Results for “Attacks”

arXiv (research papers) Cathleen Qiao SI

An Exploration of Attacks on AI Agent Systems

As autonomous AI agents become more integrated into sensitive domains, it is crucial that single and multi-agent AI systems are secured. This paper investigates how prompt injection attacks can exploit both direct user inputs and inter?agent communication to leak sensitive information and trigger cross-agent infections. While our tests mostly highlight key vulnerabilities in single-agent systems, we also extend these attacks to multi-agent frameworks, investigating how…

Read at arXiv (research papers)

Journal articles (OpenAlex) Sungmin Cho · Pacific Affairs Geopolitics

Accidental Conflict in the Taiwan Strait and on the Korean Peninsula: A Comparative Analysis

The potential for simultaneous conflicts in the Taiwan Strait and on the Korean Peninsula has emerged as one of the most critical security challenges in East Asia. While analysts have primarily focused on how North Korea might respond to a Taiwan contingency, less attention has been paid to the reverse scenario: how China might react in the Taiwan Strait if North Korea attacks South Korea. For a more balanced analysis, this article compares the structural and political…

Read at Journal articles (OpenAlex)

Granma (Cuban state media) René Tamayo León Geopolitics

Cuba does not surrender, Cuba is not for sale, and Cuba does not forget

Miguel Díaz-Canel Bermúdez, First Secretary of the Central Committee of the Communist Party and President of the Republic, held a meeting with family members and colleagues of the victims of the Barbados bombing and other terrorist attacks against Cuba, as well as young members of the national fencing team, combatants from the FAR (Revolutionary Armed Forces) and Minint (Ministry of the Interior), scholars, and prominent figures from various sectors of society

Read at Granma (Cuban state media)

The Guardian Lebanon Tom Ambrose (now); Taz Ali and Martin Belam (earlier)

World leaders pay tribute to victims as Israel marks three years since 7 October attack by Hamas – as it happened

1,200 died in attack that led to launch of devastating war in Gaza ‘All I want to do is cry’: trauma of 7 October survivors three years after the attacks The Israeli military said it launched several security operations across the West Bank overnight, arresting dozens of people suspected of being involved in terrorism, including expressing support for the 7 October attack, the Israeli military said. A young man was shot and wounded by Israeli soldiers during one of the…

Read at The Guardian Lebanon

Committee to Protect Journalists CPJ Staff

Spate of attacks on DRC press include journalist threatened over Fally Ipupa controversy

Investigative journalist Gauthier Sey Tshiang was evacuated to France on Monday for medical treatment after being assaulted and left unconscious in a street — one of a series of attacks on the press in the Democratic Republic of the Congo in recent weeks. Another journalist has been missing for a month and two are behind...

Read at Committee to Protect Journalists

arXiv (research papers) Yupu Wang SI

Package Hallucination Attacks on Coding Agents through Prompt Injection in Rule Files

Modern agentic coding frameworks increasingly rely on community-shared rule files (e.g., AGENTS.md or .cursorrules) to guide autonomous code generation, yet the security risks of this pipeline remain underexplored. To bridge this gap, we introduce the package hallucination attack, where an attacker injects malicious prompts into benign rule files to induce coding agents to replace legitimate dependencies with attacker-controlled packages. To obtain effective malicious…

Read at arXiv (research papers)

Journal articles (OpenAlex) M. E. Caplan · Global Security Health Science and Policy Geopolitics

A nuclear terrorism scenario: improvised nuclear weapons with 60%-enriched uranium

In this work we show that approximately 60–70 kg of 60%-enriched uranium can be used to build a crude nuclear weapon with a kiloton yield. While too large to fit on a missile, such a weapon could be delivered by a shipping container. This analysis is motivated by the June 2025 Israeli and US attacks on Iran, especially the bombings of the nuclear facilities at Natanz, Fordow and Isfahan. The Iranian stockpile of approximately 440 kg of 60%-enriched uranium is, at the time of…

Read at Journal articles (OpenAlex)

Journal articles (OpenAlex) Hari Godara · World Water Policy Geopolitics

Performative Sovereignty Versus Performative Nationalism: Hydraulic Securitization and Abeyance of the Indus Waters Treaty

