Cover Image

The Defense Trilemma: Why Prompt Injection Defense Wrappers Fail?

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Bhatt, Manish, Munshi, Sarthak, Narajala, Vineeth Sai, Habler, Idan, Al-Kahfah, Ammar, Huang, Ken, Webb, Joel, Gatto, Blake, Hoque, Md Tamjidul
Format:	Preprint
Published:	2026
Subjects:	Cryptography and Security Artificial Intelligence
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Manifold of Failure: Behavioral Attraction Basins in Language Models
by: Munshi, Sarthak, et al.
Published: (2026)

ETDI: Mitigating Tool Squatting and Rug Pull Attacks in Model Context Protocol (MCP) by using OAuth-Enhanced Tool Definitions and Policy-Based Access Control
by: Bhatt, Manish, et al.
Published: (2025)

Enterprise-Grade Security for the Model Context Protocol (MCP): Frameworks and Mitigation Strategies
by: Narajala, Vineeth Sai, et al.
Published: (2025)

Securing GenAI Multi-Agent Systems Against Tool Squatting: A Zero Trust Registry-Based Approach
by: Narajala, Vineeth Sai, et al.
Published: (2025)

Large Empirical Case Study: Go-Explore adapted for AI Red Team Testing
by: Bhatt, Manish, et al.
Published: (2025)

COALESCE: Economic and Security Dynamics of Skill-Based Task Outsourcing Among Team of Autonomous LLM Agents
by: Bhatt, Manish, et al.
Published: (2025)

Building A Secure Agentic AI Application Leveraging A2A Protocol
by: Habler, Idan, et al.
Published: (2025)

Agent Capability Negotiation and Binding Protocol (ACNBP)
by: Huang, Ken, et al.
Published: (2025)

Agent Name Service (ANS): A Universal Directory for Secure AI Agent Discovery and Interoperability
by: Huang, Ken, et al.
Published: (2025)

MAIF: Enforcing AI Trust and Provenance with an Artifact-Centric Agentic Paradigm
by: Narajala, Vineeth Sai, et al.
Published: (2025)

Adversarial Hubness Detector: Detecting Hubness Poisoning in Retrieval-Augmented Generation Systems
by: Habler, Idan, et al.
Published: (2026)

A2AS: Agentic AI Runtime Security and Self-Defense
by: Neelou, Eugene, et al.
Published: (2025)

A Novel Zero-Trust Identity Framework for Agentic AI: Decentralized Authentication and Fine-Grained Access Control
by: Huang, Ken, et al.
Published: (2025)

Securing Agentic AI: A Comprehensive Threat Model and Mitigation Framework for Generative AI Agents
by: Narajala, Vineeth Sai, et al.
Published: (2025)

Why Self-Awareness Fails: Coherence, Defense, and the Structure of Repair
by: Jovanovic, Vladisav
Published: (2026)

Formalizing and Benchmarking Prompt Injection Attacks and Defenses
by: Liu, Yupei, et al.
Published: (2023)

PromptArmor: Simple yet Effective Prompt Injection Defenses
by: Shi, Tianneng, et al.
Published: (2025)

ARES-LSHADE: Autoresearch-Enhanced LSHADE with Memetic Polish for the GNBG Benchmark
by: Naeem, Abdullah, et al.
Published: (2026)

Jatmo: Prompt Injection Defense by Task-Specific Finetuning
by: Piet, Julien, et al.
Published: (2023)

Evaluation of Prompt Injection Defenses in Large Language Models
by: Deep, Priyal, et al.
Published: (2026)

Backdoor-Powered Prompt Injection Attacks Nullify Defense Methods
by: Chen, Yulin, et al.
Published: (2025)

Multimodal Prompt Injection Attacks: Risks and Defenses for Modern LLMs
by: Yeo, Andrew, et al.
Published: (2025)

Defending Against Prompt Injection With a Few DefensiveTokens
by: Chen, Sizhe, et al.
Published: (2025)

Defense Against Prompt Injection Attack by Leveraging Attack Techniques
by: Chen, Yulin, et al.
Published: (2024)

A Critical Evaluation of Defenses against Prompt Injection Attacks
by: Jia, Yuqi, et al.
Published: (2025)

Defense against Prompt Injection Attacks via Mixture of Encodings
by: Zhang, Ruiyi, et al.
Published: (2025)

Beyond the Benchmark: Innovative Defenses Against Prompt Injection Attacks
by: Shaheer, Safwan, et al.
Published: (2025)

ESLD (External Surrogate Latent Defense): A Latent-Space Architecture for Faster, Stronger Prompt-Injection Defense
by: Narendra, Yash
Published: (2026)

Defense Against Indirect Prompt Injection via Tool Result Parsing
by: Yu, Qiang, et al.
Published: (2026)

PISmith: Reinforcement Learning-based Red Teaming for Prompt Injection Defenses
by: Yin, Chenlong, et al.
Published: (2026)

WARD: Adversarially Robust Defense of Web Agents Against Prompt Injections
by: Cao, Tri, et al.
Published: (2026)

One Step to the Side: Why Defenses Against Malicious Finetuning Fail Under Adaptive Adversaries
by: Zloczower, Itay, et al.
Published: (2026)

MELON: Provable Defense Against Indirect Prompt Injection Attacks in AI Agents
by: Zhu, Kaijie, et al.
Published: (2025)

A Multi-Agent LLM Defense Pipeline Against Prompt Injection Attacks
by: Hossain, S M Asif, et al.
Published: (2025)

ICON: Indirect Prompt Injection Defense for Agents based on Inference-Time Correction
by: Wang, Che, et al.
Published: (2026)

Checkpoint-GCG: Auditing and Attacking Fine-Tuning-Based Prompt Injection Defenses
by: Yang, Xiaoxue, et al.
Published: (2025)

Logic layer Prompt Control Injection (LPCI): A Novel Security Vulnerability Class in Agentic Systems
by: Atta, Hammad, et al.
Published: (2025)

AegisAgent: An Autonomous Defense Agent Against Prompt Injection Attacks in LLM-HARs
by: Wang, Yihan, et al.
Published: (2025)

FATH: Authentication-based Test-time Defense against Indirect Prompt Injection Attacks
by: Wang, Jiongxiao, et al.
Published: (2024)

LeakSealer: A Semisupervised Defense for LLMs Against Prompt Injection and Leakage Attacks
by: Panebianco, Francesco, et al.
Published: (2025)