Auto mode is now the default in Claude Code for Pro, Max, and Team plans
· Source: Simon Willison
The company Anthropic has announced that automatic mode will become the default setting in Claude Code for its Pro, Max, and Team plans. This decision is based on their confidence in automatic mode’s ability to mitigate risks associated with command injection and data exfiltration. A test involving 1,053 users found that automatic mode blocked 89% of malicious actions, whereas only 13.6% of human users rejected them. Anthropic has also conducted security tests with a third-party firm, Trajectory Labs, which was unable to successfully breach Claude Code in automatic mode. While these results are encouraging, some experts remain skeptical and consider it necessary to have independent confirmation to ensure the technology’s security. The safety of coding agents is a significant concern, as they may have access to sensitive data and tools, and it is crucial to find ways to execute them securely to prevent damage. This is particularly relevant in the context of applied artificial intelligence in commerce, where security and reliability are essential for operational success.
Read the original article on Simon Willison
This summary is an informational synthesis produced by dataqbs.com. All rights to the original content belong to its author and the cited media outlet. We act solely as curators of technology news and claim no authorship.