Spring AI Prompt Injection Vulnerability via Memory Poisoning

A prompt injection vulnerability has been identified in Spring AI versions 1.0.0 through 1.0.x and 1.1.0 through 1.1.x. This vulnerability allows a malicious user to craft input that is stored in conversation memory and later interpreted by the model in an unintended manner. Applications using the affected advisor with user-controlled input may be susceptible to manipulation of model behavior across conversation turns.

Impact

Exploitation of this vulnerability could lead to unauthorized manipulation of model behavior, allowing for prompt injection attacks that could disrupt the intended functionality of the application.

Remediation

Users of affected Spring AI versions should upgrade to version 1.0.7 or 1.1.6, depending on their current version.