Chris Xu developed the Lumo System Prompt Definition for Proton’s AI assistant, focusing on a comprehensive, policy-driven prompt architecture within the asgeirtj/system_prompts_leaks repository. Leveraging expertise in AI development, natural language processing, and documentation, Chris structured the prompt to define the assistant’s identity, personality, tool usage, file handling, and communication style. The approach emphasized maintainability and governance, ensuring consistent behavior and safer user interactions. By using Markdown for clear documentation, Chris enabled easier onboarding for future prompt iterations. The work addressed trust and safety requirements, providing a scalable foundation that improved both user experience and operational efficiency for the Lumo project.