Conference item
It's a TRAP! Task-redirecting agent persuasion benchmark for web agents
- Abstract:
- Web-based agents powered by large language models are increasingly used for tasks such as email management or professional networking. Their reliance on dynamic web content, however, makes them vulnerable to prompt injection attacks: adversarial instructions hidden in interface elements that persuade the agent to divert from its original task. We introduce the Task-Redirecting Agent Persuasion Benchmark (TRAP), a benchmark for studying how persuasion techniques misguide autonomous web agents on realistic tasks. Across six frontier models, agents are susceptible to prompt injection in 25% of tasks on average (13% for GPT-5 to 43% for DeepSeek-R1), with small interface or contextual changes often doubling success rates and revealing systemic, psychologically driven vulnerabilities in web-based agents. We also provide a modular social-engineering injection framework with controlled experiments on high-fidelity website clones, allowing for further benchmark expansion.
- Publication status:
- Published
- Peer review status:
- Peer reviewed
Actions
Access Document
- Files:
-
-
(Preview, Version of record, pdf, 11.5MB, Terms of use)
-
- Publication website:
- https://openreview.net/forum?id=Smk4fl6Q9E&utm
Authors
- Publisher:
- PMLR
- Host title:
- Proceedings of the 43rd International Conference on Machine Learning (ICML 2026)
- Series:
- Proceedings of Machine Learning Research
- Series number:
- 306
- Publication date:
- 2026-04-30
- Acceptance date:
- 2026-05-26
- Event title:
- International Conference on Machine Learning (ICML 2026)
- Event location:
- Seoul, South Korea
- Event website:
- https://icml.cc/
- Event start date:
- 2026-07-06
- Event end date:
- 2026-07-11
- ISSN:
-
2640-3498
- Language:
-
English
- Keywords:
- Pubs id:
-
2434023
- Local pid:
-
pubs:2434023
- Deposit date:
-
2026-06-16
- ARK identifier:
Terms of use
- Copyright holder:
- Korgul et al.
- Copyright date:
- 2026
- Rights statement:
- © 2026 the author(s). This work is under a Creative Commons Attribution 4.0 International License.
- Licence:
- CC Attribution (CC BY)
If you are the owner of this record, you can report an update to it here: Report update to this record