diff --git a/app/agent/policy/mcp.py b/app/agent/policy/mcp.py
index 2d9ea150f..3fa8836ac 100644
--- a/app/agent/policy/mcp.py
+++ b/app/agent/policy/mcp.py
@@ -396,6 +396,8 @@ FIELD_DESCRIPTIONS = {
     "sample_rate": "Recorded audio sample rate in hertz.",
     "save_path": "Configured downloader-side save path for the download or subscription.",
     "scrape": "Generate metadata and images after manual transfer.",
+    "search_interval": "Scheduled search interval in whole hours (1-8760); null uses the system interval.",
+    "last_search": "Read-only UTC timestamp of the most recent subscription search attempt.",
     "search_imdbid": "Use IMDb identity during subscription search when set to 1.",
     "season": "Season number used by the media, search, subscription, or transfer operation.",
     "seasons": "Season-number expression recorded in history.",
diff --git a/app/agent/policy/resources/api_mcp_schema.json b/app/agent/policy/resources/api_mcp_schema.json
index 120c28c0a..253072a09 100644
--- a/app/agent/policy/resources/api_mcp_schema.json
+++ b/app/agent/policy/resources/api_mcp_schema.json
@@ -2862,6 +2862,18 @@
           "description": "Number of episodes still missing from the subscription.",
           "title": "Lack Episode"
         },
+        "last_search": {
+          "anyOf": [
+            {
+              "type": "string"
+            },
+            {
+              "type": "null"
+            }
+          ],
+          "description": "Read-only UTC timestamp of the most recent subscription search attempt.",
+          "title": "Last Search"
+        },
         "last_update": {
           "anyOf": [
             {
@@ -3057,6 +3069,20 @@
           "description": "Use IMDb identity during subscription search when set to 1.",
           "title": "Search Imdbid"
         },
+        "search_interval": {
+          "anyOf": [
+            {
+              "maximum": 8760.0,
+              "minimum": 1.0,
+              "type": "integer"
+            },
+            {
+              "type": "null"
+            }
+          ],
+          "description": "Scheduled search interval in whole hours (1-8760); null uses the system interval.",
+          "title": "Search Interval"
+        },
         "season": {
           "anyOf": [
             {
diff --git a/app/application/subscription/contract.py b/app/application/subscription/contract.py
index 871c67d94..f33392ad8 100644
--- a/app/application/subscription/contract.py
+++ b/app/application/subscription/contract.py
@@ -21,6 +21,8 @@ _SUBSCRIPTION_FIELDS = frozenset(
         "year",
         "type",
         "keyword",
+        "search_interval",
+        "last_search",
         "media_source",
         "media_id",
         "music_type",
@@ -83,6 +85,7 @@ _SUBSCRIPTION_HISTORY_FIELDS = (_SUBSCRIPTION_FIELDS - {
     "note",
     "state",
     "last_update",
+    "last_search",
     "downloader",
     "manual_total_episode",
 }) | _CLASSIFICATION_HISTORY_FIELDS
@@ -153,6 +156,7 @@ class SubscriptionSnapshot:
     year: Optional[str] = None
     type: Optional[str] = None
     keyword: Optional[str] = None
+    search_interval: Optional[int] = None
     media_source: Optional[MediaSource] = None
     media_id: Optional[str] = None
     music_type: Optional[str] = None
@@ -179,6 +183,7 @@ class SubscriptionSnapshot:
     note: Optional[builtins.list[int]] = None
     state: str = "N"
     last_update: Optional[str] = None
+    last_search: Optional[str] = None
     date: Optional[str] = None
     username: Optional[str] = None
     sites: Optional[builtins.list[int]] = None
@@ -219,6 +224,7 @@ class SubscriptionHistorySnapshot:
     year: Optional[str] = None
     type: Optional[str] = None
     keyword: Optional[str] = None
+    search_interval: Optional[int] = None
     media_source: Optional[MediaSource] = None
     media_id: Optional[str] = None
     music_type: Optional[str] = None
diff --git a/app/application/subscription/query.py b/app/application/subscription/query.py
index 9b5f51727..9442ab028 100644
--- a/app/application/subscription/query.py
+++ b/app/application/subscription/query.py
@@ -3,6 +3,7 @@
 from __future__ import annotations
 
 from collections.abc import Mapping
+from datetime import datetime, timedelta
 from typing import Optional
 
 from app.application.subscription.contract import (
@@ -253,3 +254,25 @@ class SubscriptionQueryService:
         return any(
             subscribe.type == MediaType.MUSIC.value for subscribe in self._repository.list(searchable_states) or []
         )
+
+
+def subscription_search_due(
+    subscribe: SubscriptionSnapshot,
+    default_interval: int,
+    now: datetime,
+) -> bool:
+    """按最近一次主动搜索计算定时搜索到期；旧记录以本地创建时间起算。"""
+    if subscribe.state not in {"R", "P"}:
+        return False
+    previous = subscribe.last_search or subscribe.date
+    if not previous:
+        return True
+    try:
+        started = datetime.fromisoformat(previous)
+    except ValueError:
+        return True
+    # 旧创建时间没有时区，保持其本地时间语义；新搜索时间始终携带 UTC 时区。
+    if started.tzinfo is None:
+        started = started.astimezone()
+    interval = subscribe.search_interval or default_interval
+    return now >= started + timedelta(hours=interval)
diff --git a/app/chain/subscribe/search.py b/app/chain/subscribe/search.py
index 4ec69d975..2c405b2a2 100644
--- a/app/chain/subscribe/search.py
+++ b/app/chain/subscribe/search.py
@@ -28,7 +28,7 @@ from app.application.subscription.observability import (
     batch_progress_text,
     inline_search_result,
 )
-from app.application.subscription.query import SubscriptionQueryService
+from app.application.subscription.query import SubscriptionQueryService, subscription_search_due
 from app.application.subscription.sitebudget import (
     SubscriptionSearchCancelled,
     SubscriptionSearchDeferred,
@@ -160,11 +160,12 @@ class _SubscribeSearchQueueCoordinator(_SubscribeOwnerBase):
         manual: Optional[bool] = False,
         progress_callback: Optional[Callable[..., None]] = None,
         sids: Optional[tuple[int, ...]] = None,
+        scheduled_interval: Optional[int] = None,
     ) -> Optional[str]:
         """
         执行订阅搜索。
 
-        保持定时任务、API 和插件使用的公开签名，搜索实现委托给内部执行阶段。
+        scheduled_interval 仅供定时调度传入系统间隔；手动和指定目标搜索不受周期限制。
         """
         return self._execute_search(
             sid=sid,
@@ -172,6 +173,7 @@ class _SubscribeSearchQueueCoordinator(_SubscribeOwnerBase):
             state=state,
             manual=manual,
             progress_callback=progress_callback,
+            scheduled_interval=scheduled_interval,
         )
 
     def _execute_search(
@@ -181,6 +183,7 @@ class _SubscribeSearchQueueCoordinator(_SubscribeOwnerBase):
         manual: Optional[bool] = False,
         progress_callback: Optional[Callable[..., None]] = None,
         sids: Optional[tuple[int, ...]] = None,
+        scheduled_interval: Optional[int] = None,
     ) -> Optional[str]:
         """
         订阅搜索
@@ -202,6 +205,7 @@ class _SubscribeSearchQueueCoordinator(_SubscribeOwnerBase):
                 state=state,
                 manual=manual,
                 progress_callback=progress_callback,
+                scheduled_interval=scheduled_interval,
             )
         self._execute_inline_search(
             sid=sid,
@@ -209,6 +213,7 @@ class _SubscribeSearchQueueCoordinator(_SubscribeOwnerBase):
             state=state,
             manual=manual,
             progress_callback=progress_callback,
+            scheduled_interval=scheduled_interval,
         )
         return None
 
