This repository maintains a comprehensive collection of research papers on textual adversarial attacks and defenses in natural language processing. It organizes papers across multiple perturbation levels: sentence-level, word-level, and character-level attacks, along with defense methods, certified robustness, and benchmarks. The list also includes open-source toolkits for adversarial text generation and analysis, such as OpenAttack and RobustQA, making it a reference resource for NLP security research.