Results for “security”

arXiv (research papers) Chenlong Yin SI

Climbing the Hill: Prompt Injection Red-Teaming Against Frontier Models with Curriculum Reinforcement Learning

Prompt injection is a leading security risk for LLMs and LLM-based applications such as agents. State-of-the-art red-teaming methods for prompt injection leverage reinforcement learning (RL) to train an attacker LLM to generate effective injected prompts. However, when targeting frontier LLMs such as GPT-6-Luna, a major challenge is the cold-start problem: every attack attempt by the attacker LLM fails and thus receives zero reward, providing no signal for learning. In this…

Read at arXiv (research papers)

arXiv (research papers) Yixuan Liu SI

Evaluating System One Models for Agent Security Decisions: Reliability, Calibration, and Selective Automation

Agentic software connects language models to tools that modify files and interact with external services. Developers use model-based security checks to screen external content and user requests before agents act. System One models answer developer-defined questions with probabilities over predefined answers such as safe or unsafe, but classification accuracy alone does not establish whether these probabilities support reliable automation. We evaluate Jev, Laya, Decider, and…

Read at arXiv (research papers)

arXiv (research papers) Patrick Kenney SI

API Secrets Should Never Become Tokens in the LLM's Vocabulary: A Threat Analysis of API Credential Handling in LLM Agent Systems and an Empirical Evaluation of a Vault-Mediated Execution Boundary

Tool-using large language model (LLM) agents turn credential hygiene from a storage problem into an execution-security problem. A key pasted into a prompt, or embedded in a system prompt or tool configuration, crosses from an authentication boundary into a data pipeline, where it may persist in conversation history, logs, memory stores, generated code, and error payloads. Prompt injection and excessive agency then convert passive disclosure into unauthorized action. This…

Read at arXiv (research papers)

arXiv (research papers) Ben Hagag SI

ORBIT: A Framework for Multi-Agent Safety and Security Evaluations

Multi-agent LLM systems are increasingly deployed for complex, long-horizon tasks or emerge as a natural consequence of agents interacting in the wild. Yet they give rise to significant safety and security risks: the flexible protocols that enable task generalization also expose novel threats, from cascading prompt injection to inter-agent collusion. Progress in defending against these threats has been slowed by a lack of shared empirical infrastructure, which forces bespoke…

Read at arXiv (research papers)

Journal articles (OpenAlex) Marcela Aurora Tapia Ladino, Romina Ramos Rodríguez, Angélica Alvites Baiadera · Journal of Borderlands Studies Geopolitics

Border Reconfigurations in Twenty-First-Century Latin America: Mobile Architectures of “Sustained Exception”

This article analyzes border control and migration governance strategies in Latin America the early twenty-first century, focusing on links between security agendas and cross-border mobility. Using a multilevel qualitative methodology, we examine the principal security initiatives shaping the role of borders in Latin America before and during the COVID-19 pandemic, and situate them within the governance frameworks promoted by the United States. The pandemic acted as a…

Read at Journal articles (OpenAlex)

KabulNow Nazanin Mohseni

12 Killed, 35 Injured in Suicide Blast at Police Checkpost in Northwestern Pakistan

KABUL, AFGHANISTAN - At least 12 people were killed and 35 others injured on Saturday after a vehicle laden with explosives was rammed into the Aman Mela police checkpost in Darazinda tehsil of Dera Ismail Khan in Pakistan’s Khyber Pakhtunkhwa province, Pakistani officials said. Dera Ismail Khan Regional Police Officer Ghulam Mubashir Maken, citing an initial assessment, confirmed the casualties and said no law enforcement or security personnel were among those killed. The…

Read at KabulNow

arXiv (research papers) Animesh Shaw SI

Silent Failures in Agentic Security Evaluation: A Validated Harness for Tool-Call Mediation Under Indirect Prompt Injection

LLM agents that invoke privileged tools are vulnerable to indirect prompt injection (IPI), in which adversarial instructions embedded in retrieved data hijack the agent's actions. A growing body of work evaluates defenses against IPI, but the validity of that evaluation is rarely examined. We audit an IPI benchmark and its harness and identify four defect classes -- silent payload non-delivery, attack success scored by tool identity rather than arguments, false-rejection…

