A red-teaming framework for identifying and testing prompt injection vulnerabilities in AI agent systems
Scoring was not performed