IAMRAGBENCH: Benchmarking Authorization Expansions in LLM-Generated AWS IAM Policies under RAG Poisoning

Hasan Kaplan, Indika Kumara, Willem-Jan van den Heuvel +1

Large language models (LLMs) using retrieval-augmented generation (RAG) can synthesize Amazon Web Services (AWS) Identity and Access Management (IAM) policies from tasks and operat…