Read at arXiv (research papers)

Journal articles (OpenAlex) Gulnaz Kanafina · Space and Culture India Geopolitics

Socio-Political and Cultural Transformation of Central Asian States in the Era of Global Crises: Challenges to Security and Resilience

The study examines the socio-political transformations unfolding in Central Asia and seeks to identify the principal challenges and threats shaping regional development. The research draws on a diverse set of sources, including Deutsche Welle, the United Nations Development Programme, the World Bank Group, Tengri News, and Radio Azattyk. The methodological framework combines legal, content, comparative, and statistical analyses. The structured analytical approach employed in…

Read at Journal articles (OpenAlex)

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)

The Rio Times Florencia Belén Ruiz Geopolitics

Shield of the Americas: 2 Chile Ex-Presidents Ask Kast to Reconsider

Shield of the Americas: Bachelet and Boric asked Kast to reconsider Chile’s membership. On 25 September Security Minister Martín Arrau urged calm and said “one has to read the background”.

Read at The Rio Times

Asia Times Rommel Banlaoi South China Sea

Japan’s new ambassador will test Manila’s China balancing act

As Manila and Tokyo mark the 70th anniversary of normalized diplomatic relations, elevate their ties to a Comprehensive Strategic Partnership, and face rising security tensions in the West Philippine Sea, the arrival of Kitagawa Katsuro as Japan’s new ambassador to the Philippines comes at a pivotal moment in regional geopolitics. He arrives not merely as

Read at Asia Times

Journal articles (OpenAlex) Luz Galvis, Àlex Boso, Àlvaro Hofflinger et al. · Current Sociology Geopolitics

Everyday water insecurity in decentralized water systems: Emotions, coping, and justice across hydrosocial territories in Southern Chile

This study examines how rural households in southern Chile navigate water insecurity under the looming climate stress and industrial forestry. Drawing on an abductive thematic analysis of interviews conducted in 2022, we conceptualize insecurity as a multitemporal regime of uncertainty that jeopardizes the ontological security of daily life. We document how hydrosocial encirclement by plantations produces visible water exclusion, where water remains physically observable but…

Read at Journal articles (OpenAlex)

Global Voices Kwasi Cudjoe

Trinidad & Tobago’s Special Operations Bill and the geography of development

Identifying a security hot spot is only one part of understanding a place, [which] may also be a centre of employment, have weak infrastructure, strong local organisations, or youth vulnerabilities.

Read at Global Voices

The Guardian NATO Pjotr Sauer in Kyiv

Russia could attack a Nato country within months, Danish intelligence says

Warning comes hours after Poland says fire at Starlink satellite station was ‘act of sabotage’ Russia could launch a limited military attack against a Nato country in the coming months, Danish intelligence said, in one of the starkest warnings yet by western security services. In a report published on Thursday, the Danish defence intelligence service warned of a “low but growing risk” that Russia could launch long-range strikes on Nato infrastructure supporting Ukraine, or…

Read at The Guardian NATO

Committee to Protect Journalists CPJ Staff

3 Tunisian journalists briefly detained as media crackdown intensifies

New York, September 24, 2026—The Committee to Protect Journalists on Thursday demanded that Tunisian authorities hold accountable those responsible for detaining and allegedly mistreating three local journalists from the independent news website Nawaat. “Tunisian security forces’ alleged use of deception to lure journalists into detention, attempts to access their phones without a judicial warrant, and threats...

