Anthropic is turning Claude Code’s auto mode on by default
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Anthropic is making auto mode the default in Claude Code for Pro, Max, and Team accounts starting August 14, requiring less human oversight during programming sessions.
- In auto mode, Claude Code proceeds with actions unless they're "irreversible, destructive, or aimed outside your environment," skipping the usual per-step approval prompts.
- Anthropic said auto mode proved safer than manual review in a study of 1,053 paid testers — auto mode caught 89% of harmful actions versus just 13.6% for human review.
- Anthropic attributed humans' poor performance to habituation, noting that "manual review can become habitual: users approve 97% of permission prompts in Claude Code."
- Boris Cherny, Claude Code's head, said on X that he and his team use auto mode exclusively and "couldn't imagine going back to permission prompts."
- Anthropic has added new safety features alongside the rollout, including prompt injection screening and customizable hard deny rules to prevent things like data exfiltration.
Why it matters: For Pro, Max, and Team subscribers — the bulk of Claude Code's professional users — every coding session starting August 14 will run with minimal human checkpoints by default. Anthropic's own numbers suggest that's not reckless: a 1,053-person test found auto mode caught 89% of harmful actions versus 13.6% for human review, which the company pins on habitual rubber-stamping (97% approval rate).
Ask SkimNews
