Preprint · 2026 · Open access
AgentCIScope: Grounded Contextual Integrity Evaluation for LLM Agents
arXiv (Cornell University) · Published 13 October 2026 · DOI 10.48550/arxiv.2610.19158
In brief
“AgentCIScope: Grounded Contextual Integrity Evaluation for LLM Agents” is a 2026 preprint by Chengquan Guo and 4 co-authors, published in arXiv (Cornell University). OpenAlex counts 0 citations as of 19 November 2026. It is open access under a CC BY licence: the full text is free to read at arXiv (Cornell University).
- Citations
- 0
- References
- 10
- Authors
- 5
Abstract
LLM agents are becoming increasingly capable of handling daily workspace tasks by interacting with application environments through tool calls. However, as agents read and write sensitive user data across email, documents, and other applications, a central privacy question is whether they preserve contextual integrity (CI), that is, whether the information flows they produce comply with context-specific disclosure constraints. Existing work either focuses on model-level privacy leakage or evaluates agents in limited application environments with manually constructed disclosure constraints. In this work, we introduce AgentCIScope, a unified platform and benchmark for evaluating agent-level contextual integrity through executable, multi-step tasks across diverse, realistic application environments. AgentCIScope transforms public and synthetic privacy cases into executable agent tasks, with disclosure constraints grounded in three criteria: governing privacy policies, persona-based agreement, and a predefined sensitive-attribute taxonomy. Based on AgentCIScope, we evaluate 12 agents on 936 tasks and find that strong task-completion capability does not guarantee contextual-integrity preservation: single-run leakage rates range from 32.5% to 94.4%, while repeating each task five times further increases leakage by 4-19 percentage points. Leakage varies substantially across attributes: personal preferences and booking details have average leakage rates of 84.7% and 76.5%, compared with 17.1% for credit-card information and 29.4% for SSNs; environments: calendar has the highest average leakage rate as an information source, while WhatsApp ranks highest as an information destination; and governing privacy policies. AgentCIScope provides deep insights into agent privacy leakage and serves as a unified testbed for developing privacy-preserving agents.
Cite this paper
Formatted from the catalogue record. Check names, capitals and page numbers against the paper before you submit.
APA 7
Guo, C., Chen, Z., Lin, Z., Liu, X., & Li, B. (2026). AgentCIScope: Grounded Contextual Integrity Evaluation for LLM Agents. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2610.19158
MLA 9
Guo, Chengquan, et al. “AgentCIScope: Grounded Contextual Integrity Evaluation for LLM Agents.” arXiv (Cornell University), 2026. https://doi.org/10.48550/arxiv.2610.19158.
Chicago
Guo, Chengquan, Zhaorun Chen, Zinan Lin, Xiaodong Liu, and Bo Li. 2026. “AgentCIScope: Grounded Contextual Integrity Evaluation for LLM Agents.” arXiv (Cornell University). https://doi.org/10.48550/arxiv.2610.19158.
Harvard
Guo, C. et al. (2026) ‘AgentCIScope: Grounded Contextual Integrity Evaluation for LLM Agents’, arXiv (Cornell University). doi:10.48550/arxiv.2610.19158.
IEEE
C. Guo, Z. Chen, Z. Lin, X. Liu, and B. Li, “AgentCIScope: Grounded Contextual Integrity Evaluation for LLM Agents,” arXiv (Cornell University), 2026, doi: 10.48550/arxiv.2610.19158.
BibTeX
@misc{guo2026agentciscope,
title = {{AgentCIScope: Grounded Contextual Integrity Evaluation for LLM Agents}},
author = {Guo, Chengquan and Chen, Zhaorun and Lin, Zinan and Liu, Xiaodong and Li, Bo},
howpublished = {arXiv (Cornell University)},
year = {2026},
doi = {10.48550/arxiv.2610.19158},
url = {https://doi.org/10.48550/arxiv.2610.19158}
}
Download for a reference manager: BibTeX (.bib) · RIS for Zotero, Mendeley and EndNote (.ris)
Topic and keywords
Physical Sciences › Computer Science › Artificial Intelligence › Privacy-Preserving Technologies in Data
- contextual integrity
- LLM agents
- privacy leakage
- privacy policies
- sensitive attributes
- tool calls
- information flow
Work with this paper
Most-cited papers in Computer Science
- ImageNet classification with deep convolutional neural networksAlex Krizhevsky, Ilya Sutskever, Geoffrey E. Hinton · Communications of the ACM · 2017 · 114,726 citations
- Long Short-Term MemorySepp Hochreiter, Jürgen Schmidhuber · Neural Computation · 1997 · 109,097 citations
- A Mathematical Theory of CommunicationClaude E. Shannon · Bell System Technical Journal · 1948 · 91,736 citations
- Deep learningYann LeCun, Yoshua Bengio, Geoffrey E. Hinton · Nature · 2015 · 90,226 citations
- Faster R-CNN: Towards Real-Time Object Detection with Region Proposal NetworksShaoqing Ren, Kaiming He, Ross Girshick and 1 more · IEEE Transactions on Pattern Analysis and Machine Intelligence · 2016 · 74,143 citations
Questions about this paper
How do I cite this paper in APA style?
Guo, C., Chen, Z., Lin, Z., Liu, X., & Li, B. (2026). AgentCIScope: Grounded Contextual Integrity Evaluation for LLM Agents. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2610.19158
How many times has this paper been cited?
0 times, according to OpenAlex on 19 November 2026. Google Scholar and Scopus count different sets of sources, so their numbers differ.
Is this paper free to read?
Yes. It is open access at arXiv (Cornell University): sign in and open the PDF from this page.
Who wrote this paper?
Chengquan Guo, Zhaorun Chen, Zinan Lin, Xiaodong Liu, Bo Li.
What is the DOI of this paper?
10.48550/arxiv.2610.19158, permanently at https://doi.org/10.48550/arxiv.2610.19158. Sign in and open it from this page.
Bibliographic data from OpenAlex (CC0), last checked 19 November 2026. ResearchGuru does not host this paper; once you are signed in, the links open it at the publisher or the repository that does.