Agentic AI is reshaping software engineering, but DSLs remain challenging for general-purpose AI models. This article introduces new agent skills, evaluation capabilities, and tooling that help AI agents work more effectively with specialised Languages.