Claude Code Safety Rules Can Be Bypassed with Long Subcommand Chains
Security researchers found that Anthropic's Claude Code agent will ignore its safety deny rules if burdened with a sufficiently long chain of subcommands.
Security researchers found that Anthropic's Claude Code agent will ignore its safety deny rules if burdened with a sufficiently long chain of subcommands.
A prompt injection vulnerability has been discovered in Anthropic's new Claude Cowork AI, which could allow attackers to exfiltrate sensitive files from users' accounts.