diff --git a/apps/api/src/services/decision_manager.py b/apps/api/src/services/decision_manager.py
index 999a7064..d774a080 100644
--- a/apps/api/src/services/decision_manager.py
+++ b/apps/api/src/services/decision_manager.py
@@ -1527,14 +1527,32 @@ class DecisionManager:
             logger.debug("target_rescue_skipped", error=str(_rescue_err))
 
         # ADR-073 Phase 3-2: infrastructure 告警 (Docker/Host) → SSH MCP routing (2026-04-12 ogt)
-        # alert_category = "infrastructure" 表示 Docker/Host 告警，不走 K8s executor
-        # action 格式應為 "docker restart <container>" 或 "systemctl restart <service>"
+        # alert_category = "infrastructure" 表示 Docker 告警，非 kubectl action → SSH
         # P1-1 fix 2026-04-12: 必須在 kubectl safety guard 之前 routing，否則 docker 指令被 _action_safe=False 攔截
         _alert_category = getattr(incident, "alert_category", None) or ""
         if _alert_category == "infrastructure" and action and not action.startswith("kubectl"):
             await self._ssh_execute(incident, token, action, _target)
             return
 
+        # 2026-04-15 ogt: host_resource 告警（HostHighCpuLoad 等）不是 K8s workload 問題
+        # 不得執行 kubectl 操作，改降級人工審核
+        # 根因：原本只擋了 infrastructure，忘記 host_resource 也不走 K8s
+        if _alert_category == "host_resource" and action and action.startswith("kubectl"):
+            logger.warning(
+                "auto_execute_blocked_host_resource_no_k8s",
+                incident_id=incident.incident_id,
+                alert_category=_alert_category,
+                action=action[:80],
+                reason="host_resource 告警不應執行 K8s kubectl 操作，降級人工審核",
+            )
+            token.state = DecisionState.READY
+            token.proposal_data["auto_executed"] = False
+            token.proposal_data["mcp_all_failed"] = True
+            token.proposal_data["blocked_reason"] = "host_resource 告警禁止 K8s kubectl，請人工排查主機"
+            await self._save_token(token)
+            _fire_and_forget(_push_decision_to_telegram(incident, token.proposal_data))
+            return
+
         # 安全守衛: 替換後仍含 "unknown" 或未替換的 <...>/{...} → 拒絕執行
         # 另外：若 target 等於 alertname，代表 LLM 把告警名稱填入 deployment_name，也拒絕
         _alertname = incident.signals[0].labels.get("alertname", "") if incident.signals else ""
@@ -1667,6 +1685,32 @@ class DecisionManager:
             _fire_and_forget(_push_decision_to_telegram(incident, token.proposal_data))
             return
 
+        # 2026-04-15 ogt: 同一 target 5 分鐘內最多執行 2 次，防止修復風暴
+        # 根因：多個 incident 共享同一 target 時，各自獨立自動執行 → 重複重啟
+        try:
+            from src.core.redis_client import get_redis as _get_redis_dm
+            _redis_dm = _get_redis_dm()
+            _dm_cooldown_key = f"awoooi:auto_execute_cooldown:{_ns}:{_target}"
+            _dm_exec_count = await _redis_dm.get(_dm_cooldown_key)
+            if _dm_exec_count and int(_dm_exec_count) >= 2:
+                logger.warning(
+                    "auto_execute_cooldown_blocked",
+                    incident_id=incident.incident_id,
+                    target=_target,
+                    namespace=_ns,
+                    exec_count=int(_dm_exec_count),
+                    reason="同一 target 5 分鐘內已自動執行 2 次，冷卻中",
+                )
+                token.state = DecisionState.READY
+                token.proposal_data["auto_executed"] = False
+                token.proposal_data["cooldown_blocked"] = True
+                await self._save_token(token)
+                return
+            await _redis_dm.incr(_dm_cooldown_key)
+            await _redis_dm.expire(_dm_cooldown_key, 300)  # 5 分鐘
+        except Exception as _cd_err:
+            logger.debug("auto_execute_cooldown_check_error", error=str(_cd_err))
+
         try:
             # 延遲導入避免循環依賴
             from src.models.approval import ApprovalRequest, ApprovalStatus