MCP-38: A Comprehensive Threat Taxonomy for Model Context Protocol Systems (v1.0)

Model Context Protocol (MCP) introduces a structurally distinct attack surface that existing threat frameworks, designed for traditional software systems or generic LLM deployments, do not adequately cover. This paper presents

Stateless Yet Not Forgetful: Implicit Memory as a Hidden Channel in LLMs

supplied. We challenge this assumption by introducing implicit memory-the ability of a model to carry state across otherwise independent interactions by encoding information in its own outputs and later

Backdoor Attacks on Prompt-Driven Video Segmentation Foundation Models

Prompt-driven Video Segmentation Foundation Models (VSFMs) such as SAM2

Bilevel Optimization for Covert Memory Tampering in Heterogeneous Multi-Agent Architectures (XAMT)

inherently heterogeneous, integrating conventional Multi-Agent Reinforcement Learning (MARL) with emerging Large Language Model (LLM) agent architectures utilizing Retrieval-Augmented Generation (RAG). A critical shared vulnerability is reliance on centralized

Securing the Model Context Protocol (MCP): Risks, Controls, and Governance

Model Context Protocol (MCP) replaces static, developer-controlled API integrations with more dynamic, user-driven agent systems, which also introduces new security risks. As MCP adoption grows across community servers

vLLM is an inference and serving engine for large language

StateFence: Task-State Protection Against Evidence-Chain Contamination in Blue-Team LLM Agents

Indirect prompt injection creates a distinctive risk for blue-team

AutoRedClaw: Automated Red-Teaming for Claw-Style Agents

Agent systems are shifting from task-scoped tool-use agents

VirusCascade: Hijacking Collaborative Reflection in LLM-Powered Recommender Agents

Advancing beyond traditional static scoring models, LLM-powered agentic recommender systems (LLM-ARS) instantiate users and items as autonomous agents, whose semantic states are dynamically refined through a recurrent process

Characterizing Network Centralization and Observability in the Remote MCP Ecosystem

The Model Context Protocol (MCP) has emerged as the dominant

Salience Induction against Multi-Hop RAG Agents: Threat and Defense

Multi-Hop question answering, agents chain facts across documents. Existing defenses focus on content poisoning, which injects false facts, and prompt injection, which embeds directives. We identify a third attack

Fed-FBD: Federated Functional Block Diversification for Isolation, Privacy, and Surgical Unlearning

Federated learning (FL) enables collaborative model training without sharing raw

Provable Robustness against Backdoor Attacks via the Primal-Dual Perspective on Differential Privacy

Randomized smoothing is a powerful tool for certifying robustness to adversarial perturbations, including poisoning attacks via randomized training and evasion attacks via randomized inference. Extending these guarantees to backdoor attacks

Not What You Asked For: Typographic Attacks in Household Robot Manipulation

Open-vocabulary embodied AI agents increasingly rely on vision-language

When AI reviews science: Can we trust the referee?

The volume of scientific submissions continues to climb, outpacing the

SlowBA: An efficiency backdoor attack towards VLM-based GUI agents

Modern vision-language-model (VLM) based graphical user interface (GUI

SoK: Trust-Authorization Mismatch in LLM Agent Interactions

Large Language Models (LLMs) are evolving into autonomous agents capable of executing complex workflows via standardized protocols (e.g., MCP). However, this paradigm shifts control from deterministic code to probabilistic inference

SSCL-BW: Sample-Specific Clean-Label Backdoor Watermarking for Dataset Ownership Verification

dataset owners. Existing backdoor-based dataset ownership verification methods suffer from inherent limitations: poison-label watermarks are easily detectable due to label inconsistencies, while clean-label watermarks face high technical

Terrarium: Revisiting the Blackboard for Multi-Agent Safety, Privacy, and Security Studies

multi-agent system (MAS) powered by large language models (LLMs) can automate tedious user tasks such as meeting scheduling that requires inter-agent collaboration. LLMs enable nuanced protocols that account

Injection, Attack and Erasure: Revocable Backdoor Attacks via Machine Unlearning

networks (DNNs) due to their stealth and durability. While recent research has explored leveraging model unlearning mechanisms to enhance backdoor concealment, existing attack strategies still leave persistent traces that