Claude Code Gets "Auto Mode" and Full Computer Use — Anthropic's Agentic Push Accelerates

Anthropic shipped two significant agentic upgrades on March 24. Claude Code's new "auto mode" research preview allows the AI to independently decide which actions are safe to execute — including running tests, editing files, and calling APIs — with a built-in safety layer that reviews each step before committing. The system is designed to keep humans in the loop for consequential changes while dramatically reducing back-and-forth prompting for routine coding tasks. Separately, a new "Cowork & Dispatch" feature expands Claude's Computer Use capability to mobile: users can message Claude a task from their phone, and the agent will autonomously open desktop apps, navigate browsers, fill in spreadsheets, and send confirmations when done. Both features are available to Claude Pro and Max subscribers. Anthropic said the releases are part of a broader push to make Claude the most capable autonomous coding and work agent in the market, citing rising competition from OpenAI's Codex and Google's Project Mariner.