@@ -219,6 +224,7 @@ class _SubscribeSearchQueueCoordinator(_SubscribeOwnerBase):
         state: Optional[str],
         manual: Optional[bool],
         progress_callback: Optional[Callable[..., None]],
+        scheduled_interval: Optional[int] = None,
     ) -> None:
         """在独立 Search 通道内按订阅准入执行兼容搜索。"""
         lock_acquired = self._acquire_run_lock("search", progress_callback)
@@ -228,7 +234,10 @@ class _SubscribeSearchQueueCoordinator(_SubscribeOwnerBase):
         processed = []
         summary: Optional[SearchExecutionSummary] = None
         try:
-            subscribes = self._load_search_subscriptions(sid=sid, sids=sids, state=state)
+            subscribes = self._load_search_subscriptions(
+                sid=sid, sids=sids, state=state,
+                scheduled_interval=None if manual else scheduled_interval,
+            )
             total = len(subscribes)
             source, _priority = _search_source_and_priority(
                 sid=sid,
@@ -352,9 +361,17 @@ class _SubscribeSearchQueueCoordinator(_SubscribeOwnerBase):
         state: Optional[str],
         manual: Optional[bool],
         progress_callback: Optional[Callable[..., None]],
-    ) -> str:
+        scheduled_interval: Optional[int] = None,
+    ) -> Optional[str]:
         """将搜索转为持久任务并在无 Match 长锁的短租约中串行消费。"""
-        subscribes = self._load_search_subscriptions(sid=sid, sids=sids, state=state)
+        subscribes = self._load_search_subscriptions(
+            sid=sid, sids=sids, state=state,
+            scheduled_interval=None if manual else scheduled_interval,
+        )
+        if scheduled_interval is not None and not subscribes:
+            if progress_callback:
+                progress_callback(value=100, text="暂无到期订阅，无需搜索")
+            return None
         source, priority = _search_source_and_priority(
             sid=sid,
             sids=sids,
@@ -573,6 +590,7 @@ class SubscribeSearchOwner(_SubscribeSearchQueueOwner):
         sid: Optional[int],
         sids: Optional[tuple[int, ...]],
         state: Optional[str],
+        scheduled_interval: Optional[int] = None,
     ) -> list[SubscriptionSnapshot]:
         """按单条、指定批次或状态读取本轮搜索订阅。"""
         repository = self.subscription_repository
@@ -581,7 +599,11 @@ class SubscribeSearchOwner(_SubscribeSearchQueueOwner):
             return [subscribe] if subscribe else []
         if sids is not None:
             return [item for current_id in sids if (item := repository.get(current_id)) is not None]
-        return cast(list[SubscriptionSnapshot], repository.list(self.get_states_for_search(state or "N")))
+        subscribes = cast(list[SubscriptionSnapshot], repository.list(self.get_states_for_search(state or "N")))
+        if scheduled_interval is None:
+            return subscribes
+        now = datetime.now(timezone.utc)
+        return [item for item in subscribes if subscription_search_due(item, scheduled_interval, now)]
 
     @staticmethod
     def _recent_subscription_retry_at(
@@ -682,6 +704,11 @@ class SubscribeSearchOwner(_SubscribeSearchQueueOwner):
     ) -> Optional[SubscriptionSnapshot]:
         """处理单个订阅，并返回下载后重新读取的状态快照。"""
         _ensure_execution_active(execution_context)
+        subscribe = self._SubscribeChain__apply_subscribe_update(
+            subscribe,
+            {"last_search": datetime.now(timezone.utc).isoformat(timespec="seconds")},
+            scene="search",
+        )
         logger.debug(f"开始搜索订阅，标题：{subscribe.name} ...")
         target = prepare_search_target(
             self, subscribe, MediaChain(), partial(_ensure_execution_active, execution_context),
diff --git a/app/db/adapters/subscription.py b/app/db/adapters/subscription.py
index 6db8d42e1..131512306 100644
--- a/app/db/adapters/subscription.py
+++ b/app/db/adapters/subscription.py
@@ -77,6 +77,7 @@ def _project_subscription(record: Subscribe) -> SubscriptionSnapshot:
         year=record.year,
         type=record.type,
         keyword=record.keyword,
+        search_interval=record.search_interval,
         media_source=_media_source(record.media_source),
         media_id=record.media_id,
         music_type=record.music_type,
@@ -103,6 +104,7 @@ def _project_subscription(record: Subscribe) -> SubscriptionSnapshot:
         note=cast(Optional[builtins.list[int]], record.note),
         state=record.state,
         last_update=record.last_update,
+        last_search=record.last_search,
         date=record.date,
         username=record.username,
         sites=cast(Optional[builtins.list[int]], record.sites),
@@ -134,6 +136,7 @@ def _project_history(record: SubscribeHistory) -> SubscriptionHistorySnapshot:
         year=record.year,
         type=record.type,
         keyword=record.keyword,
+        search_interval=record.search_interval,
         media_source=_media_source(record.media_source),
         media_id=record.media_id,
         music_type=record.music_type,
diff --git a/app/db/models/subscribe.py b/app/db/models/subscribe.py
index fcafe32cf..27da6026d 100644
--- a/app/db/models/subscribe.py
+++ b/app/db/models/subscribe.py
@@ -22,6 +22,8 @@ class Subscribe(Base):
     year: Mapped[Optional[str]] = mapped_column(String)
     # 类型
     type: Mapped[Optional[str]] = mapped_column(String)
+    # 自定义定时搜索间隔（小时）；空值跟随系统
+    search_interval: Mapped[Optional[int]] = mapped_column(Integer)
     # 搜索关键字
     keyword: Mapped[Optional[str]] = mapped_column(String)
     media_source: Mapped[Optional[str]] = mapped_column(String, index=True)
@@ -72,6 +74,8 @@ class Subscribe(Base):
     note: Mapped[Optional[Any]] = mapped_column(JSON)
     # 状态：N-新建 R-订阅中 P-待定 S-暂停
     state: Mapped[str] = mapped_column(String, nullable=False, index=True, default="N")
+    # 最近一次主动搜索开始时间，使用带时区的 UTC 时间
+    last_search: Mapped[Optional[str]] = mapped_column(String)
     # 最后更新时间
     last_update: Mapped[Optional[str]] = mapped_column(String)
     # 创建时间
diff --git a/app/db/models/subscribehistory.py b/app/db/models/subscribehistory.py
index f4ee645c9..943292904 100644
--- a/app/db/models/subscribehistory.py
+++ b/app/db/models/subscribehistory.py
@@ -20,6 +20,8 @@ class SubscribeHistory(Base):
     year: Mapped[Optional[str]] = mapped_column(String)
     # 类型
     type: Mapped[Optional[str]] = mapped_column(String)
+    # 自定义定时搜索间隔（小时）；空值跟随系统
+    search_interval: Mapped[Optional[int]] = mapped_column(Integer)
     # 搜索关键字
     keyword: Mapped[Optional[str]] = mapped_column(String)
     media_source: Mapped[Optional[str]] = mapped_column(String, index=True)
diff --git a/app/locales/en-US.json b/app/locales/en-US.json
index 3e59ac9ce..40213b0d7 100644
--- a/app/locales/en-US.json
+++ b/app/locales/en-US.json
@@ -103,6 +103,7 @@
     }
   },
   "messages": {
+    "暂无到期订阅，无需搜索": "No subscriptions are due for search",
     "调用工具失败": "Tool call failed",
     "无效的媒体来源": "Invalid media source",
     "该媒体来源不支持此音乐接口": "This media source is not supported by this music endpoint",
diff --git a/app/locales/zh-TW.json b/app/locales/zh-TW.json
index 12d1e11ab..eb8e4b422 100644
--- a/app/locales/zh-TW.json
+++ b/app/locales/zh-TW.json
@@ -103,6 +103,7 @@
     }
   },
   "messages": {
+    "暂无到期订阅，无需搜索": "暫無到期訂閱，無需搜尋",
     "调用工具失败": "調用工具失敗",
     "媒体来源和媒体 ID 必须同时提供": "媒體來源和媒體 ID 必須同時提供",
     "media_source 和 media_id 必须同时提供": "media_source 和 media_id 必須同時提供",
diff --git a/app/scheduler/catalog.py b/app/scheduler/catalog.py
index 320a2c0bb..8fab04323 100644
--- a/app/scheduler/catalog.py
+++ b/app/scheduler/catalog.py
@@ -36,11 +36,14 @@ class _MediaServerSchedule(TypedDict):
     interval: int
 
