ADAM: A Systematic Data Extraction Attack on Agent Memory via Adaptive Querying
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866911584393101312 |
|---|---|
| author | Lyu, Xingyu He, Jianfeng Wang, Ning Hu, Yidan Li, Tao Chen, Danjue Li, Shixiong Chen, Yimin |
| author_facet | Lyu, Xingyu He, Jianfeng Wang, Ning Hu, Yidan Li, Tao Chen, Danjue Li, Shixiong Chen, Yimin |
| contents | Large Language Model (LLM) agents have achieved rapid adoption and demonstrated remarkable capabilities across a wide range of applications. To improve reasoning and task execution, modern LLM agents would incorporate memory modules or retrieval-augmented generation (RAG) mechanisms, enabling them to further leverage prior interactions or external knowledge. However, such a design also introduces a group of critical privacy vulnerabilities: sensitive information stored in memory can be leaked through query-based attacks. Although feasible, existing attacks often achieve only limited performance, with low attack success rates (ASR). In this paper, we propose ADAM, a novel privacy attack that features data distribution estimation of a victim agent's memory and employs an entropy-guided query strategy for maximizing privacy leakage. Extensive experiments demonstrate that our attack substantially outperforms state-of-the-art ones, achieving up to 100% ASRs. These results thus underscore the urgent need for robust privacy-preserving methods for current LLM agents. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_09747 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | ADAM: A Systematic Data Extraction Attack on Agent Memory via Adaptive Querying Lyu, Xingyu He, Jianfeng Wang, Ning Hu, Yidan Li, Tao Chen, Danjue Li, Shixiong Chen, Yimin Cryptography and Security Artificial Intelligence Large Language Model (LLM) agents have achieved rapid adoption and demonstrated remarkable capabilities across a wide range of applications. To improve reasoning and task execution, modern LLM agents would incorporate memory modules or retrieval-augmented generation (RAG) mechanisms, enabling them to further leverage prior interactions or external knowledge. However, such a design also introduces a group of critical privacy vulnerabilities: sensitive information stored in memory can be leaked through query-based attacks. Although feasible, existing attacks often achieve only limited performance, with low attack success rates (ASR). In this paper, we propose ADAM, a novel privacy attack that features data distribution estimation of a victim agent's memory and employs an entropy-guided query strategy for maximizing privacy leakage. Extensive experiments demonstrate that our attack substantially outperforms state-of-the-art ones, achieving up to 100% ASRs. These results thus underscore the urgent need for robust privacy-preserving methods for current LLM agents. |
| title | ADAM: A Systematic Data Extraction Attack on Agent Memory via Adaptive Querying |
| topic | Cryptography and Security Artificial Intelligence |
| url | https://arxiv.org/abs/2604.09747 |