Analyzing Defensive Misdirection Against Model-Guided Automated Attacks on Agentic AI Systems

Vivek Namsani; Reza Soosahabi

doi:10.5281/zenodo.22869768

RResearcharAcademic radar for researchers

Feed Venues Groups Network Inbox Profile

Your academic radar

One personalized feed of papers, calls for papers, collaborators and datasets — plus research groups and paid-in-points study recruitment.

Create account Log in

Top venues

Browse all

Loading…

Analyzing Defensive Misdirection Against Model-Guided Automated Attacks on Agentic AI Systems · Researchar

Back to feed

Paperopenalex

ZCE

Analyzing Defensive Misdirection Against Model-Guided Automated Attacks on Agentic AI Systems

Vivek Namsani, Reza Soosahabi · Zenodo (CERN European Organization for Nuclear Research) 2026 · 2026

DOI: 10.5281/zenodo.22869768

0citations

Adversarial Robustness in Machine LearningExplainable Artificial Intelligence (XAI)Ethics and Social Impacts of AIArtifact (error)Computer science

Citations by source

Synced 2026-09-23T02:31:08.701886Z

0OpenAlex · primary0Semantic Scholar0Crossref

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

Abstract

Research artifact for the paper "Analyzing Defensive Misdirection AgainstModel-Guided Automated Attacks on Agentic AI Systems" (ACSAC 2026). This artifact contains: (1) the LLMaaD defense framework implementing theCMPE (Confuse, Misdirect, Preserve, Evade) algorithm; (2) integration andevaluation scripts for AutoDAN-Turbo, AutoDAN-Reasoning, GPTFuzz, and PAIRattack frameworks; (3) pre-computed result CSVs for all experiments.

Analyzing Defensive Misdirection Against Model-Guided Automated Attacks on Agentic AI Systems

Citations by source

Abstract

Discussion (0)