Anthropic has announced a new research preview feature called Auto Mode for Claude Code, set to be released by March 12, 2026, which aims to streamline the coding process by automatically managing permission prompts. This feature addresses concerns that arose from the previous necessity of using –dangerously-skip-permissions, which removed safety checks, leading to developers opting for unsafe workarounds. With Auto Mode, Claude will intelligently handle permission choices while still blocking potential threats like prompt injections, allowing developers to focus on their tasks without constant interruptions. Since Auto Mode is still in its preview phase, it is advised to run it in isolated environments for safety.
Anthropic unveils Auto Mode for Claude code to automate permission handling
