Comment by boratanrikulu

Allowing LLMs to execute unrestricted commands without human review is risky and insecure.