LLM Integration for Threat Assessment

Develop the integration logic for Claude 3.5 Haiku using Marvin. Define the Marvin schema for extracting critical threat attributes from the fused sensor data and any external contextual data obtained via Browserless. Craft a sophisticated prompt strategy for Claude 3.5 Haiku that takes the structured threat information, environmental data, and Rules of Engagement to generate prioritized, actionable countermeasure recommendations. Provide example inputs and the expected LLM output, ensuring it aligns with the `Threat_Assessment_Recommendation_Task` output format.