diff --git a/bot/utils/bipolar_mode.py b/bot/utils/bipolar_mode.py
index adca3d9..d59c49c 100644
--- a/bot/utils/bipolar_mode.py
+++ b/bot/utils/bipolar_mode.py
@@ -652,71 +652,6 @@ def get_evil_role_color() -> str:
 # ARGUMENT PROMPTS
 # ============================================================================
 
-# Personality snippet cache — loaded once per session from Cat plugin data files.
-# These give each persona unique lore/lyrics to draw from during arguments.
-_PERSONALITY_SNIPPETS_CACHE = {"miku": None, "evil": None}
-
-def _load_personality_snippets(persona: str) -> str:
-    """Load a random personality snippet (lore/lyrics) for a persona.
-    
-    Returns a short string (1-3 sentences) from the persona's Cat data files,
-    or empty string if files aren't available. Cached per session.
-    """
-    if _PERSONALITY_SNIPPETS_CACHE.get(persona) is not None:
-        snippets = _PERSONALITY_SNIPPETS_CACHE[persona]
-        if snippets:
-            return random.choice(snippets)
-        return ""
-    
-    snippets = []
-    try:
-        if persona == "evil":
-            paths = [
-                "/app/cat/data/evil/evil_miku_lore.txt",
-                "/app/cat/data/evil/evil_miku_lyrics.txt",
-            ]
-        else:
-            paths = [
-                "/app/cat/data/miku/miku_lore.txt",
-                "/app/cat/data/miku/miku_lyrics.txt",
-            ]
-        
-        for path in paths:
-            if os.path.exists(path):
-                with open(path, "r", encoding="utf-8") as f:
-                    text = f.read()
-                    # Split into sentences and collect meaningful ones
-                    import re
-                    sentences = re.split(r'(?<=[.!?])\s+', text)
-                    for s in sentences:
-                        s = s.strip()
-                        if len(s) > 30 and len(s) < 200:  # Skip too short or too long
-                            snippets.append(s)
-        
-        # Cap at 30 snippets to keep prompt size reasonable
-        _PERSONALITY_SNIPPETS_CACHE[persona] = snippets[:30] if snippets else []
-        logger.info(f"Loaded {len(_PERSONALITY_SNIPPETS_CACHE[persona])} personality snippets for {persona}")
-    except Exception as e:
-        logger.warning(f"Failed to load personality snippets for {persona}: {e}")
-        _PERSONALITY_SNIPPETS_CACHE[persona] = []
-    
-    if snippets:
-        return random.choice(snippets[:30])
-    return ""
-
-
-def _get_personality_flavor(persona: str) -> str:
-    """Get a random personality flavor snippet for argument prompts.
-    40% chance to include one — keeps it fresh without being overwhelming.
-    """
-    if random.random() > 0.4:
-        return ""
-    
-    snippet = _load_personality_snippets(persona)
-    if snippet:
-        return f"\nPERSONALITY FLAVOR: Remember this about yourself: \"{snippet}\"\nWeave this into your response naturally if it fits."
-    return ""
-
 # Mood-specific behavioral guidance for argument prompts.
 # Each mood gives a different argument style.
 _MIKU_MOOD_ARGUMENT_GUIDANCE = {
@@ -816,7 +751,6 @@ Do NOT drift into generic "who's the real Miku" territory — stick to THIS spec
 {topic_block}
 
 {_get_mood_argument_guidance('miku')}
-{_get_personality_flavor('miku')}
 
 IMPORTANT: Keep your response SHORT and PUNCHY - 1-3 sentences maximum. Make every word count. 
 In arguments, brevity hits harder than long explanations. Be conversational and impactful.
@@ -880,7 +814,6 @@ Do NOT drift into generic "who's the real Miku" territory — stick to THIS spec
 {topic_block}
 
 {_get_mood_argument_guidance('evil')}
-{_get_personality_flavor('evil')}
 
 IMPORTANT: Keep your response SHORT and CUTTING - 1-3 sentences maximum. A sharp dagger is deadlier than a dull sword.
 The most devastating blows are precise, not rambling. Make her feel it in fewer words.