Anthropic will make auto mode the default for Claude Code on Pro, Max, and Team accounts starting August 14, 2026. The company announced the change on Friday, August 9, 2026. Auto mode lets the AI programming tool proceed without human approval at each step, unless an action is "irreversible, destructive, or aimed outside your environment."
What Auto Mode Changes
Claude Code typically requires human approval before each action. That manual checkpoint is now being replaced by default. Anthropic first unveiled a test version of auto mode in March. The company says the shift is designed to balance speed and control for developers.
Boris Cherny, Head of Claude Code at Anthropic, praised the mode on X. "The team and I use Auto mode exclusively, and have been for many months. I couldn't imagine going back to permission prompts!" he wrote.
Safety Data From Testing
Anthropic ran a study with 1,053 paid testers to compare auto mode against manual review. In that study, auto mode caught 89% of harmful actions. Human review caught only 13.6% of harmful actions in the same test. The company claims this shows auto mode proved safer than manual review in testing.
The gap is stark. Auto mode caught more than six times as many harmful actions as humans did. Anthropic suggests that manual review can become habitual, which may explain the low catch rate. "Manual review can become habitual: users approve 97% of permission prompts in Claude Code," the company said.
New Safety Features
To address risks like data exfiltration, Anthropic added two new safety features. The first is prompt injection screening, which scans for malicious instructions hidden in inputs. The second is customizable hard deny rules, which let users block specific actions outright. These rules aim to prevent data from leaving the environment without permission.
Stay ahead of the AI curve
The most important updates, news, and content — delivered weekly.
No spam. Unsubscribe anytime.
The features are designed to work alongside auto mode's existing safeguards. Actions that are irreversible, destructive, or aimed outside the environment still require human approval. That boundary remains in place even with auto mode as the default.
Why the Change Matters
The high approval rate for permission prompts suggests users rarely reject actions. Anthropic argues this makes manual review more of a formality than a real safety check. Auto mode, by contrast, applies automated screening consistently across every action.
Cherny's endorsement carries weight within the company. His team has used auto mode exclusively for months, according to his post. That internal adoption predates the public default change.
Rollout Details
The default change applies to Pro, Max, and Team accounts. Free accounts are not mentioned in the announcement. The timeline is short: the announcement came on August 9, 2026, and the change takes effect five days later on August 14, 2026.
Anthropic's announcement did not include independent analysis or criticism of the change. The company's claims rest on its own study data. Developers will be able to see the results in practice starting next week.