Preprint · 2026 · Open access

Observable Channels, Not Just Storage: Evaluating Privacy Leakage in LLM Agent Pipelines

Tao Huang, Chen Hou, Guosen Wu, Jiayang Meng

arXiv (Cornell University) · Published 24 March 2026 · DOI 10.48550/arxiv.2603.22751

Open PDFView at the publisherCite this paper

In brief

“Observable Channels, Not Just Storage: Evaluating Privacy Leakage in LLM Agent Pipelines” is a 2026 preprint by Tao Huang and 3 co-authors, published in arXiv (Cornell University). OpenAlex counts 5 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
5
Most cited in
2026 (5)
References
22
Authors
4

Abstract

Privacy leakage in LLM agents is often studied through individual storage or execution components, such as memory modules, retrieval pipelines, or tool-mediated artifacts. However, these settings are typically analyzed in isolation, making it difficult to compare how private internal dependence becomes externally recoverable across heterogeneous agent pipelines. In this paper, we present CIPL (Channel Inversion for Privacy Leakage) as a unified channel-oriented measurement interface for evaluating privacy leakage in LLM agent pipelines. Rather than claiming a universally strongest attack recipe, CIPL provides a shared way to represent a target through its sensitive source, selection, assembly, execution, observation, and extraction stages, and to measure how internal exposure is transformed into attacker-recoverable leakage under a common protocol. Using memory-based, retrieval-mediated, and tool-mediated instantiations under this shared interface, we identify a distinct cross-target risk picture. Memory behaves as a near-saturated high-risk special case, while beyond-memory leakage exhibits a different regime: retrieval-mediated targets show frequent but often incomplete leakage, and tool-mediated targets are strongly shaped by the exposed observation surface and provider behavior. We further show that leakage is governed by channel conditions rather than by a universally dominant recipe: cleaned weak controls sharply suppress leakage, and semantic annotation reveals attacker-useful leakage beyond exact-match extraction. Together, these findings suggest that privacy risk in LLM agent pipelines is better understood through \emph{observable channels}, not just storage components. More broadly, our results motivate channel-oriented privacy evaluation as a necessary complement to component-local or exact-only analyses.

Cite this paper

Formatted from the catalogue record. Check names, capitals and page numbers against the paper before you submit.

APA 7

Huang, T., Hou, C., Wu, G., & Meng, J. (2026). Observable Channels, Not Just Storage: Evaluating Privacy Leakage in LLM Agent Pipelines. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2603.22751

MLA 9

Huang, Tao, et al. “Observable Channels, Not Just Storage: Evaluating Privacy Leakage in LLM Agent Pipelines.” arXiv (Cornell University), 2026. https://doi.org/10.48550/arxiv.2603.22751.

Chicago

Huang, Tao, Chen Hou, Guosen Wu, and Jiayang Meng. 2026. “Observable Channels, Not Just Storage: Evaluating Privacy Leakage in LLM Agent Pipelines.” arXiv (Cornell University). https://doi.org/10.48550/arxiv.2603.22751.

Harvard

Huang, T. et al. (2026) ‘Observable Channels, Not Just Storage: Evaluating Privacy Leakage in LLM Agent Pipelines’, arXiv (Cornell University). doi:10.48550/arxiv.2603.22751.

IEEE

T. Huang, C. Hou, G. Wu, and J. Meng, “Observable Channels, Not Just Storage: Evaluating Privacy Leakage in LLM Agent Pipelines,” arXiv (Cornell University), 2026, doi: 10.48550/arxiv.2603.22751.

BibTeX

@misc{huang2026observable,
  title     = {{Observable Channels, Not Just Storage: Evaluating Privacy Leakage in LLM Agent Pipelines}},
  author    = {Huang, Tao and Hou, Chen and Wu, Guosen and Meng, Jiayang},
  howpublished = {arXiv (Cornell University)},
  year      = {2026},
  doi       = {10.48550/arxiv.2603.22751},
  url       = {https://doi.org/10.48550/arxiv.2603.22751}
}

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

  • privacy leakage
  • LLM agents
  • sensitive information exposure
  • privacy risk assessment

Work with this paper

Most-cited papers in Computer Science

All highly cited Computer Science papers

Questions about this paper

How do I cite this paper in APA style?

Huang, T., Hou, C., Wu, G., & Meng, J. (2026). Observable Channels, Not Just Storage: Evaluating Privacy Leakage in LLM Agent Pipelines. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2603.22751

How many times has this paper been cited?

5 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?

Tao Huang, Chen Hou, Guosen Wu, Jiayang Meng.

What is the DOI of this paper?

10.48550/arxiv.2603.22751, permanently at https://doi.org/10.48550/arxiv.2603.22751. 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.