Archive — US — techcrunch
Anthropic’s Claude Code tool had a bug that ‘bricked’ some systems
Anthropic's launch of its coding tool, Claude Code, faced issues as a bug in its auto-update function caused some systems to become unstable and 'bricked.' The bug altered system file permissions, leading to potential access issues. Anthropic has since removed the faulty commands and corrected a typo in their troubleshooting link.
Key points
- Anthropic released a coding tool named Claude Code.
- A bug in the auto-update function contained commands that destabilized systems.
- The buggy commands modified critical system file permissions.
- If installed at root/superuser levels, the tool could 'brick' systems.
- Users had to utilize a 'rescue instance' to correct broken permissions.
- Anthropic removed the faulty commands and updated their troubleshooting guide link.
Why it matters
- This incident highlights ongoing concerns with AI tools managing system-level permissions, relevant amidst increasing reliance on automation in IT.
- Bricking devices due to software issues is not new, echoing past tech failures that affected user trust.
- As organizations adopt AI coding tools, this event raises questions about safety measures and quality control in software development.
The issues with Claude Code underscore the importance of rigorous testing in AI tools before deployment, as user safety and system integrity remain paramount.