 
-def _subscription_search_job_specs(services: SchedulerServices) -> tuple[JobSpec, ...]:
+def _subscription_search_job_specs(
+    services: SchedulerServices, search_interval: int = 24,
+) -> tuple[JobSpec, ...]:
     """构造订阅搜索、新增搜索与持久队列恢复任务目录。"""
     return (
         JobSpec(
-            "subscribe_search", "订阅搜索补全", services.search_subscribe, "subscription", kwargs={"state": "R"}
+            "subscribe_search", "订阅搜索补全", services.search_subscribe, "subscription",
+            kwargs={"state": "R", "scheduled_interval": search_interval}
         ),
         JobSpec(
             "new_subscribe_search",
@@ -168,7 +171,7 @@ class SchedulerCatalogOwner(_SchedulerOwnerBase):
                 JobSpec("cookiecloud", "同步CookieCloud站点", services.sync_cookies, "site"),
                 JobSpec("mediaserver_sync", "同步媒体服务器", services.sync_mediaserver, "mediaserver"),
                 JobSpec("subscribe_tmdb", "订阅元数据更新", services.check_subscribe, "subscription"),
-                *_subscription_search_job_specs(services),
+                *_subscription_search_job_specs(services, config.subscribe_search_interval),
                 JobSpec("subscribe_refresh", "订阅刷新", services.refresh_subscribe, "subscription"),
                 JobSpec("subscribe_follow", "关注的订阅分享", services.follow_subscribe, "subscription"),
                 JobSpec(
@@ -298,14 +301,14 @@ class SchedulerCatalogOwner(_SchedulerOwnerBase):
             kwargs={"job_id": "subscribe_tmdb"},
         )
 
-        # 订阅状态每隔24小时搜索一次
+        # 每五分钟检查逐条订阅到期时间；实际搜索仍受系统或独立周期约束。
         if config.subscribe_search:
             self._scheduler.add_job(
                 self.start,
                 "interval",
                 id="subscribe_search",
                 name="订阅搜索补全",
-                hours=config.subscribe_search_interval,
+                minutes=5,
                 kwargs={"job_id": "subscribe_search"},
             )
 
diff --git a/app/schemas/subscribe.py b/app/schemas/subscribe.py
index 935c78132..c04b8e4f6 100644
--- a/app/schemas/subscribe.py
+++ b/app/schemas/subscribe.py
@@ -111,7 +111,7 @@ class Subscribe(OptionalMediaIdentityMixin, BaseModel):
         "note", "state", "last_update", "username", "current_priority", "episode_priority", "date",
         "current_audio_format", "current_bitrate", "current_bit_depth", "current_sample_rate",
         "classification_rule_id", "classification_policy_revision", "classification_source",
-        "execution_status",
+        "execution_status", "last_search",
     })
 
     id: Optional[int] = None
@@ -121,6 +121,10 @@ class Subscribe(OptionalMediaIdentityMixin, BaseModel):
     year: Optional[str] = None
     # 订阅类型 电影/电视剧
     type: Optional[str] = None
+    # 定时搜索间隔（小时）；空值跟随系统设置
+    search_interval: Optional[int] = Field(default=None, ge=1, le=8760)
+    # 最近一次主动搜索开始时间，仅供读取
+    last_search: Optional[str] = None
     # 搜索关键字
     keyword: Optional[str] = None
     media_source: Optional[MediaSource] = None
@@ -262,7 +266,7 @@ class Subscribe(OptionalMediaIdentityMixin, BaseModel):
             data = dict(data)
             for key, value in list(data.items()):
                 if isinstance(value, str) and value == "":
-                    if key in {"media_source", "media_id"} or key in cls.CLEARABLE_FILTER_FIELDS:
+                    if key in {"media_source", "media_id", "search_interval"} or key in cls.CLEARABLE_FILTER_FIELDS:
                         data[key] = None
                     else:
                         data.pop(key)
diff --git a/database/versions/f8a2c6e9b1d4_3_0_30.py b/database/versions/f8a2c6e9b1d4_3_0_30.py
new file mode 100644
index 000000000..44defbc8b
--- /dev/null
+++ b/database/versions/f8a2c6e9b1d4_3_0_30.py
@@ -0,0 +1,27 @@
+"""3.0.30 增加逐订阅搜索周期和持久搜索时间。"""
+
+import sqlalchemy as sa
+from alembic import op
+
+revision = "f8a2c6e9b1d4"
+down_revision = "e7f3a9c1d5b2"
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    """兼容首次建表和已有数据库，旧订阅默认跟随系统周期。"""
+    inspector = sa.inspect(op.get_bind())
+    for table in ("subscribe", "subscribehistory"):
+        columns = {column["name"] for column in inspector.get_columns(table)}
+        if "search_interval" not in columns:
+            op.add_column(table, sa.Column("search_interval", sa.Integer(), nullable=True))
+        if table == "subscribe" and "last_search" not in columns:
+            op.add_column(table, sa.Column("last_search", sa.String(), nullable=True))
+
+
+def downgrade() -> None:
+    """移除新增周期字段，保留原有订阅和历史数据。"""
+    op.drop_column("subscribe", "last_search")
+    for table in ("subscribe", "subscribehistory"):
+        op.drop_column(table, "search_interval")
diff --git a/docs/mcp-api.md b/docs/mcp-api.md
index 179904bb0..4102a9714 100644
--- a/docs/mcp-api.md
+++ b/docs/mcp-api.md
@@ -422,6 +422,17 @@ TMDB 缓存查询响应的 `data` 包含 `count`、`recognized`、`unrecognized`
 旧版未包含这些证据的派生缓存在升级后重新建立，不影响下载历史或订阅数据。
 名称确认规则更新时同样重建旧派生缓存，避免艺术家前后缀误截断的旧结果继续命中。
 
