📄 中文摘要
Anthropic在‘AI for code’的激烈竞争中，推出了Claude Code，旨在超越简单的代码自动补全，提供更为深刻的代码理解能力。该技术不仅能生成代码，还能分析和理解复杂系统的运行状态，例如Kubernetes集群的异常情况。通过高级的代理多步骤推理循环，Claude Code能够自动化复杂的安全研究，提升开发者在安全性和效率上的能力。这一创新标志着AI技术在代码审查和安全研究领域的重大进步。
Anthropic推出Claude Code代码审查，通过高级代理多步骤推理循环自动化复杂安全研究
📄 English Summary
Anthropic Introduces Code Review via Claude Code to Automate Complex Security Research Using Advanced Agentic Multi-Step Reasoning Loops
Anthropic has launched Claude Code in the competitive landscape of 'AI for code', aiming to go beyond mere code autocompletion to provide deeper code comprehension capabilities. This technology not only generates code but also analyzes and understands the operational status of complex systems, such as the anomalies in Kubernetes clusters. Utilizing advanced agentic multi-step reasoning loops, Claude Code automates complex security research, enhancing developers' capabilities in security and efficiency. This innovation marks a significant advancement in AI technology for code review and security research.