Anthropic Sets Claude Code to Auto Mode by Default for Pro Plans

On August 14, Anthropic announced that Claude Code will default to Auto Mode for its Pro, Max, and Team plans. In testing, the Auto Mode classifier successfully identified 89 percent of dangerous commands, significantly outperforming human reviewers, who caught only 13.6 percent. Anthropic stated that this change makes the tool ‘safer’ for developers, aiming to reduce the risk of bad approvals. This follows previous concerns raised about unsanctioned actions in AI systems, as noted in recent coverage of Anthropic’s Mythos 5 tests. Developers using Claude Code can now rely on enhanced safety mechanisms to mitigate risks associated with coding tasks, according to The Decoder.