diff --git a/backend/src/module/parser/title_parser.py b/backend/src/module/parser/title_parser.py
index 49015b60..17013e88 100644
--- a/backend/src/module/parser/title_parser.py
+++ b/backend/src/module/parser/title_parser.py
@@ -456,7 +456,20 @@ class TitleParser:
             )
             deterministic = parse_outcome.result
             if parse_outcome.engine == "classic":
-                deterministic = _project_classic_release(raw, deterministic)
+                projected = _project_classic_release(raw, deterministic)
+                if (
+                    projected is None
+                    and deterministic is not None
+                    and deterministic.episode is None
+                    and deterministic.media_type is MediaType.UNKNOWN
+                    and persistence_target(deterministic) is not None
+                ):
+                    # 字幕组未标记特别篇/剧场版等字样的无集数单发资源（#1092）：
+                    # 旧版投影契约会整条拒绝，导致订阅静默漏抓。与 Preview 的
+                    # 准入策略对齐，按 eps_collect 整理集接纳；带集数的旧版
+                    # 拒绝（如小数集）不受影响。
+                    projected = deterministic
+                deterministic = projected
             if (
                 deterministic is None
                 and parse_outcome.trace is not None
diff --git a/backend/src/test/test_issue_bugs.py b/backend/src/test/test_issue_bugs.py
index 7fb6565c..b5b8cece 100644
--- a/backend/src/test/test_issue_bugs.py
+++ b/backend/src/test/test_issue_bugs.py
@@ -555,3 +555,77 @@ class TestIssue1005SearchOfficialTitle:
         db = BangumiDatabase(db_session)
         result = await db.search_official_title("不存在的番剧")
         assert result is None
+
+
+# ---------------------------------------------------------------------------
+# Issue #1092: 字幕组未标记特别篇等字样时，无集数标题无法解析
+# https://github.com/EstrellaXD/Auto_Bangumi/issues/1092
+#
+# LoliHouse 发布的 TV 特别篇（如 ONE PIECE HEROINES）标题中没有集数也没有
+# 特别篇/剧场版标记。Preview 引擎的准入策略按 eps_collect 整理集接纳这类
+# 带发布证据的单发资源，但 classic 引擎的旧版兼容投影会整条拒绝，导致
+# 默认配置下订阅静默漏抓。
+# ---------------------------------------------------------------------------
+
+
+class TestIssue1092UnmarkedEpisodelessRelease:
+    """Issue #1092: classic admission now matches the Preview policy."""
+
+    REPORTED_TITLE = (
+        "[LoliHouse] 海贼王：女英雄们的故事 / ONE PIECE HEROINES "
+        "[WebRip 1080p HEVC-10bit AAC][简繁内封字幕]"
+    )
+
+    async def test_classic_admits_reported_title_as_collect_bangumi(self):
+        from unittest.mock import patch
+
+        from module.conf import settings
+        from module.models import Bangumi
+        from module.models.config import LLM
+        from module.parser import TitleParser
+
+        with (
+            patch.object(settings, "llm", LLM(enable=False)),
+            patch.object(settings.rss_parser, "engine", "classic"),
+        ):
+            result = await TitleParser.raw_parser(self.REPORTED_TITLE)
+
+        assert isinstance(result, Bangumi)
+        assert result.official_title == "海贼王：女英雄们的故事"
+        assert result.group_name == "LoliHouse"
+        assert result.episode_type == "episode"
+        assert result.eps_collect is True
+
+    async def test_classic_and_preview_agree_on_reported_title(self):
+        from unittest.mock import patch
+
+        from module.conf import settings
+        from module.models import Bangumi
+        from module.models.config import LLM
+        from module.parser import TitleParser
+
+        with (
+            patch.object(settings, "llm", LLM(enable=False)),
+            patch.object(settings.rss_parser, "engine", "tokenizer"),
+        ):
+            preview_result = await TitleParser.raw_parser(self.REPORTED_TITLE)
+
+        assert isinstance(preview_result, Bangumi)
+        assert preview_result.official_title == "海贼王：女英雄们的故事"
+
+    async def test_classic_still_rejects_title_without_release_evidence(self):
+        from unittest.mock import patch
+
+        from module.conf import settings
+        from module.models.config import LLM
+        from module.parser import TitleParser
+
+        with (
+            patch.object(settings, "llm", LLM(enable=False)),
+            patch.object(settings.rss_parser, "engine", "classic"),
+        ):
+            result = await TitleParser.raw_parser(
+                "Random Non Matching Title No Digits At All"
+            )
+
+        assert result is None
diff --git a/backend/src/test/test_title_parser.py b/backend/src/test/test_title_parser.py
index e944dfd9..450069ca 100644
--- a/backend/src/test/test_title_parser.py
+++ b/backend/src/test/test_title_parser.py
@@ -233,14 +233,8 @@ class TestConfiguredParserEngine:
 
         assert result is None
 
-    @pytest.mark.parametrize(
-        "raw",
-        (
-            "[Group] Anime [1080p]",
-            "[Group] Anime EP07.5 [1080p]",
-        ),
-    )
-    async def test_classic_preserves_legacy_compatibility_rejections(self, raw):
+    async def test_classic_preserves_legacy_compatibility_rejections(self):
+        raw = "[Group] Anime EP07.5 [1080p]"
         parsed = parse_classic_release_title(raw)
 
         assert parsed is not None
@@ -254,6 +248,24 @@ class TestConfiguredParserEngine:
 
         assert result is None
 
+    async def test_classic_admits_episodeless_release_with_evidence(self):
+        """无集数且无类型标记但带发布证据的单发资源（#1092）：旧版 Episode
+        投影仍拒绝（冻结契约），但 classic 准入与 Preview 策略对齐后放行。"""
+        raw = "[Group] Anime [1080p]"
+        parsed = parse_classic_release_title(raw)
+
+        assert parsed is not None
+        assert to_legacy_episode(parsed) is None
+
+        with (
+            patch.object(settings, "llm", LLM(enable=False)),
+            patch.object(settings.rss_parser, "engine", "classic"),
+        ):
+            result = await TitleParser.raw_parser(raw)
+
+        assert isinstance(result, Bangumi)
+        assert result.eps_collect is True
+
     async def test_classic_projects_admitted_result_through_legacy_episode(self):
         raw = "[Group] Anime - 01 [2024][1080p]"
         parsed = parse_classic_release_title(raw)