Modern robots operate in complex and unpredictable environments and increasingly rely on AI to learn and make decisions. This makes robotics software harder to design, verify, and maintain with traditional engineering methods. ASTIR addresses this challenge by integrating trustworthy generative AI into every stage of the robotics software engineering lifecycle, from requirements engineering to system maintenance. The open-source, AI-augmented workbench aims to improve productivity, collaboration, and reliability in robotics development. The project will validate its technology through three open-access use cases in service robotics, manufacturing robotics, and drone systems.