The launch of Anthropic’s coding device, Claude Code, is off to a rocky begin.
In line with reviews on GitHub, Claude Code’s auto-update perform contained buggy instructions that rendered some workstations unstable and damaged. When Claude Code was put in on the “root” or “superuser” ranges — permissions that give packages the flexibility to make working system-level adjustments — the buggy instructions would let functions modify usually restricted file directories and, within the worst-case situation, “brick” programs.
The problematic Claude Code auto-update instructions modified the entry permissions of sure essential system recordsdata. Permissions outline which packages and customers can learn or modify recordsdata, or run sure apps. One GitHub person mentioned that they have been compelled to make use of a “rescue occasion” to repair the permissions of recordsdata Claude Code’s instructions inadvertently broke.
Anthropic instructed information.killnetswitch it eliminated the problematic instructions from Claude Code and added a hyperlink in this system directing customers to a troubleshooting information. The hyperlink initially had a typo — however Anthropic says that’s been mounted, too.