+### 单条订阅搜索周期
+
+`POST /api/v1/subscribe/` 和 `PUT /api/v1/subscribe/` 支持 `search_interval`：
+取值为 1–8760 的整数小时数，`null` 表示跟随系统的 `SUBSCRIBE_SEARCH_INTERVAL`。
+更新时省略该字段保留原值，显式传入 `null` 恢复系统周期。默认订阅规则同样可保存此字段。
+
+仅在启用 `SUBSCRIBE_SEARCH` 时执行定时搜索，每五分钟检查到期订阅，再通过原有搜索队列和站点限流执行；
+实际开始时间可能因站点忙而延后。`last_search` 为系统维护的 UTC 搜索尝试开始时间，重启后仍有效，公共写接口忽略它。
+旧订阅尚无搜索时间时，以添加时间计算到期。新订阅首次搜索、手动搜索和 RSS 刷新不受周期过滤影响；
+手动主动搜索会更新最近搜索时间。电影、电视剧和音乐均支持独立周期。
+
 ### 插件补充接口
 
 **GET** `/api/v1/plugin/history/{plugin_id}`
diff --git a/skills/database-operation/SKILL.md b/skills/database-operation/SKILL.md
index b84a3e904..4d3a6ff28 100644
--- a/skills/database-operation/SKILL.md
+++ b/skills/database-operation/SKILL.md
@@ -228,13 +228,13 @@ python scripts/mp-db.py write "UPDATE subscribe SET state = 'S' WHERE id = 123"
 - Purpose: Stores active movie, TV, or music subscriptions, filters, progress, and download targets.
 - Useful queries: Inspecting state, missing episodes/tracks, quality rules, site scope, and match progress.
 - Write boundary: Create, update, search, or delete through the subscription API to preserve state-machine consistency.
-- Columns: `id`, `name`, `year`, `type`, `keyword`, `media_source`, `media_id`, `music_type`, `total_tracks`, `season`, `poster`, `backdrop`, `vote`, `description`, `filter`, `include`, `exclude`, `quality`, `resolution`, `effect`, `audio_quality`, `audio_format`, `min_bitrate`, `min_bit_depth`, `min_sample_rate`, `total_episode`, `start_episode`, `lack_episode`, `note`, `state`, `last_update`, `date`, `username`, `sites`, `downloader`, `best_version`, `best_version_full`, `current_priority`, `current_audio_format`, `current_bitrate`, `current_bit_depth`, `current_sample_rate`, `episode_priority`, `save_path`, `search_imdbid`, `manual_total_episode`, `custom_words`, `media_category_id`, `media_category`, `filter_groups`, `episode_group`
+- Columns: `id`, `name`, `year`, `type`, `search_interval`, `keyword`, `media_source`, `media_id`, `music_type`, `total_tracks`, `season`, `poster`, `backdrop`, `vote`, `description`, `filter`, `include`, `exclude`, `quality`, `resolution`, `effect`, `audio_quality`, `audio_format`, `min_bitrate`, `min_bit_depth`, `min_sample_rate`, `total_episode`, `start_episode`, `lack_episode`, `note`, `state`, `last_search`, `last_update`, `date`, `username`, `sites`, `downloader`, `best_version`, `best_version_full`, `current_priority`, `current_audio_format`, `current_bitrate`, `current_bit_depth`, `current_sample_rate`, `episode_priority`, `save_path`, `search_imdbid`, `manual_total_episode`, `custom_words`, `media_category_id`, `media_category`, `filter_groups`, `episode_group`
 
 ### `subscribehistory`
 - Purpose: Stores snapshots of completed or archived subscriptions and their final filter state.
 - Useful queries: Auditing historical subscriptions, media identity, completion criteria, and filter configuration.
 - Write boundary: Generated by subscription completion and archival; restore or delete through its business API.
-- Columns: `id`, `name`, `year`, `type`, `keyword`, `media_source`, `media_id`, `music_type`, `total_tracks`, `season`, `poster`, `backdrop`, `vote`, `description`, `filter`, `include`, `exclude`, `quality`, `resolution`, `effect`, `audio_quality`, `audio_format`, `min_bitrate`, `min_bit_depth`, `min_sample_rate`, `total_episode`, `start_episode`, `date`, `username`, `sites`, `best_version`, `best_version_full`, `current_priority`, `current_audio_format`, `current_bitrate`, `current_bit_depth`, `current_sample_rate`, `episode_priority`, `save_path`, `search_imdbid`, `custom_words`, `media_category_id`, `media_category`, `classification_rule_id`, `classification_policy_revision`, `classification_source`, `filter_groups`, `episode_group`
+- Columns: `id`, `name`, `year`, `type`, `search_interval`, `keyword`, `media_source`, `media_id`, `music_type`, `total_tracks`, `season`, `poster`, `backdrop`, `vote`, `description`, `filter`, `include`, `exclude`, `quality`, `resolution`, `effect`, `audio_quality`, `audio_format`, `min_bitrate`, `min_bit_depth`, `min_sample_rate`, `total_episode`, `start_episode`, `date`, `username`, `sites`, `best_version`, `best_version_full`, `current_priority`, `current_audio_format`, `current_bitrate`, `current_bit_depth`, `current_sample_rate`, `episode_priority`, `save_path`, `search_imdbid`, `custom_words`, `media_category_id`, `media_category`, `classification_rule_id`, `classification_policy_revision`, `classification_source`, `filter_groups`, `episode_group`
 
