diff --git a/src/daemon/ticker.py b/src/daemon/ticker.py
index 5576488..38880ba 100644
--- a/src/daemon/ticker.py
+++ b/src/daemon/ticker.py
@@ -640,7 +640,7 @@ class Ticker:
         reclaimed: List[str] = []
         now = datetime.utcnow()  # UTC，与 SQLite datetime('now') 一致
 
-        # claimed 超时 → 重置为 pending
+        # claimed 超时 → 重置为 pending（如果 retry_count >= 3 则升级庞统）
         claimed = queries.tasks_by_status("claimed")
         for task in claimed:
             if not task.claimed_at:
@@ -649,20 +649,37 @@ class Ticker:
                 claimed_time = datetime.fromisoformat(task.claimed_at)
                 elapsed = (now - claimed_time).total_seconds() / 60.0
                 if elapsed > self.claim_timeout_minutes:
-                    conn = get_connection(db_path)
-                    try:
-                        ok = self._transition_status(
-                            conn, task.id, "pending",
-                            agent="daemon",
-                            detail={"reason": "claim_timeout",
-                                    "elapsed_minutes": round(elapsed, 1)},
-                        )
-                        if ok:
+                    retry_count = getattr(task, 'retry_count', 0) or 0
+                    # 司马懿建议：复用 retry_count，>=3 则升级庞统
+                    if retry_count >= 3:
+                        conn = get_connection(db_path)
+                        try:
+                            self._transition_status(
+                                conn, task.id, "escalated",
+                                agent="daemon",
+                                detail={"reason": "claim_timeout_no_taker",
+                                        "retry_count": retry_count},
+                            )
                             reclaimed.append(task.id)
-                            logger.info("Reclaimed %s: claimed → pending (timeout %.1fm)",
-                                        task.id, elapsed)
-                    finally:
-                        conn.close()
+                            logger.warning("Escalated %s: no taker after %d broadcasts",
+                                            task.id, retry_count)
+                        finally:
+                            conn.close()
+                    else:
+                        conn = get_connection(db_path)
+                        try:
+                            ok = self._transition_status(
+                                conn, task.id, "pending",
+                                agent="daemon",
+                                detail={"reason": "claim_timeout",
+                                        "elapsed_minutes": round(elapsed, 1)},
+                            )
+                            if ok:
+                                reclaimed.append(task.id)
+                                logger.info("Reclaimed %s: claimed → pending (timeout %.1fm)",
+                                            task.id, elapsed)
+                        finally:
+                            conn.close()
             except (ValueError, TypeError):
                 pass