TY  - RPRT
TI  - Evo-Attacker: Memory-Augmented Reinforcement Learning for Long-Horizon Tool Attacks on LLM-MAS
AU  - Bingyu Yan
AU  - Xiaoming Zhang
AU  - Jinyu Hou
AU  - Chaozhuo Li
AU  - Ziyi Zhou
AU  - Yiming Hei
AU  - Litian Zhang
PY  - 2026
UR  - https://arxiv.org/abs/2605.25389
ID  - 2605.25389
ER  -