 ### `subscriptionsearchbatch`
 - Purpose: Stores durable subscription search batches, source, aggregate state, counts, and cancellation requests.
diff --git a/skills/moviepilot-api/SKILL.md b/skills/moviepilot-api/SKILL.md
index 76879d0b3..3b60bb36f 100644
--- a/skills/moviepilot-api/SKILL.md
+++ b/skills/moviepilot-api/SKILL.md
@@ -1129,7 +1129,7 @@ Purpose: Read configured directory or storage settings.
 Purpose: Create one movie, TV, or music subscription.
 - `path_params`: none
 - `query`: none
-- `body`: `audio_format` (string|null): Requested or recorded audio container or codec, such as FLAC or MP3.; `audio_quality` (string|null): Subscription audio-quality rule, such as hires, lossless, or lossy.; `backdrop` (string|null): Backdrop image URL stored with the media or subscription.; `best_version` (integer|null): Enable normal best-version upgrading when set to 1.; `best_version_full` (integer|null): Enable full best-version upgrading when set to 1.; `classification_policy_revision` (integer|null): Policy revision that produced the persisted classification snapshot.; `classification_rule_id` (string|null): Stable rule ID that selected the persisted classification category.; `classification_source` (string|null): Selection source recorded with the persisted classification snapshot.; `completed_episode` (integer|null): Highest episode number already completed for the subscription.; `current_audio_format` (string|null): Audio format of the best version currently held.; `current_bit_depth` (integer|null): Bit depth of the best version currently held.; `current_bitrate` (integer|null): Bitrate of the best version currently held.; `current_priority` (integer|null): Calculated priority of the best version currently held.; `current_sample_rate` (integer|null): Sample rate of the best version currently held.; `custom_words` (string|null): Custom recognition or rename words applied to this media workflow.; `date` (string|null): Record creation or completion timestamp used by the history item.; `description` (string|null): Human-readable media, torrent, or subscription description.; `downloader` (string|null): Configured downloader instance name.; `effect` (string|null): Video or release-effect filter expression used by the subscription.; `episode_group` (string|null): TMDB episode-group identifier used for alternate episode ordering.; `episode_priority` (object|null): Per-episode best-version priority state.; `exclude` (string|null): Regular expression or filter expression that rejects matching releases.; `execution_status` (SubscriptionExecutionStatus|null): Current subscription execution status returned with the subscription snapshot.; `filter` (string|null): Named filter rule or rule expression applied to this site or subscription.; `filter_groups` (array<string>|null): Ordered filter-rule group names applied to the subscription.; `id` (integer|null): Persistent database identifier of the supplied record.; `include` (string|null): Regular expression or filter expression that a release must match.; `keyword` (string|null): Case-insensitive substring used to discover settings or filter storage entries.; `lack_episode` (integer|null; default `0`): Number of episodes still missing from the subscription.; `last_update` (string|null): Timestamp of the subscription's most recent update.; `media_category` (string|null): MoviePilot library category assigned to the media.; `media_category_id` (string|null): Stable classification category ID; preserve it separately from the current category path snapshot.; `media_id` (string|null): Source-native media ID. Always pair it with the exact media_source returned by search.; `media_source` (MediaSource|null): Metadata source identifier. Preserve the exact value returned with media_id.; `min_bit_depth` (integer|null): Minimum acceptable audio bit depth in bits.; `min_bitrate` (integer|null): Minimum acceptable audio bitrate in bits per second.; `min_sample_rate` (integer|null): Minimum acceptable audio sample rate in hertz.; `music_type` (string|null): Music identity level: recording, album, or artist where supported.; `name` (string|null): Human-readable name of the site, storage item, subscription, or rule group.; `note` (array<integer>|null): Structured auxiliary metadata stored with the record.; `poster` (string|null): Poster image URL stored with the media or subscription.; `quality` (string|null): Video or release quality filter expression.; `resolution` (string|null): Video resolution filter expression, such as 1080p or 2160p.; `save_path` (string|null): Configured downloader-side save path for the download or subscription.; `search_imdbid` (integer|null; default `0`): Use IMDb identity during subscription search when set to 1.; `season` (integer|null): Season number used by the media, search, subscription, or transfer operation.; `sites` (array<integer>|null): Exact site IDs included in the search or subscription scope.; `start_episode` (integer|null; default `0`): First episode number requested by the subscription.; `state` (string|null): Current site, subscription, marketplace, or transfer state filter.; `total_episode` (integer|null; default `0`): Expected total episode count for the subscription.; `total_tracks` (integer|null): Expected or recorded track count for a music item.; `type` (string|null): MoviePilot media or storage item type required by the selected operation.; `username` (string|null): MoviePilot or site username required by the selected operation.; `vote` (number|null; default `0.0`): Media vote average stored with the subscription.; `year` (string|null): Release or premiere year used to disambiguate the media title.
+- `body`: `audio_format` (string|null): Requested or recorded audio container or codec, such as FLAC or MP3.; `audio_quality` (string|null): Subscription audio-quality rule, such as hires, lossless, or lossy.; `backdrop` (string|null): Backdrop image URL stored with the media or subscription.; `best_version` (integer|null): Enable normal best-version upgrading when set to 1.; `best_version_full` (integer|null): Enable full best-version upgrading when set to 1.; `classification_policy_revision` (integer|null): Policy revision that produced the persisted classification snapshot.; `classification_rule_id` (string|null): Stable rule ID that selected the persisted classification category.; `classification_source` (string|null): Selection source recorded with the persisted classification snapshot.; `completed_episode` (integer|null): Highest episode number already completed for the subscription.; `current_audio_format` (string|null): Audio format of the best version currently held.; `current_bit_depth` (integer|null): Bit depth of the best version currently held.; `current_bitrate` (integer|null): Bitrate of the best version currently held.; `current_priority` (integer|null): Calculated priority of the best version currently held.; `current_sample_rate` (integer|null): Sample rate of the best version currently held.; `custom_words` (string|null): Custom recognition or rename words applied to this media workflow.; `date` (string|null): Record creation or completion timestamp used by the history item.; `description` (string|null): Human-readable media, torrent, or subscription description.; `downloader` (string|null): Configured downloader instance name.; `effect` (string|null): Video or release-effect filter expression used by the subscription.; `episode_group` (string|null): TMDB episode-group identifier used for alternate episode ordering.; `episode_priority` (object|null): Per-episode best-version priority state.; `exclude` (string|null): Regular expression or filter expression that rejects matching releases.; `execution_status` (SubscriptionExecutionStatus|null): Current subscription execution status returned with the subscription snapshot.; `filter` (string|null): Named filter rule or rule expression applied to this site or subscription.; `filter_groups` (array<string>|null): Ordered filter-rule group names applied to the subscription.; `id` (integer|null): Persistent database identifier of the supplied record.; `include` (string|null): Regular expression or filter expression that a release must match.; `keyword` (string|null): Case-insensitive substring used to discover settings or filter storage entries.; `lack_episode` (integer|null; default `0`): Number of episodes still missing from the subscription.; `last_search` (string|null): Read-only UTC timestamp of the most recent subscription search attempt.; `last_update` (string|null): Timestamp of the subscription's most recent update.; `media_category` (string|null): MoviePilot library category assigned to the media.; `media_category_id` (string|null): Stable classification category ID; preserve it separately from the current category path snapshot.; `media_id` (string|null): Source-native media ID. Always pair it with the exact media_source returned by search.; `media_source` (MediaSource|null): Metadata source identifier. Preserve the exact value returned with media_id.; `min_bit_depth` (integer|null): Minimum acceptable audio bit depth in bits.; `min_bitrate` (integer|null): Minimum acceptable audio bitrate in bits per second.; `min_sample_rate` (integer|null): Minimum acceptable audio sample rate in hertz.; `music_type` (string|null): Music identity level: recording, album, or artist where supported.; `name` (string|null): Human-readable name of the site, storage item, subscription, or rule group.; `note` (array<integer>|null): Structured auxiliary metadata stored with the record.; `poster` (string|null): Poster image URL stored with the media or subscription.; `quality` (string|null): Video or release quality filter expression.; `resolution` (string|null): Video resolution filter expression, such as 1080p or 2160p.; `save_path` (string|null): Configured downloader-side save path for the download or subscription.; `search_imdbid` (integer|null; default `0`): Use IMDb identity during subscription search when set to 1.; `search_interval` (integer|null): Scheduled search interval in whole hours (1-8760); null uses the system interval.; `season` (integer|null): Season number used by the media, search, subscription, or transfer operation.; `sites` (array<integer>|null): Exact site IDs included in the search or subscription scope.; `start_episode` (integer|null; default `0`): First episode number requested by the subscription.; `state` (string|null): Current site, subscription, marketplace, or transfer state filter.; `total_episode` (integer|null; default `0`): Expected total episode count for the subscription.; `total_tracks` (integer|null): Expected or recorded track count for a music item.; `type` (string|null): MoviePilot media or storage item type required by the selected operation.; `username` (string|null): MoviePilot or site username required by the selected operation.; `vote` (number|null; default `0.0`): Media vote average stored with the subscription.; `year` (string|null): Release or premiere year used to disambiguate the media title.
 