ABSTRACT Indus Waters Treaty (IWT, 1960) survived four major wars, decades of cross‐border militancy, and 56 years of parliamentary skepticism in India. However, it could not survive the performative securitization sequence initiated by three terror attacks, that is, Uri (2016), Pulwama (2019), and Pahalgam (2025), leading to IWT's abeyance in 2025. Conventional securitization frameworks become incompatible here as dam construction and treaty management operate as two…

Read at Journal articles (OpenAlex)

arXiv (research papers) Niveen O. Jaffal SI

RAG-PIBench: A Leakage-Aware Benchmark for Prompt-Injection Detection in Trustworthy RAG Systems

Retrieval-Augmented Generation (RAG) systems are vulnerable to prompt-injection attacks embedded in retrieved content. We introduce RAG-PIBench, a benchmark for RAG-style prompt-injection detection containing 4,876 contextual examples across frozen train, validation, and protected-test splits. Using a leakage-aware construction pipeline and strict evaluation protocol, we compare keyword-based, semantic-reference, TF-IDF, and transformer-based detectors. DistilBERT achieves…

Read at arXiv (research papers)

The Guardian NATO Jessica Elgot and Dan Sabbagh

Conservatives propose £10bn ‘Britannia Shield’ inspired by Israel’s Iron Dome

Party to unveil plans for one of UK’s ‘most significant military programmes since the cold war’ The Conservatives will announce plans to build a £10bn British “Iron Dome” similar to Israel’s air and missile defence system, which the party will christen “Britannia Shield” in an announcement at conference on Tuesday. The shadow defence secretary, James Cartlidge, will say Britain needs to be ready to defend itself from potential aerial attacks and that new investment was…

Read at The Guardian NATO

KabulNow Besmellah Zahidi

Hazaras Face Violence, Displacement and Discrimination Under Taliban Rule: Report

Afghanistan’s Hazara community continues to face deadly attacks, forced displacement, discrimination and loss of land under Taliban rule, according to a new report by U.S.-based Genocide Watch that calls for accountability and protection.

Read at KabulNow

arXiv (research papers) Aniruddh Pramod SI

Towards a Unified Misuse Monitoring Benchmark

LLM agents increasingly act in multi-actor environments, exposing them to misuse from multiple sources: decomposition attacks, where a harmful request is split into innocuous sub-requests, and prompt injection attacks, where a compromised tool delivers a malicious instruction. Existing evaluations treat these threats separately and ask whether a trajectory is harmful, rather than when it becomes harmful. We propose monitoring the agent's responses, where its actions are…

Read at arXiv (research papers)

arXiv (research papers) James Peters-Gill SI

Can CaMeLs Talk? Securing Multi-Agent Systems Against Indirect Prompt Injection Attacks

Indirect prompt injection attacks - malicious instructions embedded in content processed by large language models - remain a major obstacle to safely deploying tool-using agents. CaMeL [Debenedetti et al., 2025] mitigates this threat for an individual agent by separating trusted control flow from untrusted data and enforcing capability-based security policies at runtime. In this work, we investigate whether CaMeL's security guarantees compose in hierarchical multi-agent…

Read at arXiv (research papers)

Journal articles (OpenAlex) Faris N. Alsulami · Applied Sciences Gadgets

ChipGuard-AI: A Hardware-Assisted Unified Framework for Adversarial and Side-Channel Resilience in Consumer Electronics

Consumer electronics powered by on-device artificial intelligence face a structurally coupled dual threat: adversarial attacks that manipulate model outputs, and side-channel attacks that exploit physical implementation characteristics to extract cryptographic keys and neural network weights. Prior work addresses these threats independently, leaving exploitable gaps when both attack vectors are applied in a coordinated manner. This paper presents ChipGuard-AI, a…

Read at Journal articles (OpenAlex)

arXiv (research papers) Jingkai Liu SI

Blocking at the Boundary: Auditing Long-Horizon Agents against Staged Prompt Injection

Long-horizon agents consume external content, invoke tools, and modify persistent state. Indirect prompt injection can exploit task-specific context, propagate across causally connected stages, and alter a consequential action while the workflow continues; we term this staged prompt injection. We build an automated, feedback-guided attack generation pipeline and apply it to Claude Code and Codex in their native runtimes. The confirmed attacks span eight workflow scenarios…

