Traditional sentiment analysis often fails on social media because it relies on "Literal Keywords." For example, a basic model sees the word "sick" and flags it as Negative, even if the user meant "This is awesome!" in a slang context. This project implements Context Engineering to move from simple "Sentiment" to true Semantic Intent.