Read at Committee to Protect Journalists

Journal articles (OpenAlex) Marzhan S. Orumbayeva, Toktokuchuk B. Mamytov, Kurmangali Aimen Kuanyshbaykyzy · Asian Journal of Political Science Geopolitics

Counter-terrorism policy coordination in Central Asia: formal convergence and practical cooperation

At the beginning of the twenty-first century, the activity of international terrorist organisations intensified, prompting Central Asian countries to strengthen their counter-terrorism systems. This study examined terrorist threats in Central Asia in the context of global security trends. The main methods were statistical analysis, used to assess quantitative indicators such as the number of residents involved in extremist organisations in 2023, and the historical method…

Read at Journal articles (OpenAlex)

POMEPS (Project on Middle East Political Science) Prerna BalaEddy Geopolitics

Job Opening: Université Laval Tenure-track Faculty Position in International Security and Strategic Studies

The Department of Political Science at Université Laval is inviting applications for a full-time tenure-track position of Professor in International Security and Strategic Studies. The selected candidate will teach, generate original knowledge and help consolidate the Department’s scientific and academic leadership in this important field. Responsibilities include teaching at the undergraduate and the graduate (M.A.

Read at POMEPS (Project on Middle East Political Science)

Journal articles (OpenAlex) Rasmus Leander Nielsen, Jeppe Strandsbjerg · Journal of the British Academy Geopolitics

The evolution of Greenlandic security and defence policy: a quest for recognition

This article traces the recent development of security and defence policy in Greenland. Although constitutionally, questions of security and defence have been for Denmark to decide, Greenland has recently begun to formulate its own security and defence policies, as part of its quest to secure international recognition as an actor in its own right, and to play a part in shaping, and benefitting from, vast increases in defence investments. Traditionally, Greenland has cleaved…

Read at Journal articles (OpenAlex)

arXiv (research papers) Reshabh K Sharma SI

Ajar: Measuring Open Privilege in Agent Defenses

A language model agent acts through the tools it is given. The data it reads while working on a task can redirect what it does with those tools. A growing set of techniques for safe and secure agent execution therefore sits between the agent and its tools, aiming to enforce access control, information flow or isolation at that boundary. Today these techniques are evaluated on agent-security benchmarks built around indirect prompt injection. Those benchmarks judge a defense…

Read at arXiv (research papers)

Journal articles (OpenAlex) Gіzem Melek, Kübra Geysi, Zeynep Nur Sarı · Global Media and Communication Geopolitics

Framing the Ukrainian refugee crisis: A cross-country media analysis during the early Russia-Ukraine conflict

This study explores how media in the United States, United Kingdom, Germany and Türkiye framed the Ukrainian refugee crisis during the early phase of the Russia-Ukraine conflict. A content analysis of eight news outlets highlights a bias favouring Ukrainian refugees, with positive frames – including economic benefit, cultural affinity, victimhood, humanitarian aid and military support – dominating negative frames related to cultural, economic and security threats. The…

Read at Journal articles (OpenAlex)

Journal articles (OpenAlex) Yating Pan, Zhan Shu, Haoyang Long · Telecommunications Policy Geopolitics

Securitizing digital platforms: TikTok, algorithmic governance, and the construction of national security threats in the United States

This article examines how commercial digital platforms become framed as national security threats in contemporary policy environments. Focusing on the U.S. regulation of TikTok, it analyzes how a widely used social media platform was transformed into an object of exceptional governance requiring qualified divestiture or potential exclusion. The study draws on a multi-source corpus, including congressional hearings, executive orders, legislative texts, judicial rulings, and a…

Read at Journal articles (OpenAlex)

Journal articles (OpenAlex) Manish Barma, S. K. Sinha · European Politics and Society Geopolitics

The European Union between norms and power: strategic recalibration across China, Russia, and India

The European Union's external action has entered a period of significant strategic recalibration struggling between its longstanding normative identity and emerging security imperatives. With China's rapid economic expansion and growing clout and competition, the EU has reassessed its engagement with China, as Beijing's deployment of soft power through cultural diplomacy and economic initiatives like the Belt and Road Initiative (BRI) has acquired broader strategic and…

Read at Journal articles (OpenAlex)

Journal articles (OpenAlex) Kate Sullivan de Estrada · Geopolitics Geopolitics

Island State Agency as ‘Project Enactment’: The Regional Maritime Security Architecture of the Western Indian Ocean

