Semantic search using three-axis scoring (recency + importance + relevance). Updates last_accessed_at on returned results.
AI agents call agent_search to retrieve information from Mybrain 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 personal knowledge base without creating, modifying irreversibly, or deleting substantive content. The metadata update (last_accessed_at) is a standard read-side-effect and does not constitute destructive or write-category action. The three-axis scoring (recency, importance, relevance) is part of the retrieval mechanism.
From the tool's definition Tool name 'agent_search' and description indicate it performs 'semantic search' and 'returns results'. The only side effect mentioned is updating 'last_accessed_at' metadata on returned results, which is a benign timestamp update typical of access logging.
Documented attack patterns abuse exactly the kind of access agent_search gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Mybrain, and nothing reaches the server without passing your rules. This is the rule we recommend for agent_search:
{
"version": "1",
"default": "deny",
"tools": {
"agent_search": {}
}
}agent_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.
Semantic search using three-axis scoring (recency + importance + relevance). Updates last_accessed_at on returned results. It is categorised as a Read tool in the Mybrain MCP Server, which means it retrieves data without modifying state.
Register the Mybrain MCP server in PolicyLayer and add a rule for agent_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 Mybrain. Nothing to install.
agent_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 agent_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 agent_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.
agent_search is provided by the Mybrain MCP server (robertsfeir/mybrain). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from Mybrain, 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.
8 Mybrain tools catalogued and risk-classified — across an index of 43,000+ MCP servers.