Boris Cherny on Claude Code Auto Mode: When the Classifier Replaces the Click-Through

On August 7, 2026, Anthropic announced that auto mode will become the default permission setting in Claude Code for Pro, Max, and Team plans starting August 14. Instead of asking developers to approve every shell command, Claude Code routes tool calls through a classifier that blocks actions deemed irreversible, destructive, or aimed outside the user's environment.

This article is compiled from Anthropic's public blog posts and Boris Cherny's public statements. It is not a private interview with Cherny.

Why auto mode is becoming the default

Claude Code Head Boris Cherny said 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!"

Anthropic reported that users approve 97% of permission prompts in Claude Code. In a controlled study with 1,053 paid testers, human review caught 13.6% of dangerous commands, while auto mode blocked 89%.

What changes on August 14

New sessions on Pro, Max, and Team plans will start in auto mode unless a user or admin pinned a different default. Enterprise and API users remain on manual mode for now; Anthropic plans to extend the default within the coming month.

The company also stopped charging Pro, Max, and Team users for classifier token overhead, effective immediately.

How the classifier works

Auto mode evaluates each tool call before execution. Broad allow rules that grant arbitrary code execution are set aside while auto mode is active. When blocked, Claude finds a safer path or asks the user; after three consecutive blocks or twenty in a session, Claude Code falls back to manual approvals.

Anthropic documented three internal incidents the classifier prevented, including an attempted mass process kill across roughly two thousand pods.

Production customers

Teams at Adobe, Nuro, Gusto, and Garner Health already run auto mode as their production default. Tomislav Reil, Director of Engineering at Adobe, said auto mode let Claude build and verify UI before engineers reviewed it.

Anthropic cautioned that auto mode relies on classification and does not eliminate risk entirely.

Sources

• Anthropic — Auto mode is now the default in Claude Code for Pro, Max, and Team plans (Aug. 7, 2026)
• Anthropic — Running auto mode in production (Aug. 7, 2026)
• TechCrunch — Anthropic is turning Claude Code's auto mode on by default (Aug. 9, 2026)