 ### `subscription.delete`
 `DELETE /api/v1/subscribe/{subscribe_id}`; policy effect: `destructive_write`.
@@ -1303,7 +1303,7 @@ Purpose: Set one accessible subscription to running, paused, or stopped state.
 Purpose: Update one existing movie, TV, or music subscription.
 - `path_params`: none
 - `query`: none
-- `body`: `audio_format` (string|null): Requested or recorded audio container or codec, such as FLAC or MP3.; `audio_quality` (string|null): Subscription audio-quality rule, such as hires, lossless, or lossy.; `backdrop` (string|null): Backdrop image URL stored with the media or subscription.; `best_version` (integer|null): Enable normal best-version upgrading when set to 1.; `best_version_full` (integer|null): Enable full best-version upgrading when set to 1.; `classification_policy_revision` (integer|null): Policy revision that produced the persisted classification snapshot.; `classification_rule_id` (string|null): Stable rule ID that selected the persisted classification category.; `classification_source` (string|null): Selection source recorded with the persisted classification snapshot.; `completed_episode` (integer|null): Highest episode number already completed for the subscription.; `current_audio_format` (string|null): Audio format of the best version currently held.; `current_bit_depth` (integer|null): Bit depth of the best version currently held.; `current_bitrate` (integer|null): Bitrate of the best version currently held.; `current_priority` (integer|null): Calculated priority of the best version currently held.; `current_sample_rate` (integer|null): Sample rate of the best version currently held.; `custom_words` (string|null): Custom recognition or rename words applied to this media workflow.; `date` (string|null): Record creation or completion timestamp used by the history item.; `description` (string|null): Human-readable media, torrent, or subscription description.; `downloader` (string|null): Configured downloader instance name.; `effect` (string|null): Video or release-effect filter expression used by the subscription.; `episode_group` (string|null): TMDB episode-group identifier used for alternate episode ordering.; `episode_priority` (object|null): Per-episode best-version priority state.; `exclude` (string|null): Regular expression or filter expression that rejects matching releases.; `execution_status` (SubscriptionExecutionStatus|null): Current subscription execution status returned with the subscription snapshot.; `filter` (string|null): Named filter rule or rule expression applied to this site or subscription.; `filter_groups` (array<string>|null): Ordered filter-rule group names applied to the subscription.; `id` (integer|null): Persistent database identifier of the supplied record.; `include` (string|null): Regular expression or filter expression that a release must match.; `keyword` (string|null): Case-insensitive substring used to discover settings or filter storage entries.; `lack_episode` (integer|null; default `0`): Number of episodes still missing from the subscription.; `last_update` (string|null): Timestamp of the subscription's most recent update.; `media_category` (string|null): MoviePilot library category assigned to the media.; `media_category_id` (string|null): Stable classification category ID; preserve it separately from the current category path snapshot.; `media_id` (string|null): Source-native media ID. Always pair it with the exact media_source returned by search.; `media_source` (MediaSource|null): Metadata source identifier. Preserve the exact value returned with media_id.; `min_bit_depth` (integer|null): Minimum acceptable audio bit depth in bits.; `min_bitrate` (integer|null): Minimum acceptable audio bitrate in bits per second.; `min_sample_rate` (integer|null): Minimum acceptable audio sample rate in hertz.; `music_type` (string|null): Music identity level: recording, album, or artist where supported.; `name` (string|null): Human-readable name of the site, storage item, subscription, or rule group.; `note` (array<integer>|null): Structured auxiliary metadata stored with the record.; `poster` (string|null): Poster image URL stored with the media or subscription.; `quality` (string|null): Video or release quality filter expression.; `resolution` (string|null): Video resolution filter expression, such as 1080p or 2160p.; `save_path` (string|null): Configured downloader-side save path for the download or subscription.; `search_imdbid` (integer|null; default `0`): Use IMDb identity during subscription search when set to 1.; `season` (integer|null): Season number used by the media, search, subscription, or transfer operation.; `sites` (array<integer>|null): Exact site IDs included in the search or subscription scope.; `start_episode` (integer|null; default `0`): First episode number requested by the subscription.; `state` (string|null): Current site, subscription, marketplace, or transfer state filter.; `total_episode` (integer|null; default `0`): Expected total episode count for the subscription.; `total_tracks` (integer|null): Expected or recorded track count for a music item.; `type` (string|null): MoviePilot media or storage item type required by the selected operation.; `username` (string|null): MoviePilot or site username required by the selected operation.; `vote` (number|null; default `0.0`): Media vote average stored with the subscription.; `year` (string|null): Release or premiere year used to disambiguate the media title.
+- `body`: `audio_format` (string|null): Requested or recorded audio container or codec, such as FLAC or MP3.; `audio_quality` (string|null): Subscription audio-quality rule, such as hires, lossless, or lossy.; `backdrop` (string|null): Backdrop image URL stored with the media or subscription.; `best_version` (integer|null): Enable normal best-version upgrading when set to 1.; `best_version_full` (integer|null): Enable full best-version upgrading when set to 1.; `classification_policy_revision` (integer|null): Policy revision that produced the persisted classification snapshot.; `classification_rule_id` (string|null): Stable rule ID that selected the persisted classification category.; `classification_source` (string|null): Selection source recorded with the persisted classification snapshot.; `completed_episode` (integer|null): Highest episode number already completed for the subscription.; `current_audio_format` (string|null): Audio format of the best version currently held.; `current_bit_depth` (integer|null): Bit depth of the best version currently held.; `current_bitrate` (integer|null): Bitrate of the best version currently held.; `current_priority` (integer|null): Calculated priority of the best version currently held.; `current_sample_rate` (integer|null): Sample rate of the best version currently held.; `custom_words` (string|null): Custom recognition or rename words applied to this media workflow.; `date` (string|null): Record creation or completion timestamp used by the history item.; `description` (string|null): Human-readable media, torrent, or subscription description.; `downloader` (string|null): Configured downloader instance name.; `effect` (string|null): Video or release-effect filter expression used by the subscription.; `episode_group` (string|null): TMDB episode-group identifier used for alternate episode ordering.; `episode_priority` (object|null): Per-episode best-version priority state.; `exclude` (string|null): Regular expression or filter expression that rejects matching releases.; `execution_status` (SubscriptionExecutionStatus|null): Current subscription execution status returned with the subscription snapshot.; `filter` (string|null): Named filter rule or rule expression applied to this site or subscription.; `filter_groups` (array<string>|null): Ordered filter-rule group names applied to the subscription.; `id` (integer|null): Persistent database identifier of the supplied record.; `include` (string|null): Regular expression or filter expression that a release must match.; `keyword` (string|null): Case-insensitive substring used to discover settings or filter storage entries.; `lack_episode` (integer|null; default `0`): Number of episodes still missing from the subscription.; `last_search` (string|null): Read-only UTC timestamp of the most recent subscription search attempt.; `last_update` (string|null): Timestamp of the subscription's most recent update.; `media_category` (string|null): MoviePilot library category assigned to the media.; `media_category_id` (string|null): Stable classification category ID; preserve it separately from the current category path snapshot.; `media_id` (string|null): Source-native media ID. Always pair it with the exact media_source returned by search.; `media_source` (MediaSource|null): Metadata source identifier. Preserve the exact value returned with media_id.; `min_bit_depth` (integer|null): Minimum acceptable audio bit depth in bits.; `min_bitrate` (integer|null): Minimum acceptable audio bitrate in bits per second.; `min_sample_rate` (integer|null): Minimum acceptable audio sample rate in hertz.; `music_type` (string|null): Music identity level: recording, album, or artist where supported.; `name` (string|null): Human-readable name of the site, storage item, subscription, or rule group.; `note` (array<integer>|null): Structured auxiliary metadata stored with the record.; `poster` (string|null): Poster image URL stored with the media or subscription.; `quality` (string|null): Video or release quality filter expression.; `resolution` (string|null): Video resolution filter expression, such as 1080p or 2160p.; `save_path` (string|null): Configured downloader-side save path for the download or subscription.; `search_imdbid` (integer|null; default `0`): Use IMDb identity during subscription search when set to 1.; `search_interval` (integer|null): Scheduled search interval in whole hours (1-8760); null uses the system interval.; `season` (integer|null): Season number used by the media, search, subscription, or transfer operation.; `sites` (array<integer>|null): Exact site IDs included in the search or subscription scope.; `start_episode` (integer|null; default `0`): First episode number requested by the subscription.; `state` (string|null): Current site, subscription, marketplace, or transfer state filter.; `total_episode` (integer|null; default `0`): Expected total episode count for the subscription.; `total_tracks` (integer|null): Expected or recorded track count for a music item.; `type` (string|null): MoviePilot media or storage item type required by the selected operation.; `username` (string|null): MoviePilot or site username required by the selected operation.; `vote` (number|null; default `0.0`): Media vote average stored with the subscription.; `year` (string|null): Release or premiere year used to disambiguate the media title.
 
