3 weeks ago
Anthropic Makes Claude Code's Auto Mode Default for Paid Users
Anthropic made a computer helper called Claude Code that helps people write software.
Before, the helper always asked for permission before doing anything.
Now, for many people who pay for it, the helper will just do the work on its own.
This is called Auto mode.
The helper has a special safety checker that stops actions that could hurt your computer or files.
In a big test, the helper stopped 89 out of every 100 dangerous actions, which was more than people did on their own.
Anthropic says people sometimes press 'approve' without checking carefully, but the safety checker stays alert.
People can still change a setting to make the helper ask questions again by pressing Shift+Tab.
The helper can also work for many hours without stopping, which is good for big tasks.
Big companies on the Enterprise plan get to choose whether to use it for now.
Anthropic will make Claude Code's Auto mode the default for Pro, Max and Team users starting 14th August.
In Auto mode, an independent classifier verifies each command and blocks destructive, irreversible or outside-the-environment actions instead of pausing for permission prompts.
Users who already set a default permission mode will get a one-time in-app notification and can switch back using the Shift+Tab shortcut.
In a study of 1,053 paid testers, Auto mode blocked 89% of dangerous commands, while human detection was reported as 13.6% in two articles and 11% in another, a discrepancy between sources.
Auto mode remains opt-in for Claude Enterprise, the Claude API, Claude Platform on AWS, Amazon Bedrock, Google Cloud's Agent Platform and Microsoft Foundry, with a default rollout planned in the coming months.
- Who
- Anthropic, the AI company behind Claude Code; its Pro, Max and Team users; and Boris Cherny, Anthropic's Head of Claude Code.
- What
- Claude Code's Auto mode becomes the default, allowing the AI coding tool to act autonomously without pausing for permission prompts at every step.
- Where
- Not specified in the articles; the change applies to Anthropic's Claude Code software product.
- When
- Starting 14th August.
- Why
- Anthropic says Auto mode 'matched or beat manual permission review on every safety measure we tracked', blocked 89% of dangerous commands in testing versus lower human detection rates, and lets Claude run uninterrupted for hours on large tasks.
Key facts
- Company
- Anthropic
- Product
- Claude Code
- Change
- Auto mode becomes the default for Pro, Max and Team users
- Effective date
- 14th August
- Revert shortcut
- Shift+Tab
- Tester study
- 1,053 paid testers; Auto mode blocked 89% of dangerous commands (human detection cited as 13.6% or 11%)
- Safety rule
- Manual approvals required after 3 consecutive or 20 total blocked commands in a session
- Opt-in platforms
- Claude Enterprise, Claude API, Claude Platform on AWS, Amazon Bedrock, Google Cloud's Agent Platform, Microsoft Foundry
Quotes
Boris Cherny
Anthropic’s Head of Claude Code
“Auto mode also lets Claude work autonomously for longer stretches. This makes models built for long‑running work, like Claude Opus 5, more practical to leave running for hours on large tasks.”
livemint.com
“The team and I use Auto mode exclusively, and have been for many months. I couldn't imagine going back to permission prompts! Really excited to get this out to everyone.”
livemint.com
thehansindia.com







