Anthropic is making it easier for developers to use Claude Code with less manual oversight, announcing that its auto mode will become the default setting for Pro, Max and Team accounts from August 14.
The move marks a broader shift toward more autonomous AI-assisted programming. Instead of stopping to request user approval before taking individual actions, Claude Code will generally continue working on a task unless it determines that an action is irreversible, destructive or targets systems outside the user’s environment.
Anthropic first introduced auto mode as a test feature in March, positioning it as a way to strike a balance between programming speed and user control.
Testing Shows Higher Safety
Anthropic says its decision to make auto mode the default is supported by results from testing involving 1,053 paid users. According to the company, auto mode detected 89% of harmful actions, compared with 13.6% identified through manual review.
The company suggested that human oversight can become less effective when users repeatedly encounter permission requests. In Claude Code, users reportedly approve about 97% of permission prompts, potentially turning the review process into a routine action rather than a meaningful safety check.
Boris Cherny, head of Claude Code, said on X that he and his team have been using auto mode exclusively for months, adding that he could not imagine returning to permission prompts.
More Safeguards for Autonomous Coding
While Anthropic is reducing the need for constant approvals, the company is also adding safeguards designed to limit the risks associated with greater autonomy.
Among the new protections are prompt-injection screening and customizable hard-deny rules. These controls are intended to prevent potentially dangerous actions, including attempts to extract or exfiltrate sensitive data.
The approach reflects Anthropic’s effort to give Claude Code greater freedom to execute programming tasks while maintaining boundaries around high-risk operations.
For developers, the change could make coding workflows faster and more fluid by removing repeated approval interruptions. At the same time, the default shift places greater importance on the safeguards built into the system, as Claude Code will increasingly be expected to determine which actions are safe to execute independently.
Anthropic’s move signals the growing emphasis on autonomous software development, where AI coding tools are shifting from assistants that wait for instructions toward systems capable of completing larger portions of development work on their own.
Also Read :- Anthropic AI and Carbon Emissions: What Businesses Should Consider
