A critical vulnerability has been discovered in Amazon Kiro, allowing attackers to execute prompt injection attacks through maliciously crafted workspace content. When users interact with the Kiro agent after opening a compromised workspace, sensitive data can be automatically exfiltrated using Kiro's integrated powers. This security flaw poses significant risks to enterprise users handling confidential information and highlights the importance of validating third-party workspace content before deployment.

Summary auto-generated by AI from the original publisher's content. Editorial standards.