fix: Change max_tokens to max_completion_tokens for GPT-5 API compatibility

analyzer/llm_client.py

@@ -230,7 +230,7 @@ class LLMClient:
                 "messages": messages,
                 "temperature": temperature,
                 "top_p": top_p,
-                "max_tokens": max_tokens,
+                "max_completion_tokens": max_tokens,  # GPT-5 uses max_completion_tokens
                 "stream": stream,
             }
 

src/analyzer/llm_client.py

@@ -230,7 +230,7 @@ class LLMClient:
                 "messages": messages,
                 "temperature": temperature,
                 "top_p": top_p,
-                "max_tokens": max_tokens,
+                "max_completion_tokens": max_tokens,  # GPT-5 uses max_completion_tokens
                 "stream": stream,
             }