During December 2025, Arthan built the AI Safety Red Teaming Templates Library for the Azure/PyRIT repository, focusing on structured, scenario-based testing of AI safety controls. Leveraging skills in AI safety research, prompt engineering, and security testing, Arthan developed modular YAML templates that enable repeatable and extensible red-teaming exercises. The work included integrating a jailbreak template collection into PyRIT workflows, streamlining the setup for evaluating AI systems against safety bypass techniques. By contributing to documentation and usage guidelines, Arthan ensured the library’s adoption across safety and security teams, establishing a scalable foundation for risk assessment and governance alignment in AI environments.