agentgate

The security gateway for AI agents. Enforce authentication, rate limiting, and policy controls between users and AI agents — in one line of code.

Install

npm install agentgate

Quickstart

Get your free API key at https://agent-gate-rho.vercel.app/

import { withAgentGate } from "agentgate";

const secured = withAgentGate(myAgent, {
  apiKey: "ag_your_key_here",
});

// Use exactly like before — every tool call is now policy-checked
await secured.executeTool("delete_file", { path: "/etc/passwd" });

MCP middleware

import { createAgentGateMiddleware } from "agentgate";

const gate = createAgentGateMiddleware({ apiKey: "ag_your_key_here" });

server.setRequestHandler(CallToolRequestSchema, async (req) => {
  return gate(req.params.name, req.params.arguments, () =>
    myToolHandler(req)
  );
});

Fail-open vs Fail-closed

// Development default — allow if AgentGate unreachable
withAgentGate(agent, { apiKey: "ag_xxx", onNetworkError: "fail-open" });

// Production recommended — block if AgentGate unreachable
withAgentGate(agent, { apiKey: "ag_xxx", onNetworkError: "fail-closed" });

Policy decisions

| Decision | Behaviour | | -------- | --------- | | ALLOW | Tool executes normally | | DENY | Tool blocked, error returned to agent | | REQUIRE_APPROVAL | Tool paused, human approves in dashboard |

AgentGate Connect (multi-tenant)

If you are a SaaS platform managing agents for your customers, use Connect sub-keys (agc_ prefix) to enforce per-customer policies.

Links

Dashboard: https://agent-gate-rho.vercel.app/ GitHub: https://github.com/wiserautomation/agentgate Issues: https://github.com/wiserautomation/agentgate/issues