LLMs famously aren’t that good at using new frameworks/languages. Sure they can get by with the right context, but most people are pointing them at standard frameworks in common languages to maximize the quality of their output.