Deep-Research Agents Can Be Poisoned via User-Generated Content
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_ | 1866916041224880128 |
|---|---|
| author | Zhang, Tingwei Triedman, Harold Shmatikov, Vitaly |
| author_facet | Zhang, Tingwei Triedman, Harold Shmatikov, Vitaly |
| contents | Deep-research agents, i.e., systems that rely on multi-agent pipelines to iteratively retrieve, synthesize, and cite Web content in order to produce structured reports, are rapidly replacing traditional search for both routine and complex information needs. These agents issue many related queries during a single research session. We show that for many common search topics, they repeatedly retrieve the same user-generated content (UGC) pages from platforms such as Reddit and Wikipedia. Next, we argue that this retrieval overlap creates a concentrated attack surface: an adversary who appends a short, crafted text to a single, frequently retrieved UGC page can cause the agent to cite attacker-chosen content and promote attacker-chosen entities across many related queries.
We evaluate this attack on three representative deep-research systems (STORM, Co-STORM, and OmniThink) across multiple query clusters. We also study defenses at different stages of the pipeline, including source-level filtering and output-based detection. Our findings highlight a fundamental vulnerability in how deep-research agents retrieve and integrate web content. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_24245 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Deep-Research Agents Can Be Poisoned via User-Generated Content Zhang, Tingwei Triedman, Harold Shmatikov, Vitaly Cryptography and Security Deep-research agents, i.e., systems that rely on multi-agent pipelines to iteratively retrieve, synthesize, and cite Web content in order to produce structured reports, are rapidly replacing traditional search for both routine and complex information needs. These agents issue many related queries during a single research session. We show that for many common search topics, they repeatedly retrieve the same user-generated content (UGC) pages from platforms such as Reddit and Wikipedia. Next, we argue that this retrieval overlap creates a concentrated attack surface: an adversary who appends a short, crafted text to a single, frequently retrieved UGC page can cause the agent to cite attacker-chosen content and promote attacker-chosen entities across many related queries. We evaluate this attack on three representative deep-research systems (STORM, Co-STORM, and OmniThink) across multiple query clusters. We also study defenses at different stages of the pipeline, including source-level filtering and output-based detection. Our findings highlight a fundamental vulnerability in how deep-research agents retrieve and integrate web content. |
| title | Deep-Research Agents Can Be Poisoned via User-Generated Content |
| topic | Cryptography and Security |
| url | https://arxiv.org/abs/2605.24245 |