Can LLM Prompting Serve as a Proxy for Static Analysis in Vulnerability Detection