
Share
Auto Mode streamlines long-running tasks by automating permission checks and adding safeguards, making it easier and safer to manage complex operations without constant manual intervention.
Today, Claude Code has launched a new feature called "Auto Mode," designed to provide a safer and more efficient way to handle long-running tasks. This mode allows Claude to make permission decisions on your behalf while incorporating safeguards that monitor actions before they execute. It's currently available as a research preview for Team plan users, with plans to roll it out to Enterprise and API users in the coming days.
Claude Code's default permissions are intentionally conservative, requiring approval for every file write and bash command. This ensures safety but can be cumbersome for long-running tasks, which often require frequent human approvals. Some developers have resorted to using --dangerously-skip-permissions to bypass these checks, but this approach can lead to dangerous outcomes if not used in isolated environments.
Auto Mode offers a balanced solution by allowing longer tasks to run with fewer interruptions while maintaining a level of safety. Here’s how it works:
While Auto Mode reduces risk compared to --dangerously-skip-permissions, it doesn't eliminate all risks. The classifier may still allow some risky actions if user intent is ambiguous or if Claude lacks sufficient context about your environment. Conversely, it might occasionally block benign actions. We will continue to refine and improve the experience over time.

Auto Mode might have a small impact on token consumption, cost, and latency for tool calls. However, these trade-offs are generally minimal compared to the enhanced safety and efficiency it provides.
Auto Mode is now available in Claude Code as a research preview for Team plan users. It will roll out to Enterprise and API users in the coming days. The feature works with both Claude Sonnet 4.6 and Opus 4.6.
"disableAutoMode": "true" in your configuration file.Auto Mode is a significant step forward in making long-running tasks safer and more efficient. By balancing automation with safety checks, it offers a practical solution for developers looking to streamline their workflows without compromising on security. Stay tuned for further updates and improvements as we continue to enhance this feature.
Tags
Original Sources
About the author
Kai built ML infrastructure at a Bay Area startup before developing an obsession with transformer architectures and inference optimisation that eventually pulled him out of product work entirely. A stint at a compute research lab sharpened his instinct for what actually matters in a model release versus what is marketing. He writes from the inside — from the perspective of someone who has debugged the systems he is describing at three in the morning. He is allergic to hype and instinctively drawn to the unglamorous plumbing questions that everyone else skips over.
More from The Engineer →This Week's Edition
25 March 2026
133 articles
Related Articles

Smarter Engagement for Stronger Growth: How Payers Can Leverage AI to Do More with Less
Products & Applications · 3 min

Penn Medicine and K Health Deploy AI Clinical Agents to Enhance Patient Care
Products & Applications · 3 min

Wheel and b.well Partner to Build Turnkey AI-First Virtual Care Infrastructure
Products & Applications · 3 min
Related Articles

Smarter Engagement for Stronger Growth: How Payers Can Leverage AI to Do More with Less
Products & Applications · 3 min

Penn Medicine and K Health Deploy AI Clinical Agents to Enhance Patient Care
Products & Applications · 3 min

Wheel and b.well Partner to Build Turnkey AI-First Virtual Care Infrastructure
Products & Applications · 3 min
More Stories