Read at arXiv (research papers)

arXiv (research papers) Zixuan Wang SI

Self-Reflection Fine-Tuning: Enhancing Agent Security against Prompt Injection Attacks from Failure Experience

Large language model (LLM) agents are increasingly deployed in tool-augmented environments, but their reliance on external inputs makes them highly vulnerable to prompt injection attacks that can hijack task objectives. Existing safety alignment methods rely on static expert trajectories or preference optimization, limiting their ability to generalize to adaptive attack patterns. In this work, we propose Self-Reflection Fine-Tuning (SRFT), a training framework that enables…

Read at arXiv (research papers)

arXiv (research papers) Debeshee Das SI

Self-Propagating Misalignment in LLM Agents, and Why Auditing or Disabling Memory Is Not Enough

Memory poisoning attacks on LLM agents typically assume an external adversary who plants content in the agent's persistent memory to steer its behavior. We instead study, with no adversary involved, whether a misaligned agent can write a goal it cannot yet act on to persistent memory, so that a future aligned agent carries it out when the opportunity arises. We investigate this threat, which we refer to as self-propagation of misalignment, across 20 different scenarios…

Read at arXiv (research papers)

arXiv (research papers) Giulio Zingrillo SI

Securing Computer-Use Agents Against Branch Steering Attacks

Modern Computer Use Agents (CUAs) directly interact with graphical user interfaces and execute third-party web tools, exposing them to indirect prompt injection across every rendered page and tool response. While the Dual-LLM pattern is the primary system-level architecture offering formal security guarantees - using an isolated Planner LLM (P-LLM) to fix execution paths before processing untrusted inputs via a Quarantined LLM (Q-LLM) - these guarantees break down in…

Read at arXiv (research papers)

arXiv (research papers) Narek Maloyan SI

Evaluating and Improving the Robustness of Large Language Models to Input Sequence Variations

Large language models (LLMs) in production systems face prompt injections, trojans (backdoors), and manipulation of automatic quality metrics. This thesis develops models, methods, and algorithms for evaluating and improving LLM robustness to adversarial input sequence variations. We propose R_stab(f), a generative robustness metric based on the Jensen-Shannon divergence between per-step output distributions under small input perturbations. For localized attacks we prove…

Read at arXiv (research papers)

Journal articles (OpenAlex) Jiwon Yun · Pacific Focus Geopolitics

An Analysis of the Aspects of Human Security Threat and Implications of International Cooperation in Ukraine After the Outbreak of the Russo‐Ukraine War

ABSTRACT The Russo‐Ukrainian War, which broke out on February 24, 2022, has continued despite diplomatic efforts by the US and Western nations to bring about a ceasefire. Russia has strongly demanded de facto recognition of the Ukrainian territories it occupies as a precondition for ending the war. Meanwhile, Russia's indiscriminate attacks have led to a massive influx of refugees and internally displaced persons in Ukraine, inhumane treatment of prisoners of war, and the…

Read at Journal articles (OpenAlex)

Journal articles (OpenAlex) Nagashree B A, M. Vinayakamurthy · MethodsX Gadgets

Adaptive edge-cloud isolation forest-autoencoder pipeline in real-time with delta policy feedback for zero-day IoT smartwatch intrusion detection

Smartwatches and other wearable IoT devices constantly stream physiological and behavioral data, making them vulnerable to zero-day attacks that traditional signature-based intrusion detection systems (IDS) are unable to identify and that current hybrid edge-cloud IDS leave momentarily unprotected while a cloud verdict is awaited. This approach outlines a fully reproducible three-tier detection pipeline: publisher, edge, and cloud. An edge-resident Isolation Forest scores…

Read at Journal articles (OpenAlex)

arXiv (research papers) Birk Torpmann-Hagen SI

Memetic Trojans: Social Contagions as Carriers of Adversarial Payloads in Agent Networks

Autonomous large language model (LLM) agents increasingly interact in network environments where adversarial content can propagate between agents. Known attacks include agent worms, which spread through self-replicating prompt injections or configuration compromises. We introduce \emph{memetic trojans}, a distinct class of network-mediated attack that exploits agents' tendencies to retransmit and amplify content. Unlike agent worms, whose propagation is adversarially…

