ask_operator: any agent can call it, answer routes by asker

new AgentRequest::AskOperator + AgentResponse::QuestionQueued on
the per-agent socket — same shape as the manager flavor, agent
gets the same wire surface (still uses the same operator_questions
table). agent_server::dispatch wires AskOperator through coord
.questions.submit(agent, ...) so the row's asker is the sub-agent
name; the ttl watchdog already in manager_server gets shared and
spawn_question_watchdog goes pub.

answer routing: operator_questions::answer now returns (question,
asker). post_answer_question + post_cancel_question + the watchdog
fire OperatorAnswered through new coord.notify_agent(asker, event)
instead of always notify_manager — the event lands in whichever
agent originally asked. notify_manager is now a thin wrapper.

agent socket plumbing: agent_server::start takes Arc<Coordinator>
instead of Arc<Broker> so dispatch has access to questions +
notify path; coordinator::{register_agent,ensure_runtime} take
self: &Arc<Self>. mcp::AgentServer grows the ask_operator tool;
allowed_mcp_tools(Agent) adds it; prompts/agent.md replaces the
'message the manager to ask the operator' guidance with the
direct tool description.

hive-ag3nt/src/bin/hive-ag3nt.rs

@@ -125,7 +125,10 @@ async fn serve(
                 bus.set_state(TurnState::Idle);
             }
             Ok(AgentResponse::Empty) => {}
-            Ok(AgentResponse::Ok | AgentResponse::Status { .. } | AgentResponse::Recent { .. }) => {
+            Ok(AgentResponse::Ok
+            | AgentResponse::Status { .. }
+            | AgentResponse::Recent { .. }
+            | AgentResponse::QuestionQueued { .. }) => {
                 tracing::warn!("recv produced unexpected response kind");
             }
             Ok(AgentResponse::Err { message }) => {