Anthropic, a leading AI developer, is changing its default settings for the widely used coding tool Claude Code. Starting August 14, the tool will automatically enable Auto Mode for most users, which is designed to detect and prevent potentially hazardous commands. According to internal tests, the Auto Mode was more effective than human reviewers in identifying problematic inputs. This shift is expected to alter the way developers interact with the tool, with more focus on monitoring AI-generated output rather than writing code. The change aims to improve the overall safety of the platform, reducing the risk of unintended consequences.