IntelliJ IDEA

April flew by. The pace of tech development didn’t slow, and the flow of news and knowledge didn’t either. This month, Emily Bache joins us to share some sharp finds about AI agents and test-driven development. Java stays busy with fresh updates and practical tips, and Kotlin keeps pushing forwar…

Learn how AI Agent Skills turn one-off prompts into reusable expertise. This article shows how to combine Skills with developer tools to reliably track down flaky test root causes and make AI debugging more deterministic.

向企业后端添加 AI 智能体不应以破坏架构为代价。如果您的核心系统是使用 Java 构建的，编排 LLM 并不需要您引入单独的 Python 微服务或重写技术栈。 今天，我们正式发布 Koog for Java。Koog 最初是为了满足 JetBrains 自身活动扩展需求而构建的，它使用结构化、可观测且具备容错能力的工作流取代了不可预测、临时的提示更改方法。 如今，JVM 最强大的智能体框架之一拥有了完全符合 Java 惯用风格的 API。您的 Java 团队可以直接在现有后端中构建可靠的 AI 智能体，借助流畅的构建器风格 API、线程池执行器和原生 Java 抽象，完全不需要担心任何与 …

Agent Skills is a new initiative that solves both problems by managing context progressively and extending AI agent capabilities on demand.

Build reliable AI agents natively on the JVM. Koog for Java provides idiomatic builders, persistence, and observability – no Python microservices required.