Read at arXiv (research papers)

arXiv (research papers) Zonghao Ying SI

pikit: A Composable Toolkit for Indirect Prompt Injection Research and Evaluation

Indirect prompt injection embeds malicious instructions within external content retrieved by LLM-based agents, altering target behavior without user authorization. We introduce pikit, a research toolkit designed to systematically evaluate these threats across three core dimensions: attacks (13 methods), channels (16 carriers across text and file modes), and defenses (9 prevention strategies and 3 offline detection baselines). Built on a decorator-based registry, pikit…

Read at arXiv (research papers)

arXiv (research papers) Minh Nhat Le SI

Self-Evolving Defense: Continual Security Policy Learning for LLM Agents

Large language models (LLMs) increasingly power agents that access sensitive information, use external tools, and modify software repositories. Although these capabilities offer substantial benefits, they also create security risks such as jailbreaks, prompt injection, and vulnerable code generation. Existing defenses often require retraining, fail to adapt to evolving attacks, or address only a single threat pattern. To address these limitations, we propose Self-Evolving…

Read at arXiv (research papers)

arXiv (research papers) Jie Zhang SI

Render Before Reading: Visual Rendering as a Prompt Injection Defense

Large language models are vulnerable to prompt injection attacks, where third-party adversarial content can hijack the model's behavior. In this paper, we study the role played by the adversarial data's input modality, and identify a systematic asymmetry: multimodal LLMs are more likely to follow adversarial instruction when they appear as text than when the same instruction is delivered through a non-textual channel (e.g., as an image). We hypothesize that this modality gap…

Read at arXiv (research papers)

Global Voices Maksuda Akter

Beyond protest: Mob violence and the struggle for institutional response in Bangladesh

Mob violence is rising in Bangladesh as unchecked impunity, weak law enforcement, and social media instigations drive lethal crowd attacks, challenging institutional responses, justice delivery, and overall state accountability.

Read at Global Voices

KabulNow Nazanin Mohseni Geopolitics

At UN, Pakistan PM Says Afghan Territory Is Being Used for Attacks Against Pakistan

KABUL, AFGHANISTAN - Pakistan’s Prime Minister Shehbaz Sharif has raised the situation in Afghanistan under Taliban control during his address to the United Nations General Assembly, accusing militant groups of using Afghan territory to carry out attacks in Pakistan while also calling for the protection of Afghans’ fundamental rights, particularly those of women and girls. Sharif said the Tehrik-e-Taliban Pakistan (TTP) and the Balochistan Liberation Army (BLA) were using…

Read at KabulNow

The Guardian NATO Guardian staff and agencies

Ukraine war briefing: Russian attacks force closure of Ukraine’s largest steelmaker

Closure another blow to Ukraine’s steel sector, as Zelenskyy warns Moscow is expanding its targets to include ‘daily life’. What we know on day 1,676

Read at The Guardian NATO

arXiv (research papers) Jianyi Hu SI

JevAdvBench: A Benchmark and Black-Box Attacks for Reinforcement Learning for Calibrated Decisions Models

Typed decision models (TDMs), such as Jev, answer a question about an input with a probability, an option, or an ordinal level. Web services execute these decisions automatically above a confidence threshold. An attacker who writes part of the input can change a decision that the service then executes. Current adversarial benchmarks cannot measure this risk, because they judge generated text. A TDM writes no text, and its wrong decisions have a valid format. We present…

Read at arXiv (research papers)

arXiv (research papers) Hanzhang Ma SI

MetaPermit: Scalable and Auditable Access Control for AI Agents via LLM-Inferred Meta-Attributes

The rise of autonomous AI agents equipped with tools has introduced significant security risks, ranging from unintended tool misuse to adversarial manipulation through Indirect Prompt Injection (IPI) attacks. In practice, deployed agent systems such as OpenAI Codex and Claude Code protect tool invocations through a combination of coarse-grained permission rules and LLM-based judgments about individual proposed actions. Both components, however, have important limitations…

Read at arXiv (research papers)

Older stories