Fix query appearing twice in chat history

The initial query appears twice in the prompt, this is ignored by OpenAI
models, however it breaks with Gemini models are they fail with an error
stating that AI and User prompts need to alternate.

Tested all search modes with both OpenAI GTP3 turbo and Vertex Gemini
1.0 and this changes appears to now function correctly with both.

src/agents/wolframAlphaSearchAgent.ts

@@ -165,7 +165,6 @@ const createBasicWolframAlphaSearchAnsweringChain = (llm: BaseChatModel) => {
     ChatPromptTemplate.fromMessages([
       ['system', basicWolframAlphaSearchResponsePrompt],
       new MessagesPlaceholder('chat_history'),
-      ['user', '{query}'],
     ]),
     llm,
     strParser,