 ### `subscription.user.list`
 `GET /api/v1/subscribe/user/{username}`; policy effect: `safe_read`.
diff --git a/tests/test_database_backup_scheduler.py b/tests/test_database_backup_scheduler.py
index 3e5c32b42..338d6f6ee 100644
--- a/tests/test_database_backup_scheduler.py
+++ b/tests/test_database_backup_scheduler.py
@@ -148,6 +148,29 @@ def test_scheduled_backup_uses_registered_database_governance(monkeypatch) -> No
     governance.create_backup.assert_called_once_with()
 
 
+@pytest.mark.parametrize("enabled", [False, True])
+def test_subscription_search_scans_due_items_only_when_enabled(monkeypatch, enabled) -> None:
+    """系统开关控制到期扫描，五分钟扫描节奏与实际搜索间隔分别传递。"""
+    scheduler = _scheduler()
+    scheduler._services = Mock()
+    background_scheduler = Mock()
+    monkeypatch.setattr(scheduler_catalog, "BackgroundScheduler", lambda **_kwargs: background_scheduler)
+    monkeypatch.setattr(scheduler_catalog, "get_plugin_manager", lambda: Mock())
+    monkeypatch.setattr(scheduler_catalog, "get_mediaserver_configs", lambda **_kwargs: [])
+    monkeypatch.setattr(scheduler, "init_workflow_jobs", lambda: None)
+    monkeypatch.setattr(scheduler, "init_agent_task_jobs", lambda: None)
+    monkeypatch.setattr(scheduler, "init_plugin_jobs", lambda: None)
+
+    scheduler._initialize_catalog(_config(subscribe_search=enabled, subscribe_search_interval=48))
+
+    calls = [call for call in background_scheduler.add_job.call_args_list
+             if call.kwargs.get("id") == "subscribe_search"]
+    assert bool(calls) is enabled
+    if enabled:
+        assert calls[0].kwargs["minutes"] == 5
+        assert scheduler._jobs["subscribe_search"]["kwargs"]["scheduled_interval"] == 48
+
+
 def test_scheduler_database_dependencies_are_explicit_module_imports() -> None:
     scheduler_root = Path(__file__).parents[1] / "app" / "scheduler"
     trees = [ast.parse(path.read_text(encoding="utf-8")) for path in scheduler_root.glob("*.py")]
diff --git a/tests/test_subscribe_search_state.py b/tests/test_subscribe_search_state.py
index dbdb27fee..5e514412a 100644
--- a/tests/test_subscribe_search_state.py
+++ b/tests/test_subscribe_search_state.py
@@ -125,8 +125,9 @@ def test_new_subscribe_search_marks_state_after_attempt(monkeypatch) -> None:
         chain.search(state="N", manual=False)
 
     media_chain.recognize_media.assert_called_once()
-    assert len(_SubscribeOper.updates) == 1
-    subscribe_id, subscription_patch = _SubscribeOper.updates[0]
+    assert len(_SubscribeOper.updates) == 2
+    assert "last_search" in _SubscribeOper.updates[0][1].to_payload()
+    subscribe_id, subscription_patch = _SubscribeOper.updates[1]
     assert subscribe_id == 31
     assert subscription_patch == SubscriptionPatch({"state": "R"})
 
@@ -150,6 +151,7 @@ def test_targeted_batch_searches_all_ids_without_state_scan(monkeypatch) -> None
     with patch.object(subscribe_search, "MediaChain", return_value=media_chain):
         chain = object.__new__(SubscribeChain)
         chain.subscription_repository = subscribe_oper
+        monkeypatch.setattr(chain, "_SubscribeChain__apply_subscribe_update", lambda sub, *_args, **_kwargs: sub)
         chain.search(sids=(31, 32), state=None, manual=False)
 