Attention to Indian Ocean island states and questions around their ‘agency’ have grown recently, in tandem with renewed policy and scholarly interest in the Indian Ocean as a site of strategic competition. A challenge, however, is to move beyond framing island states only in relation to the great or major powers who seek political influence, territorial footholds, and unimpeded flows of goods and energy across this maritime space. In this article, I theorise agency as ‘the…

Read at Journal articles (OpenAlex)

The Guardian Haiti Agence France-Presse Geopolitics

More than a dozen arrested in connection with 2021 assassination of Haitian president

Jovenel Moïse, 53, was shot dead in July 2021 in a killing which sparked chaos in Haiti Eighteen people have been arrested in association with the 2021 assassination of the Haitian president Jovenel Moïse, according to US officials. Moïse, 53, was shot dead on 7 July 2021 at his private residence by a hired group of about 20 military-trained Colombian mercenaries. His security detail did not intervene to protect him.

Read at The Guardian Haiti

The Guardian Chile Reuters in Geneva Geopolitics

Michelle Bachelet withdraws from race to become UN secretary general

Former president of Chile drops out after support collapsed in security council straw poll, leaving seven candidates in contest Michelle Bachelet, a former president of Chile, has withdrawn her candidacy to become the next secretary general of the ⁠UN, ending an attempt ​that had struggled to gain traction among security council members. Bachelet, 74, announced ⁠her decision in a video posted on Instagram on Saturday, saying: “I will now continue, with the ⁠same…

Read at The Guardian Chile

arXiv (research papers) Fayeq Jeelani Syed SI

Trustworthy Agentic AI: Failure Modes, Mitigation Strategies, and a Lifecycle Framework for Autonomous LLM Systems

Agentic AI systems built on large language models can plan over multiple steps, use external tools, retain information in memory, and coordinate with other agents. These capabilities make them more useful than static language models, but they also introduce new security and operational risks. Untrusted content from websites, emails, documents, and databases can enter the same context as system instructions; persistent memory can carry compromised information across sessions…

Read at arXiv (research papers)

Article 19 callum

UN: Address transnational repression of Belarusian civil society

ARTICLE 19 expresses our steadfast support for the work of the Group of Independent Experts on Belarus. We remain deeply concerned about the continued transnational repression of Belarusian journalists, human rights defenders, and civil society actors in exile, including through digital means. Our recent research with Belarusian human rights defenders, journalists, and digital security experts

Read at Article 19

Journal articles (OpenAlex) Abeer Aqeel Mohammad Alsrour · Frontiers in Political Science Geopolitics

Recalibrated strategic hedging: explaining the evolution of Saudi Arabia’s regional foreign policy (2010–2025) through strategic hedging and Regional Security Complex Theory

Introduction Saudi Arabia’s regional foreign policy has undergone significant recalibration amid changing regional security dynamics, intensifying great-power competition, and evolving patterns of interdependence in the Middle East. This study examines how Saudi Arabia has managed these uncertainties between 2010 and 2025 through the combined analytical lenses of Strategic Hedging Theory and Regional Security Complex Theory (RSCT). Methods The study employs a qualitative…

Read at Journal articles (OpenAlex)

arXiv (research papers) Yuxuan Zhang SI

Origin Is All You Need: Provenance-Aware Transformers for Structural Trust-Boundary Separation

Indirect prompt injection (IPI) remains a central safety and security challenge for large language model (LLM) systems because standard transformers lack architectural notion of source authority. Retrieved documents, user inputs, and system instructions are all processed through the same undifferentiated attention mechanism, forcing the model to infer from wording alone what should be obeyed and what should be treated as data. We propose Provenance-Aware Transformers, a…

Read at arXiv (research papers)

FDD's Long War Journal Ahmad Sharawi & Justin Leopold-Cohen Geopolitics

Syria’s military is rebuilding, but foreign influence may slow security progress

Syria’s post-Assad military is rebuilding rapidly, but its structure and foreign dependencies could limit Damascus’s ability to advance US security interests. If security cooperation from Turkey and Russia comes with strings attached, it may jeopardize Syria’s potential as a future partner against the Islamic State and Hezbollah.

Read at FDD's Long War Journal

Older stories