Deep-Research Agents Can Be Poisoned via User-Generated Content
arXiv SecurityArchived May 26, 2026✓ Full text saved
arXiv:2605.24245v1 Announce Type: new Abstract: 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 pl
Full text archived locally
✦ AI Summary· Claude Sonnet
Computer Science > Cryptography and Security
[Submitted on 22 May 2026]
Deep-Research Agents Can Be Poisoned via User-Generated Content
Tingwei Zhang, Harold Triedman, Vitaly Shmatikov
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.
Subjects: Cryptography and Security (cs.CR)
Cite as: arXiv:2605.24245 [cs.CR]
(or arXiv:2605.24245v1 [cs.CR] for this version)
https://doi.org/10.48550/arXiv.2605.24245
Focus to learn more
Submission history
From: Tingwei Zhang [view email]
[v1] Fri, 22 May 2026 21:46:32 UTC (1,527 KB)
Access Paper:
HTML (experimental)
view license
Current browse context:
cs.CR
< prev | next >
new | recent | 2026-05
Change to browse by:
cs
References & Citations
NASA ADS
Google Scholar
Semantic Scholar
Export BibTeX Citation
Bookmark
Bibliographic Tools
Bibliographic and Citation Tools
Bibliographic Explorer Toggle
Bibliographic Explorer (What is the Explorer?)
Connected Papers Toggle
Connected Papers (What is Connected Papers?)
Litmaps Toggle
Litmaps (What is Litmaps?)
scite.ai Toggle
scite Smart Citations (What are Smart Citations?)
Code, Data, Media
Demos
Related Papers
About arXivLabs
Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)