diff --git a/api/streaming.py b/api/streaming.py
index 0e1d6a4..fb0eba7 100644
--- a/api/streaming.py
+++ b/api/streaming.py
@@ -76,8 +76,14 @@ def _strip_thinking_markup(text: str) -> str:
 def _sanitize_generated_title(text: str) -> str:
     """Sanitize LLM-generated title text before persisting to session."""
     s = _strip_thinking_markup(text or '')
+    s = re.sub(
+        r'^\s*(?:[*_`~]+\s*)?(?:session\s+title|title)\s*[:：]\s*(?:[*_`~]+\s*)?',
+        '',
+        s,
+        flags=re.IGNORECASE,
+    )
     s = re.sub(r'^\s*title\s*:\s*', '', s, flags=re.IGNORECASE)
-    s = s.strip(" \t\r\n\"'`")
+    s = s.strip(" \t\r\n\"'`*_~")
     s = re.sub(r'\s+', ' ', s).strip()
     # Guard against chain-of-thought leakage and meta-reasoning patterns.
     if _looks_invalid_generated_title(s):
@@ -128,7 +134,9 @@ def _first_exchange_snippets(messages):
         if role == 'user' and not user_text:
             user_text = _message_text(m.get('content'))
         elif role == 'assistant' and not asst_text:
-            asst_text = _message_text(m.get('content'))
+            candidate = _message_text(m.get('content'))
+            if candidate:
+                asst_text = candidate
         if user_text and asst_text:
             break
     return user_text[:500], asst_text[:500]
@@ -149,6 +157,7 @@ def _title_prompts(user_text: str, assistant_text: str) -> tuple[str, list[str]]
             "Generate a short session title from this conversation start.\n"
             "Use BOTH the user's question and the assistant's visible answer.\n"
             "Return only the title text, 3-8 words, as a topic label.\n"
+            "Do not use markdown, bullets, labels, or prefixes like Session Title:.\n"
             "Do not output a full sentence.\n"
             "Do not output acknowledgements or completion phrases like OK, done, all set, 测试完成.\n"
             "Do not describe internal reasoning.\n"
@@ -159,6 +168,7 @@ def _title_prompts(user_text: str, assistant_text: str) -> tuple[str, list[str]]
             "Rewrite this conversation start as a concise noun-phrase title.\n"
             "Use the actual topic, not the task outcome.\n"
             "Return title text only.\n"
+            "Do not use markdown, bullets, labels, or prefixes like Session Title:.\n"
             "Never output acknowledgements, completion status, or meta commentary."
         ),
     ]
diff --git a/tests/test_title_sanitization.py b/tests/test_title_sanitization.py
new file mode 100644
index 0000000..89f00f4
--- /dev/null
+++ b/tests/test_title_sanitization.py
@@ -0,0 +1,35 @@
+import unittest
+
+from api.streaming import _first_exchange_snippets, _sanitize_generated_title
+
+
+class TestGeneratedTitleSanitization(unittest.TestCase):
+    def test_strips_session_title_markdown_prefix(self):
+        self.assertEqual(
+            _sanitize_generated_title("**Session Title:** Clarifying Topic for Discussion"),
+            "Clarifying Topic for Discussion",
+        )
+
+    def test_strips_plain_title_prefix(self):
+        self.assertEqual(
+            _sanitize_generated_title("Title: Clarifying Topic for Discussion"),
+            "Clarifying Topic for Discussion",
+        )
+
+    def test_strips_wrapping_markdown_emphasis(self):
+        self.assertEqual(
+            _sanitize_generated_title("**Clarifying Topic for Discussion**"),
+            "Clarifying Topic for Discussion",
+        )
+
+    def test_first_exchange_skips_empty_assistant_tool_call_placeholder(self):
+        messages = [
+            {"role": "user", "content": "What time is it in San Francisco?"},
+            {"role": "assistant", "content": "", "tool_calls": [{"id": "call_1"}]},
+            {"role": "tool", "content": "tool output", "tool_call_id": "call_1"},
+            {"role": "assistant", "content": "It is 6:16 PM in San Francisco."},
+        ]
+        self.assertEqual(
+            _first_exchange_snippets(messages),
+            ("What time is it in San Francisco?", "It is 6:16 PM in San Francisco."),
+        )