Secure AI Agent Tools at the Server Boundary: Authorization, Idempotency, and Audit Trails A developer outlined a server-boundary design for securing AI agent tool calls, arguing that model-generated explanations and tool schemas are not authorization decisions. The approach carries a trusted principal alongside model-supplied arguments, resolves the resource and checks tenant ownership and permissions before any mutation, and pairs writes with durable audit trails, idempotency and human approval bound to the concrete operation. The writeup recommends narrow, purpose-specific tools over universal ones like run_sql or execute_command and treating every model-produced argument as untrusted input. An agent can produce a convincing explanation for an action. That explanation is not an authorization decision. This distinction matters as soon as a model can call tools that change state: create a ticket, update a record, issue a refund, deploy a service, or send a message. A prompt can describe the intended workflow, and the model can select the right function, while the application still needs to prove that this caller may perform this exact operation on this exact resource. A safe design puts that proof in ordinary application code at the tool boundary. The model proposes an action. A server-side policy decides whether the action is allowed, and the domain service performs it with a durable audit trail and retry semantics. A tool schema limits the shape of a request. It does not establish who the caller is, which tenant owns a record, or whether a requested change is permitted. The same is true of system prompts and tool descriptions. They influence model behavior, but they are not enforceable access controls. For example, an agent might receive a tool input like this: js const toolInput = { operation: "close case", caseId: "case 781", reason: "resolved" }; The caller identity should come from the authenticated application request, not from a field the model can fill. If a tool accepts userId, tenantId, or role as model-supplied values and trusts them, the tool has delegated identity to generated text. Instead, carry a trusted principal alongside the model-generated arguments: type Principal = { subject: string; tenantId: string; permissions: ReadonlySet