Powered by OpenAIRE graph
Found an issue? Give us feedback
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/ ZENODOarrow_drop_down
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/
ZENODO
Preprint . 2026
License: CC BY NC ND
Data sources: ZENODO
ZENODO
Preprint . 2026
License: CC BY NC ND
Data sources: Datacite
ZENODO
Preprint . 2026
License: CC BY NC ND
Data sources: Datacite
versions View all 2 versions
addClaim

RAG Shield: A Multi-Layer Defense System Against Poisoning Attacks in Retrieval-Augmented Generation

Authors: Petti, Fabio;

RAG Shield: A Multi-Layer Defense System Against Poisoning Attacks in Retrieval-Augmented Generation

Abstract

This whitepaper presents RAG Shield, a security-focused framework fordefending Retrieval-Augmented Generation (RAG) pipelines againstpoisoning and adversarial manipulation at the retrieval layer. The work introduces a multi-layer defense architecture combiningcryptographic document provenance validation, semantic anomaly detection,and secure, authority-weighted retrieval control. A realistic threatmodel is defined, focusing on poisoning of retrieval corpora rather thanprompt or model-level attacks. The system is evaluated against multipleattack scenarios under controlled conditions. RAG Shield is designed as a framework-agnostic security control layerthat operates independently of the underlying language model and vectordatabase, enabling deployment in enterprise and regulated environmentswithout modification of existing RAG architectures. This document is released as a technical preprint to establish prior artand support open discussion in the areas of AI security, adversarialmachine learning, and secure enterprise RAG deployment. Project website and system overview:https://sentinelrag.com Contact:info@sentinelrag.com

Keywords

poisoning attacks, retrieval-augmented generation, AI security, RAG security, adversarial machine learning

  • BIP!
    Impact byBIP!
    selected citations
    These citations are derived from selected sources.
    This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    0
    popularity
    This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
    Average
    influence
    This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    Average
    impulse
    This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
    Average
Powered by OpenAIRE graph
Found an issue? Give us feedback
selected citations
These citations are derived from selected sources.
This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Citations provided by BIP!
popularity
This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
BIP!Popularity provided by BIP!
influence
This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Influence provided by BIP!
impulse
This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
BIP!Impulse provided by BIP!
0
Average
Average
Average
Green