The paper investigates whether hidden representations in latent-based multi‑agent systems can carry attack information that remains effective during normal operation. A latent attack framework is introduced, reactivating attack effects through latent interventions without using adversarial text. Experiments show that these latent attacks can significantly degrade task performance, especially when targeting inter‑agent KV‑cache handoffs, and that the degradation cannot be explained by simple perturbations or invalid generation.
arXiv:2608. 01085v1 Announce Type: cross Abstract: LLM-based multi-agent systems (MAS) extend LLM capabilities through iterative communication and shared contexts.
arXiv:2605. 01133v3 Announce Type: replace-cross Abstract: Large language model (LLM)-powered multi-agent systems (MAS) enable agents to communicate and share information, achieving strong performance on complex tasks.
arXiv:2606. 12737v1 Announce Type: cross Abstract: Large Language Models (LLMs) are rapidly evolving into agentic systems that interact with external tools and environments, introducing new security risks such as indirect prompt injection attacks through untrusted external sources.
The paper investigates latent communication in multi‑agent systems, where agents exchange information in internal representation space via lightweight trainable links. It demonstrates that even benign training of these links can increase harmful compliance, and that attackers can exploit or poison the links to amplify this effect. A reinforcement‑learning attack further boosts harmful compliance while maintaining benign task performance, but adjusting rewards toward safer behavior can repair compromised links without updating the agents.
The paper introduces AgentLeak, a black‑box attack that clones the task‑solving capabilities of a strong LLM agent onto a weaker one by exploiting differences between successful and failed executions. Unlike prior skill‑stealing methods that only recover explicit skill artifacts, AgentLeak identifies and incorporates missing procedural behaviors, boosting task pass rates by over 40% and closing more than 80% of the capability gap across 20 scenarios. The study demonstrates that observable execution behavior can leak proprietary procedural knowledge, posing a confidentiality risk for LLM agents.
arXiv:2607. 14611v1 Announce Type: cross Abstract: A growing class of agentic systems maintain persistent state across sessions through memory files, behavioral preferences, and knowledge bases.
arXiv:2608.30207v1 Announce Type: cross Abstract: Computer use agents (CUAs) are vision-language models that perceive a screen and act on a real operating system through mouse, keyboard, and terminal...
arXiv:2606. 13994v1 Announce Type: cross Abstract: LLM-based Agents are becoming increasingly capable and widely deployed, creating growing incentives for adversarial misuse in the real-world.
arXiv:2608. 02657v1 Announce Type: cross Abstract: Agentic LLMs are vulnerable to indirect prompt injection (IPI) attacks, e.
arXiv:2607. 06595v1 Announce Type: cross Abstract: Personal AI agents powered by large language models can reason and act using available tools to access emails, manage calendars, and push code to remote repositories, all with minimal oversight.
DUMA-Bench is a new benchmark that evaluates the security of large language model agents in dual‑control settings, where both the agent and the user can modify the shared environment. It builds on the existing τ²‑bench by adding adversarial environments that cover eight vulnerability classes, such as RAG poisoning and unsafe output handling. The authors tested 14 models from five families and found that dual‑control interaction raises attack success rates from 26.9% to 41.1%, demonstrating that agent security depends on the interaction between model, user, and environment.
AgentXploit is a two‑role auditing system that separates repository‑level attack‑path discovery from runtime exploitation for AI agents. The Analyzer Agent traces attacker‑controlled inputs to sensitive operations and records candidate attack paths, while the Exploiter Agent turns these paths into concrete attacks and refines them using runtime feedback. The system is evaluated on AgentXploit‑Bench, a benchmark of 72 reproducible vulnerabilities across 12 open‑source AI‑agent systems, achieving 59.3% end‑to‑end success compared to 38.4% for Codex, and 79.2% attack success on AgentDojo versus 52.7% for AgentVigil.