An in-depth benchmark dedicated to evaluating LLM Agents. Built on an isolated sandbox environment with 30 advanced tasks, it comprehensively covers five core business scenarios: office collaboration, information retrieval, content creation, data analysis, and software engineering.