     assert [item.args for item in subscribe_oper.get.call_args_list] == [
diff --git a/tests/test_subscription_schedule.py b/tests/test_subscription_schedule.py
new file mode 100644
index 000000000..80a960a88
--- /dev/null
+++ b/tests/test_subscription_schedule.py
@@ -0,0 +1,136 @@
+"""逐订阅搜索周期的到期、写入和数据库迁移合同。"""
+
+import importlib
+from dataclasses import replace
+from datetime import datetime, timedelta, timezone
+from types import SimpleNamespace
+from unittest.mock import Mock
+
+import pytest
+import sqlalchemy as sa
+from alembic.migration import MigrationContext
+from alembic.operations import Operations
+from pydantic import ValidationError
+
+from app.application.subscription.contract import SubscriptionSnapshot
+from app.application.subscription.query import subscription_search_due
+from app.chain.subscribe.facade import SubscribeChain
+from app.db.adapters.subscription import TransactionalSubscriptionRepository
+from app.db.models.subscribe import Subscribe
+from app.db.session import SessionFactory, async_session_scope
+from app.scheduler.catalog import _subscription_search_job_specs
+from app.schemas.subscribe import Subscribe as SubscribeSchema
+
+NOW = datetime(2026, 9, 8, 12, tzinfo=timezone.utc)
+
+
+@pytest.mark.parametrize("interval,elapsed,due", [
+    (None, 23, False), (None, 24, True), (1, 0.99, False), (1, 1, True),
+    (48, 24, False), (48, 48, True), (1, -1, False),
+])
+def test_search_due_uses_custom_or_system_interval(interval, elapsed, due):
+    """独立周期可短于或长于系统周期，并在精确到期时放行。"""
+    subscribe = SubscriptionSnapshot(
+        id=1, name="周期测试", state="R", search_interval=interval,
+        last_search=(NOW - timedelta(hours=elapsed)).isoformat(),
+    )
+    assert subscription_search_due(subscribe, 24, NOW) is due
+    assert subscription_search_due(replace(subscribe, state="P"), 24, NOW) is due
+    assert not subscription_search_due(replace(subscribe, state="S"), 24, NOW)
+    assert not subscription_search_due(replace(subscribe, state="N"), 24, NOW)
+
+
+def test_legacy_creation_time_and_missing_timestamp():
+    """旧本地时间按本地时区解析，缺失或损坏时间允许恢复搜索。"""
+    recent = (NOW - timedelta(hours=1)).astimezone().strftime("%Y-%m-%d %H:%M:%S")
+    subscribe = SubscriptionSnapshot(id=1, name="旧订阅", state="R", date=recent)
+    assert not subscription_search_due(subscribe, 24, NOW)
+    for value in (None, "invalid"):
+        assert subscription_search_due(replace(subscribe, date=value), 24, NOW)
+
+
+def test_targeted_and_manual_selection_bypasses_schedule():
+    """自动调度仅选到期记录，手动及指定目标保持原有立即搜索语义。"""
+    recent = datetime.now(timezone.utc).isoformat()
+    subscribe = SubscriptionSnapshot(id=1, name="测试", state="R", last_search=recent)
+    chain = object.__new__(SubscribeChain)
+    chain.subscription_repository = SimpleNamespace(
+        list=Mock(return_value=[subscribe]), get=Mock(return_value=subscribe),
+    )
+    assert chain._load_search_subscriptions(None, None, "R", scheduled_interval=24) == []
+    assert chain._load_search_subscriptions(None, None, "R") == [subscribe]
+    assert chain._load_search_subscriptions(1, None, "R", scheduled_interval=24) == [subscribe]
+    assert chain._load_search_subscriptions(None, (1,), "R", scheduled_interval=24) == [subscribe]
+
+
+def test_scheduled_scan_does_not_create_empty_batches():
+    """尚未到期的订阅不产生空批次，也不触发队列消费。"""
+    subscribe = SubscriptionSnapshot(
+        id=1, name="测试", state="R", last_search=datetime.now(timezone.utc).isoformat(),
+    )
+    chain = object.__new__(SubscribeChain)
+    chain.subscription_repository = SimpleNamespace(list=Mock(return_value=[subscribe]))
+    chain.subscription_search_repository = Mock()
+    progress = Mock()
+    assert chain.search(state="R", scheduled_interval=24, progress_callback=progress) is None
+    chain.subscription_search_repository.enqueue.assert_not_called()
+    chain.subscription_search_repository.claim_next.assert_not_called()
+    assert progress.call_args.kwargs["value"] == 100
+
+
+@pytest.mark.parametrize("interval", [0, -1, 1.5, 8761])
+def test_invalid_search_interval_is_rejected(interval):
+    """接口拒绝零值、负数、小数和超出范围的独立周期。"""
+    with pytest.raises(ValidationError):
+        SubscribeSchema(search_interval=interval)
+
+
+def test_public_write_can_clear_interval_but_cannot_forge_search_time():
+    """恢复系统周期必须保留 null，客户端不能覆盖内部搜索时钟。"""
+    for value in (None, ""):
+        payload = SubscribeSchema(search_interval=value, last_search=NOW.isoformat())
+        assert payload.to_public_write_payload(exclude_unset=True) == {"search_interval": None}
+    assert SubscribeSchema().to_public_write_payload(exclude_unset=True) == {}
+
+
+def test_schedule_persists_across_repository_recreation(db):
+    """周期和搜索时间在 Session 关闭及仓储重建后仍可用于到期判断。"""
+    row = db.add(Subscribe(
+        name="持久周期", state="R", search_interval=48, last_search=NOW.isoformat(),
+    ))
+    for _ in range(2):
+        repository = TransactionalSubscriptionRepository(
+            sync_session=SessionFactory, async_session=async_session_scope,
+        )
+        snapshot = repository.get(row.id)
+        assert snapshot.search_interval == 48
+        assert snapshot.last_search == NOW.isoformat()
+        assert not subscription_search_due(snapshot, 24, NOW + timedelta(hours=24))
+
+
+def test_scheduler_passes_system_interval_only_to_periodic_search():
+    """调度目录传递可热重载的系统间隔，新增订阅保持首次立即搜索。"""
+    specs = {spec.job_id: spec for spec in _subscription_search_job_specs(Mock(), 12)}
+    assert specs["subscribe_search"].kwargs == {"state": "R", "scheduled_interval": 12}
+    assert specs["new_subscribe_search"].kwargs == {"state": "N"}
+
+
+def test_schedule_migration_preserves_existing_rows_and_is_reversible(monkeypatch):
+    """旧数据默认跟随系统，重复升级与回滚不损坏已有订阅。"""
+    migration = importlib.import_module("database.versions.f8a2c6e9b1d4_3_0_30")
+    engine = sa.create_engine("sqlite://")
+    with engine.begin() as connection:
+        for table in ("subscribe", "subscribehistory"):
+            connection.execute(sa.text(f"CREATE TABLE {table} (id INTEGER PRIMARY KEY, name TEXT)"))
+            connection.execute(sa.text(f"INSERT INTO {table} VALUES (1, '旧订阅')"))
+        monkeypatch.setattr(migration, "op", Operations(MigrationContext.configure(connection)))
+        migration.upgrade()
+        migration.upgrade()
+        for table in ("subscribe", "subscribehistory"):
+            row = connection.execute(sa.text(f"SELECT name, search_interval FROM {table}")).one()
+            assert row == ("旧订阅", None)
+        assert connection.execute(sa.text("SELECT last_search FROM subscribe")).scalar_one() is None
+        migration.downgrade()
+        migration.upgrade()
+        assert connection.execute(sa.text("SELECT name FROM subscribe")).scalar_one() == "旧订阅"
+    engine.dispose()