimoire). 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 →