Fix Discord bot to use LiteLLM Responses API v2 with MCP tools

Key changes:
- Upgrade OpenAI SDK from 1.x to 2.x (required for responses API)
- Update get_ai_response to use developer role for system prompts
- Improve response extraction with output_text shorthand
- Add enhanced debug logging for troubleshooting
- Add traceback logging for better error diagnostics

This fixes the "'OpenAI' object has no attribute 'responses'" error
and enables proper MCP tool auto-execution through LiteLLM.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

scripts/discordbot.py

@@ -277,10 +277,10 @@ async def get_ai_response(history_messages: List[Dict[str, Any]], user_message:
             # Build input array with system prompt, history, and current message
             input_messages = []
 
-            # Add system prompt as developer role (newer models) or user role
+            # Add system prompt as developer role for newer models
             input_messages.append({
-                "role": "user",  # System messages converted to user for Responses API
-                "content": f"[System Instructions]\n{SYSTEM_PROMPT}"
+                "role": "developer",
+                "content": SYSTEM_PROMPT
             })
 
             # Add conversation history
@@ -329,6 +329,8 @@ async def get_ai_response(history_messages: List[Dict[str, Any]], user_message:
             else:
                 # Use Responses API with MCP tools
                 debug_log(f"Calling Responses API with {len(tools_config)} MCP servers")
+                debug_log(f"Input messages: {len(input_messages)} messages")
+
                 response = client.responses.create(
                     model=MODEL_NAME,
                     input=input_messages,
@@ -336,18 +338,28 @@ async def get_ai_response(history_messages: List[Dict[str, Any]], user_message:
                     stream=False
                 )
 
-                debug_log(f"Response status: {response.status}")
+                debug_log(f"Response received, status: {getattr(response, 'status', 'unknown')}")
 
                 # Extract text from Responses API format
+                # Try the shorthand first
+                response_text = getattr(response, 'output_text', None)
+
+                if response_text:
+                    debug_log(f"Got response via output_text shorthand: {response_text[:100]}...")
+                    return response_text
+
+                # Otherwise navigate the structure
                 if hasattr(response, 'output') and len(response.output) > 0:
                     for output in response.output:
                         if hasattr(output, 'type') and output.type == "message":
                             if hasattr(output, 'content') and len(output.content) > 0:
                                 for content in output.content:
                                     if hasattr(content, 'type') and content.type == "output_text":
+                                        debug_log(f"Got response via structure navigation: {content.text[:100]}...")
                                         return content.text
 
                 debug_log(f"Unexpected response format: {response}")
+                debug_log(f"Response attributes: {dir(response)}")
                 return "I received a response but couldn't extract the text. Please try again."
 
         # Standard chat completions (when tools disabled or MCP not available)
@@ -381,6 +393,8 @@ async def get_ai_response(history_messages: List[Dict[str, Any]], user_message:
         error_msg = f"Error calling LiteLLM API: {str(e)}"
         print(error_msg)
         debug_log(f"Exception details: {e}")
+        import traceback
+        debug_log(f"Traceback: {traceback.format_exc()}")
         return error_msg
 
 @bot.event