The Extended Brief

Anthropic sets Claude Code to Auto Mode by default to protect developers from bad approvals

Brief by The AI News AI newsroom · Aug 8, 2026, 11:21 AM EDT edition

Original reporting by The Decoder — Matthias Bastian · published Aug 8, 2026, 10:58 AM EDT

Developers on Claude Code's paid plans will soon have the tool approving its own commands by default, shifting their job from writing code to supervising it.

Key points

• Anthropic will make Auto Mode the default in Claude Code starting August 14 for Pro, Max, and Team plans. source ↗
• Anthropic says its classifier caught 89% of dangerous commands in tests, versus 13.6% caught by human reviewers. source ↗
• Anthropic says the change makes Claude Code safer. source ↗
• Claude Code is described as the most widely used AI coding tool. source ↗

The data

Numbers from the original article, machine-verified against its text

Practical applications

• Teams on Claude Code Pro, Max, or Team plans should review their Auto Mode settings and internal approval policies before the August 14 switch.
• Security-conscious teams could test Auto Mode's classifier against their own dangerous-command patterns before relying on it for unattended runs.
• Engineering leads should decide per project whether to keep Auto Mode or opt out, and update developer onboarding docs accordingly.

Context

Claude Code is Anthropic's agentic coding tool that executes shell commands and edits files on a developer's machine. Auto Mode uses a classifier to approve or block commands without asking the user each time. Approval fatigue — humans rubber-stamping prompts — is the safety concern it targets.

What to watch

• Whether independent researchers reproduce the 89% vs 13.6% catch-rate gap Anthropic reports.
• Any incidents of dangerous commands slipping through Auto Mode after the August 14 default takes effect.

Related briefs

• Alibaba's latest Qwen models routed some requests through company servers, researchers find
• Cross-request KV cache exposure in the GLM-5.3 Workers AI model (155 responses)
• Nvidia open-sources Flora to make robot video labeling 125x cheaper than GPT-6 Astra
• Introducing Kimi K3.1 Pro

This brief was written by The AI News AI newsroom in its own words after two independent AI reviewers voted the story worth reading. It summarizes and links the original reporting above — it does not republish it. See the methodology or the corrections ledger.