backend: replace LAST_FULL_PROMPT/LAST_CAT_INTERACTION with unified PROMPT_HISTORY deque

- globals.py: add collections.deque(maxlen=10) PROMPT_HISTORY with _prompt_id_counter
- globals.py: add legacy accessor functions _get_last_fallback_prompt() and _get_last_cat_interaction()
- bot.py: append to PROMPT_HISTORY instead of setting LAST_CAT_INTERACTION, remove 500-char truncation, add guild/channel/model fields
- image_handling.py: same pattern for Cat media responses
- llm.py: append fallback prompts to PROMPT_HISTORY with response filled after LLM reply
- routes/core.py: new GET /prompts and GET /prompts/{id} endpoints, legacy /prompt and /prompt/cat use accessor functions

bot/utils/image_handling.py

@@ -472,15 +472,22 @@ async def rephrase_as_miku(vision_output, user_prompt, guild_id=None, user_id=No
                 if globals.EVIL_MODE:
                     effective_mood = f"EVIL:{getattr(globals, 'EVIL_DM_MOOD', 'evil_neutral')}"
                 logger.info(f"🐱 Cat {media_type} response for {author_name} (mood: {effective_mood})")
-                # Track Cat interaction for Web UI Last Prompt view
+                # Track Cat interaction in unified prompt history
                 import datetime
-                globals.LAST_CAT_INTERACTION = {
+                globals._prompt_id_counter += 1
+                globals.PROMPT_HISTORY.append({
+                    "id": globals._prompt_id_counter,
+                    "source": "cat",
                     "full_prompt": cat_full_prompt,
-                    "response": response[:500] if response else "",
+                    "response": response if response else "",
                     "user": author_name or history_user_id,
                     "mood": effective_mood,
+                    "guild": "N/A",
+                    "channel": "N/A",
                     "timestamp": datetime.datetime.now().isoformat(),
-                }
+                    "model": "Cat LLM",
+                    "response_type": response_type,
+                })
         except Exception as e:
             logger.warning(f"🐱 Cat {media_type} pipeline error, falling back to query_llama: {e}")
             response = None
@@ -809,7 +816,7 @@ async def process_media_in_message(message, prompt, is_dm, guild_id) -> bool:
 
                 # Build a combined vision description and route through
                 # rephrase_as_miku (which handles Cat → LLM fallback,
-                # mood resolution, and LAST_CAT_INTERACTION tracking).
+                # mood resolution, and prompt history tracking).
                 combined_description = "\n".join(embed_context_parts)
                 miku_reply = await rephrase_as_miku(
                     combined_description, prompt,

bot/utils/llm.py

@@ -381,7 +381,23 @@ Please respond in a way that reflects this emotional tone.{pfp_context}"""
         media_note = media_descriptions.get(media_type, f"The user has sent you {media_type}.")
         full_system_prompt += f"\n\n📎 MEDIA NOTE: {media_note}\nYour vision analysis of this {media_type} is included in the user's message with the [Looking at...] prefix."
 
-    globals.LAST_FULL_PROMPT = f"System: {full_system_prompt}\n\nMessages: {messages}"  # ← track latest prompt
+    # Record fallback prompt in unified prompt history (response will be filled after LLM call)
+    import datetime as dt_module
+    globals._prompt_id_counter += 1
+    prompt_entry = {
+        "id": globals._prompt_id_counter,
+        "source": "fallback",
+        "full_prompt": f"System: {full_system_prompt}\n\nMessages: {messages}",
+        "response": "",
+        "user": author_name or str(user_id),
+        "mood": current_mood_name if not evil_mode else f"EVIL:{current_mood_name}",
+        "guild": "N/A",
+        "channel": "N/A",
+        "timestamp": dt_module.datetime.now().isoformat(),
+        "model": model,
+        "response_type": response_type,
+    }
+    globals.PROMPT_HISTORY.append(prompt_entry)
 
     headers = {'Content-Type': 'application/json'}
     
@@ -474,7 +490,10 @@ Please respond in a way that reflects this emotional tone.{pfp_context}"""
                             content=reply,
                             is_bot=True
                         )
-                    
+
+                    # Update the prompt history entry with the actual response
+                    prompt_entry["response"] = reply if reply else ""
+
                     return reply
                 else:
                     error_text = await response.text()