diff --git a/astrbot/builtin_stars/astrbot/main.py b/astrbot/builtin_stars/astrbot/main.py
index cc2c368a23..ee3ae13a48 100644
--- a/astrbot/builtin_stars/astrbot/main.py
+++ b/astrbot/builtin_stars/astrbot/main.py
@@ -58,6 +58,13 @@ async def handle_empty_mention(self, event: AstrMessageEvent):
             cfg = self.context.get_config(umo=event.unified_msg_origin)
             p_settings = cfg["platform_settings"]
             wake_prefix = cfg.get("wake_prefix", [])
+            provider_wake_prefix = cfg.get("provider_settings", {}).get(
+                "wake_prefix",
+                "",
+            )
+            for bot_wake_prefix in wake_prefix:
+                if provider_wake_prefix.startswith(bot_wake_prefix):
+                    provider_wake_prefix = provider_wake_prefix[len(bot_wake_prefix) :]
             if len(messages) != 1:
                 return
 
@@ -74,7 +81,10 @@ async def handle_empty_mention(self, event: AstrMessageEvent):
             if not (is_empty_mention or is_wake_prefix_only):
                 return
 
-            if p_settings.get("empty_mention_waiting_need_reply", True):
+            if (
+                p_settings.get("empty_mention_waiting_need_reply", True)
+                and not provider_wake_prefix
+            ):
                 try:
                     curr_cid = await self.context.conversation_manager.get_curr_conversation_id(
                         event.unified_msg_origin,
diff --git a/tests/unit/test_empty_mention_wake_prefix.py b/tests/unit/test_empty_mention_wake_prefix.py
new file mode 100644
index 0000000000..99c537bed0
--- /dev/null
+++ b/tests/unit/test_empty_mention_wake_prefix.py
@@ -0,0 +1,79 @@
+from types import SimpleNamespace
+from unittest.mock import AsyncMock, MagicMock
+
+import pytest
+
+import astrbot.builtin_stars.astrbot.main as main_module
+from astrbot.api.message_components import At, Plain
+
+
+@pytest.mark.asyncio
+@pytest.mark.parametrize(
+    ("provider_wake_prefix", "messages", "should_request_llm"),
+    [
+        ("chat", [At(qq="bot")], False),
+        ("/chat", [At(qq="bot")], False),
+        ("chat", [Plain(text="/")], False),
+        ("", [At(qq="bot")], True),
+        ("/", [At(qq="bot")], True),
+    ],
+)
+async def test_empty_mention_reply_respects_provider_wake_prefix(
+    monkeypatch,
+    provider_wake_prefix,
+    messages,
+    should_request_llm,
+):
+    """Verify that empty-mention replies respect the additional LLM wake prefix.
+
+    Args:
+        monkeypatch: Pytest fixture used to replace the 60-second session waiter.
+        provider_wake_prefix: Additional LLM wake prefix configured for the provider.
+        messages: Message components passed to the empty-mention handler.
+        should_request_llm: Whether an immediate LLM request should be produced.
+    """
+
+    def skip_waiting(_timeout):
+        def decorator(_callback):
+            async def wait(*_args, **_kwargs):
+                raise TimeoutError
+
+            return wait
+
+        return decorator
+
+    monkeypatch.setattr(main_module, "session_waiter", skip_waiting)
+
+    conversation_manager = SimpleNamespace(
+        get_curr_conversation_id=AsyncMock(return_value="conversation-id"),
+        get_conversation=AsyncMock(return_value=None),
+    )
+    main = main_module.Main.__new__(main_module.Main)
+    main.context = MagicMock()
+    main.context.conversation_manager = conversation_manager
+    main.context.get_config.return_value = {
+        "wake_prefix": ["/"],
+        "provider_settings": {"wake_prefix": provider_wake_prefix},
+        "platform_settings": {
+            "empty_mention_waiting": True,
+            "empty_mention_waiting_need_reply": True,
+        },
+    }
+
+    event = MagicMock()
+    event.unified_msg_origin = "aiocqhttp:GroupMessage:group"
+    event.get_messages.return_value = messages
+    event.get_self_id.return_value = "bot"
+    event.get_platform_id.return_value = "aiocqhttp"
+    llm_request = object()
+    event.request_llm.return_value = llm_request
+
+    results = [item async for item in main.handle_empty_mention(event)]
+
+    assert results == ([llm_request] if should_request_llm else [])
+    if should_request_llm:
+        event.request_llm.assert_called_once()
+        conversation_manager.get_curr_conversation_id.assert_awaited_once()
+    else:
+        event.request_llm.assert_not_called()
+        conversation_manager.get_curr_conversation_id.assert_not_awaited()