Copertina

FlowSteer: Prompt-Only Workflow Steering Exposes Planning-Time Vulnerabilities in Multi-Agent LLM Systems

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autori principali:	Li, Fanxiao, Wu, Jiaying, Fu, Tingchao, Jaques, Natasha, Zhou, Wei, Kan, Min-Yen
Natura:	Preprint
Pubblicazione:	2026
Soggetti:	Cryptography and Security
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

From Storage to Steering: Memory Control Flow Attacks on LLM Agents
di: Xu, Zhenlin, et al.
Pubblicazione: (2026)

Cross-Modal Content Optimization for Steering Web Agent Preferences
di: Jiang, Tanqiu, et al.
Pubblicazione: (2025)

ChainFuzzer: Greybox Fuzzing for Workflow-Level Multi-Tool Vulnerabilities in LLM Agents
di: Wu, Jiangrong, et al.
Pubblicazione: (2026)

Harmless Yet Harmful: Neutral Prompting Attacks for Stealthy Hallucination Steering in Agent Skills
di: Hsu, Chia-Yi, et al.
Pubblicazione: (2026)

SafeSteer: Adaptive Subspace Steering for Efficient Jailbreak Defense in Vision-Language Models
di: Zeng, Xiyu, et al.
Pubblicazione: (2025)

Trust Me, Import This: Dependency Steering Attacks via Malicious Agent Skills
di: Liu, Yiyong, et al.
Pubblicazione: (2026)

Interpretable LLM Guardrails via Sparse Representation Steering
di: He, Zeqing, et al.
Pubblicazione: (2025)

Adaptive Probe-based Steering for Robust LLM Jailbreaking
di: Chen, Junxi, et al.
Pubblicazione: (2026)

Exposing Vulnerabilities in Counterfeit Prevention Systems Utilizing Physically Unclonable Surface Features
di: Nakra, Anirudh, et al.
Pubblicazione: (2025)

The Vulnerability of LLM Rankers to Prompt Injection Attacks
di: Yin, Yu, et al.
Pubblicazione: (2026)

Supply Chain Insecurity: Exposing Vulnerabilities in iOS Dependency Management Systems
di: Schmidt, David, et al.
Pubblicazione: (2026)

GradingAttack: Exposing Security Vulnerabilities in LLM Based Educational Grading Agents
di: Li, Xueyi, et al.
Pubblicazione: (2026)

PatchIsland: Orchestration of LLM Agents for Continuous Vulnerability Repair
di: Kim, Wonyoung, et al.
Pubblicazione: (2026)

Atomicity for Agents: Exposing, Exploiting, and Mitigating TOCTOU Vulnerabilities in Browser-Use Agents
di: Jiang, Linxi, et al.
Pubblicazione: (2026)

ARGUS: Defending Against Multimodal Indirect Prompt Injection via Steering Instruction-Following Behavior
di: Lu, Weikai, et al.
Pubblicazione: (2025)

The Trust Paradox in LLM-Based Multi-Agent Systems: When Collaboration Becomes a Security Vulnerability
di: Xu, Zijie, et al.
Pubblicazione: (2025)

Multi-Faceted Attack: Exposing Cross-Model Vulnerabilities in Defense-Equipped Vision-Language Models
di: Yang, Yijun, et al.
Pubblicazione: (2025)

AdaSteer: Your Aligned LLM is Inherently an Adaptive Jailbreak Defender
di: Zhao, Weixiang, et al.
Pubblicazione: (2025)

Steering Externalities: Benign Activation Steering Unintentionally Increases Jailbreak Risk for Large Language Models
di: Xiong, Chen, et al.
Pubblicazione: (2026)

The System Prompt Is the Attack Surface: How LLM Agent Configuration Shapes Security and Creates Exploitable Vulnerabilities
di: Litvak, Ron
Pubblicazione: (2026)

From Prompt Injections to Protocol Exploits: Threats in LLM-Powered AI Agents Workflows
di: Ferrag, Mohamed Amine, et al.
Pubblicazione: (2025)

ASPI: Seeking Ambiguity Clarification Amplifies Prompt Injection Vulnerability in LLM Agents
di: Sehwag, Udari Madhushani, et al.
Pubblicazione: (2026)

MALF: A Multi-Agent LLM Framework for Intelligent Fuzzing of Industrial Control Protocols
di: Ning, Bowei, et al.
Pubblicazione: (2025)

Vulnerability Mitigation System (VMS): LLM Agent and Evaluation Framework for Autonomous Penetration Testing
di: Abdulzada, Farzana
Pubblicazione: (2025)

Analysing the Safety Pitfalls of Steering Vectors
di: Li, Yuxiao, et al.
Pubblicazione: (2026)

Prompt Infection: LLM-to-LLM Prompt Injection within Multi-Agent Systems
di: Lee, Donghyun, et al.
Pubblicazione: (2024)

AlphaSteer: Learning Refusal Steering with Principled Null-Space Constraint
di: Sheng, Leheng, et al.
Pubblicazione: (2025)

MASCing: Configurable Mixture-of-Experts Behavior via Activation Steering Masks
di: Lintelo, Jona te, et al.
Pubblicazione: (2026)

Prompt Control-Flow Integrity: A Priority-Aware Runtime Defense Against Prompt Injection in LLM Systems
di: Alam, Md Takrim Ul, et al.
Pubblicazione: (2026)

Mind the Gap: Time-of-Check to Time-of-Use Vulnerabilities in LLM-Enabled Agents
di: Lilienthal, Derek, et al.
Pubblicazione: (2025)

Synthesizing Multi-Agent Harnesses for Vulnerability Discovery
di: Liu, Hanzhi, et al.
Pubblicazione: (2026)

The Blind Spot of Agent Safety: How Benign User Instructions Expose Critical Vulnerabilities in Computer-Use Agents
di: Ding, Xuwei, et al.
Pubblicazione: (2026)

Steering in the Shadows: Causal Amplification for Activation Space Attacks in Large Language Models
di: Xu, Zhiyuan, et al.
Pubblicazione: (2025)

PII Jailbreaking in LLMs via Activation Steering Reveals Personal Information Leakage
di: Nakka, Krishna Kanth, et al.
Pubblicazione: (2025)

Exposing LLM User Privacy via Traffic Fingerprint Analysis: A Study of Privacy Risks in LLM Agent Interactions
di: Zhang, Yixiang, et al.
Pubblicazione: (2025)

A Duty to Forget, a Right to be Assured? Exposing Vulnerabilities in Machine Unlearning Services
di: Hu, Hongsheng, et al.
Pubblicazione: (2023)

Exposing Vulnerabilities in RL: A Novel Stealthy Backdoor Attack through Reward Poisoning
di: Zhang, Bokang, et al.
Pubblicazione: (2025)

Transient Turn Injection: Exposing Stateless Multi-Turn Vulnerabilities in Large Language Models
di: Rayhan, Naheed, et al.
Pubblicazione: (2026)

Eyes-on-Me: Scalable RAG Poisoning through Transferable Attention-Steering Attractors
di: Chen, Yen-Shan, et al.
Pubblicazione: (2025)

Exposing and Defending Membership Leakage in Vulnerability Prediction Models
di: Liao, Yihan, et al.
Pubblicazione: (2025)