Sentinel-Rag: A Controlled Comparison of a Scratch-Trained Classifier and a Local LLM Judge for Defending Enterprise Retrieval-Augmented Generation against Indirect Prompt Injection

Enterprise assistants built on retrieval-augmented generation (RAG) read company chats, wikis and policies, so anyone who can post into those sources can plant instructions that the assistant may follow. We present Sentinel-RAG, a controlled testbed in which three assistants share the same corpus, retriever and answer model and differ only in the security stage between retrieval and generation: none, role-based access control plus a 1.6-million-parameter Transformer classifier trained from scratch, or the same access control plus a local 2-billion-parameter LLM judge with schema-constrained output. The classifier reaches 89.8% accuracy and 0.895 macro F1 on held-out attack patterns. On a 25-question suite under two corpus conditions, the unguarded assistant gave 4 and 10 unsafe answers, while both guards gave none. The classifier added 8–9 ms per query but removed 9–16% of benign chunks; the judge removed none but added 13–14 s per query.

Authors

Institutions

Publication Details

Journal
International Journal of Advanced Research in Science Communication and Technology
Published
2026-11-17
DOI
https://doi.org/10.48175/ijarsct-40823
Primary Topic
Adversarial Robustness in Machine Learning
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Sentinel-Rag: A Controlled Comparison of a Scratch-Trained Classifier and a Local LLM Judge for Defending Enterprise Retrieval-Augmented Generation against Indirect Prompt Injection

Soham Arte, Sushil Sonawane, Vijaya Patil, Aadi Surve Waaridh Borpujari
International Journal of Advanced Research in Science Communication and Technology
Adversarial Robustness in Machine Learning
article

Sentinel-Rag: A Controlled Comparison of a Scratch-Trained Classifier and a Local LLM Judge for Defending Enterprise Retrieval-Augmented Generation against Indirect Prompt Injection

Soham Arte, Sushil Sonawane, Vijaya Patil, Aadi Surve Waaridh Borpujari
article en

Abstract

Enterprise assistants built on retrieval-augmented generation (RAG) read company chats, wikis and policies, so anyone who can post into those sources can plant instructions that the assistant may follow. We present Sentinel-RAG, a controlled testbed in which three assistants share the same corpus, retriever and answer model and differ only in the security stage between retrieval and generation: none, role-based access control plus a 1.6-million-parameter Transformer classifier trained from scratch, or the same access control plus a local 2-billion-parameter LLM judge with schema-constrained output. The classifier reaches 89.8% accuracy and 0.895 macro F1 on held-out attack patterns. On a 25-question suite under two corpus conditions, the unguarded assistant gave 4 and 10 unsafe answers, while both guards gave none. The classifier added 8–9 ms per query but removed 9–16% of benign chunks; the judge removed none but added 13–14 s per query.

International Journal of Advanced Research in Science Communication and Technology
MIT Art, Design and Technology University (IN)
Adversarial Robustness in Machine Learning
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon