Prompt Injection Experiments with Opus-5 in Claude Code – Auto-Mode Edition

4 points by veganmosfet 6 days ago | 2 comments

Very interesting! Can a kernel-level policy mitigate this?

veganmosfet 6 days ago | |

Thanks! Imho using sandboxing would mitigate most of the issues. In the presence of adversarial content, auto-mode is not enough.

Very interesting! Can a kernel-level policy mitigate this?

veganmosfet 6 days ago | |

Thanks! Imho using sandboxing would mitigate most of the issues. In the presence of adversarial content, auto-mode is not enough.