Anthropic Turns On Auto Mode by Default for Claude Code as AI Coding Goes Autonomous

Anthropic turned on Auto Mode by default for Claude Code as AI coding goes autonomous. Anthropic announced the change in a quiet but consequential update rolling out this week. Previously, users had to explicitly enable Auto Mode, a setting that lets Claude act as an autonomous agent rather than a turn-by-turn assistant. Now, Auto Mode is on by default for all new sessions. In practice, that means Claude Code no longer waits for you to approve every file edit, terminal command, or debugging step. Give it a high-level prompt like refactor the authentication module to support OAuth 2.0 and add tests, and it will create a plan, explore your codebase, write the code, run tests, fix failures, and loop until the task is complete, all without pausing for permission. Users can still drop back into manual, step-by-step control or set permission boundaries, but the default behavior is now full autonomy. Anthropic says the change reflects how most developers were already using the tool. According to the company, power users had overwhelmingly opted into Auto Mode and reported significantly faster task completion when Claude was allowed to run uninterrupted.