🤖 Paperclip AI import flaws enable host command execution

🤖 Paperclip AI import flaws enable host command execution Researchers identified vulnerabilities in Paperclip AI that let attackers execute commands on a host system by importing malicious agents. The issue centers on the agent import path, where crafted content can trigger command execution during handling. The finding is significant because agent import is a routine trust boundary in AI tooling. If malicious imports reach developer or production environments, the platform can become an…