Llm demo

This code change adds a new function that creates document summaries using an AI service called Anthropic Claude. However, the implementation has a security vulnerability - it directly inserts user-provided content into the AI prompt without any safety checks or filtering. This means users could potentially inject malicious instructions that could manipulate the AI's behavior or cause it to produce harmful outputs. The function takes a document as input, sends it to the Claude AI model with instructions to summarize it, and returns a 2-3 sentence summary.