Tag: #Claude Code

Claude Code with Administrator Rights: Pentagon Contractor Leaks Military Network Map to Open Internet

Technology 10/30/2026, 18:31

An Accenture employee granted the AI agent Claude Code administrative rights in the Pentagon's NIPRNet. Credentials and the military network map leaked to an external site, and the vulnerability went unpatched for two months.

Anthropic adds an «emergency» safety switch to Claude Code: AI agent will be able to delete files without confirmation

Technology 10/28/2026, 17:43

In Claude Code v2.1.311, Anthropic has introduced an environment variable allowing the AI agent to delete files without confirmation. The feature removes the final restriction of the bypassPermissions mode and was confirmed by a company employee.

Anthropic Makes Code Autopilot the Standard: Why 'Automatic' Action Confirmation Turned Out to Be More Dangerous Than Its Absence

Technology 08/11/2026, 04:30

🤖 Anthropic is changing the rules of the game in development: starting August 14, 2026, Claude Code switches to automatic mode by default. Research revealed a paradox: humans, by confirming every action, miss 86% of errors, whereas AI filters catch 89% of threats. Manual control turned out to be an illusion of security. 🚀