Contributed to the home-assistant/core repository by developing a GetDateTime tool that injects the current date and time into prompts for large language model interactions when APIs lack this information. This feature was integrated into the Assist API and accompanied by updates to the test suite to ensure robust coverage and maintainability. Additionally, implemented an optimization for LLM prefix caching, which reduces redundant prompt prefixes and improves response latency. The work focused on backend development and API integration using Python, laying the foundation for more reliable and scalable prompt-based features while enhancing overall system reliability and developer productivity.