diff --git a/agent-office/app/agents/lotto.py b/agent-office/app/agents/lotto.py
index f6aa461..2301fc1 100644
--- a/agent-office/app/agents/lotto.py
+++ b/agent-office/app/agents/lotto.py
@@ -30,8 +30,7 @@ class LottoAgent(BaseAgent):
     async def run_signal_check(self, source: str = "light") -> dict:
         """비-LLM 시그널 평가 (light/sim) 또는 deep_check (LLM 호출 후).
 
-        Phase 2: 텔레그램 발송 안 함. lotto_signals INSERT만.
-        Phase 3 (Task 9): urgent 시그널 텔레그램 발송 + throttle 추가.
+        Phase 3 (Task 9): urgent 시그널 텔레그램 발송 + throttle/daily-cap 추가.
         """
         from ..curator.signal_runner import run_signal_check
         from ..config import LOTTO_Z_NORMAL, LOTTO_Z_URGENT
@@ -62,17 +61,86 @@ class LottoAgent(BaseAgent):
                 self.agent_id,
                 f"signal_check({source}) → overall={outcome['overall_fire']} results={len(outcome['results'])}",
             )
+
+            # --- Throttle + 텔레그램 urgent 발송 ---
+            from ..config import LOTTO_THROTTLE_HOURS, LOTTO_URGENT_DAILY_MAX
+            from ..db import (
+                get_last_signal_notification, get_recent_urgent_count,
+                mark_signal_notified,
+            )
+            from ..notifiers.telegram_lotto import send_urgent_signal
+
+            if outcome["overall_fire"] == "urgent":
+                if get_recent_urgent_count(hours=24) >= LOTTO_URGENT_DAILY_MAX:
+                    add_log(
+                        self.agent_id,
+                        "urgent daily cap 도달 → normal로 강등 (digest 합류)",
+                        level="warning",
+                    )
+                else:
+                    blocked = False
+                    for r in outcome["results"]:
+                        if r["fire_level"] in ("normal", "urgent"):
+                            last = get_last_signal_notification(
+                                metric=r["metric"], fire_level=r["fire_level"],
+                                hours=LOTTO_THROTTLE_HOURS,
+                            )
+                            if last:
+                                blocked = True
+                                break
+                    if not blocked:
+                        from datetime import datetime, timezone
+                        event = {
+                            "fire_level": "urgent",
+                            "triggered_at": datetime.now(timezone.utc).isoformat(),
+                            "results": outcome["results"],
+                        }
+                        await send_urgent_signal(event)
+                        for r in outcome["results"]:
+                            if r["fire_level"] in ("normal", "urgent"):
+                                mark_signal_notified(r["signal_id"])
+                        add_log(self.agent_id, f"urgent 텔레그램 발송 완료 (시그널 {len(outcome['results'])}개 마킹)")
+
             return {"ok": True, **outcome}
         except Exception as e:
             add_log(self.agent_id, f"signal_check 예외: {e}", level="error")
             return {"ok": False, "message": f"{type(e).__name__}: {e}"}
 
     async def run_daily_digest(self) -> dict:
-        """Phase 2: 발화 카운트만 반환. Phase 3 (Task 9)에서 텔레그램 발송 추가."""
-        from ..db import get_recent_lotto_signals, add_log
+        """일일 요약 — 지난 24h normal/urgent 발화를 묶어 텔레그램 1통."""
+        from ..db import (
+            get_recent_lotto_signals, get_signals_history, add_log,
+            get_baseline,
+        )
+        from ..notifiers.telegram_lotto import send_signal_summary
+
         sigs = get_recent_lotto_signals(hours=24, min_fire="normal")
-        add_log(self.agent_id, f"daily_digest: 지난 24h 발화 {len(sigs)}건")
-        return {"ok": True, "count": len(sigs), "signals": sigs}
+        total_24h = get_signals_history(days=1)
+        evaluated = len(total_24h)
+
+        # weights_trend: drift_weights_cache의 prev/curr 차이
+        trend = {}
+        try:
+            cache = get_baseline("drift_weights_cache")
+            if cache and isinstance(cache["window_values"], list) and len(cache["window_values"]) >= 2:
+                prev_w = cache["window_values"][-2]
+                curr_w = cache["window_values"][-1]
+                trend = {
+                    k: curr_w.get(k, 0.0) - prev_w.get(k, 0.0)
+                    for k in (set(prev_w) | set(curr_w))
+                }
+        except Exception as e:
+            add_log(self.agent_id, f"weights_trend 계산 실패: {e}", level="warning")
+
+        digest = {
+            "evaluated": evaluated,
+            "fired": len(sigs),
+            "signals": sigs,
+            "weights_trend": trend,
+        }
+        await send_signal_summary(digest)
+        add_log(self.agent_id, f"daily_digest 발송: 평가 {evaluated} / 발화 {len(sigs)}")
+        return {"ok": True, **digest}
 
     async def _run(self, source: str) -> dict:
         task_id = create_task(self.agent_id, "curate_weekly", {"source": source})
diff --git a/agent-office/app/curator/signal_runner.py b/agent-office/app/curator/signal_runner.py
index 013cca6..e7775f3 100644
--- a/agent-office/app/curator/signal_runner.py
+++ b/agent-office/app/curator/signal_runner.py
@@ -80,6 +80,8 @@ def evaluate_metric_and_persist(
         "signal_id": sid,
         "metric": metric,
         "value": value,
+        "baseline_mu": bl.mu if bl.size > 0 else None,
+        "baseline_sigma": bl.sigma if bl.size >= 2 else None,
         "z_score": z,
         "fire_level": fire,
         "payload": payload or {},
diff --git a/agent-office/app/notifiers/telegram_lotto.py b/agent-office/app/notifiers/telegram_lotto.py
index f900c7b..8673e51 100644
--- a/agent-office/app/notifiers/telegram_lotto.py
+++ b/agent-office/app/notifiers/telegram_lotto.py
@@ -88,10 +88,11 @@ def _format_urgent_signal(event: Dict[str, Any]) -> str:
         mu = r.get("baseline_mu")
         sigma = r.get("baseline_sigma")
         z = r.get("z_score")
+        v_text = f"{v:.2f}" if v is not None else "N/A"
         if mu is not None and sigma is not None and z is not None:
-            lines.append(f"• {label} {v:.2f} (μ={mu:.2f}, σ={sigma:.2f}) z={z:.1f}")
+            lines.append(f"• {label} {v_text} (μ={mu:.2f}, σ={sigma:.2f}) z={z:.1f}")
         else:
-            lines.append(f"• {label} {v:.2f}")
+            lines.append(f"• {label} {v_text}")
 
     # drift 페이로드 — 어떤 전략이 변동했는지 한 줄
     for r in fired: