[Feature]: Stablize prompts for all QType

[rootCircle]

Description

Recently, we introduced langchain output parser in #33, which will help give us the desired output in JSON, Date object, or String wherever required! Since, introducing that, the extensive prompt that we had in promptEngine.ts will become redundant and can act as source to hallucinate the LLM as well in some cases!

Suggested Solution

We remove unnecessary output instructions from the prompts and vet the output as well! Some prompts also have example as output as well like MULTI_CORRECT etc, in that the response and input and prompt vary a lot, we need to fix that as well! Please don't club this issue with any other!

Alternatives

No response

Additional Context

No response