AI agents call mem_search to retrieve information from Crossmem without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool queries/retrieves information from a cross-project memory system. While the description is empty, the name and context indicate a search/retrieval operation (Read category). Severity is medium because unauthorized memory searches could leak sensitive cross-project knowledge, training data, or private context that AI assistants have stored, even though the action itself is non-destructive.
From the tool's definition Tool name 'mem_search' combined with server description stating 'search across all Claude Code and Gemini CLI memories' and sibling tools 'mem_recall', 'mem_save', 'mem_forget', 'mem_ingest'.
Attacks that exploit this kind of access
mem_search. It is categorised as a Read tool in the Crossmem MCP Server, which means it retrieves data without modifying state.
Register the Crossmem MCP server in PolicyLayer and add a rule for mem_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 Crossmem. Nothing to install.
mem_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 mem_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 mem_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.
mem_search is provided by the Crossmem MCP server (jamiemoles/crossmem). 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 →