Agentic AI Security: Threats, Defenses, Evaluation, and Open Challenges

Agentic AI systems powered by large language models LLMs and endowed with planning, tool use, memory, and autonomy, are emerging as powerful, flexible platforms for automation. Their ability to autonomously execute tasks across web, software, and physical environments creates new and amplified...

Breaking Agent Backbones: Evaluating the Security of Backbone LLMs in AI Agents

AI agents powered by large language models LLMs are being deployed at scale, yet we lack a systematic understanding of how the choice of backbone LLM affects agent security. The non-deterministic sequential nature of AI agents complicates security modeling, while the integration of traditional...

SecureLearn - an Attack-Agnostic Defense for Multiclass Machine Learning against Data Poisoning Attacks

Data poisoning attacks are a potential threat to machine learning ML models, aiming to manipulate training datasets to disrupt their performance. Existing defenses are mostly designed to mitigate specific poisoning attacks or are aligned with particular ML algorithms. Furthermore, most defenses a...

Expression Language Injection

Spring Cloud Gateway Server Webflux is vulnerable to Expression Language Injection. The vulnerability is due to unsafe SpEL evaluation in routes due to the actuator gateway endpoint being exposed and accessible to untrusted users; attackers can create routes that use SpEL to read environment...

Remote Code Execution (RCE)

Flowise is vulnerable to remote code execution RCE. The vulnerability is due to unsanitized evaluation of user input in the “Supabase RPC Filter” field, which allows an attacker to execute arbitrary code on the affected system...

Jailbreak Mimicry: Automated Discovery of Narrative-Based Jailbreaks for Large Language Models

Large language models LLMs remain vulnerable to sophisticated prompt engineering attacks that exploit contextual framing to bypass safety mechanisms, posing significant risks in cybersecurity applications. We introduce Jailbreak Mimicry, a systematic methodology for training compact attacker mode...

Enhanced MLLM Black-Box Jailbreaking Attacks and Defenses

Multimodal large language models MLLMs comprise of both visual and textual modalities to process vision language tasks. However, MLLMs are vulnerable to security-related issues, such as jailbreak attacks that alter the model's input to induce unauthorized or harmful responses. The incorporation o...

The Qey: Implementation and Performance Study of Post Quantum Cryptography in FIDO2

Authentication systems have evolved a lot since the 1960s when Fernando Corbato first proposed the password-based authentication. In 2013, the FIDO Alliance proposed using secure hardware for authentication, thus marking a milestone in the passwordless authentication era 1. Passwordless...

Security Bulletin: multiple vulerability in IBM Spectrum Symphony with Bouncy Castle Java TLS API

Summary multiple vulerability in IBM Spectrum Symphony with Bouncy Castle Java TLS API Vulnerability Details CVEID:CVE-2024-30172 DESCRIPTION: An issue was discovered in Bouncy Castle Java Cryptography APIs before 1.78. An Ed25519 verification code infinite loop can occur via a crafted signature...

CVE-2023-53708

In the Linux kernel, the following vulnerability has been resolved: ACPI: x86: s2idle: Catch multiple ACPITYPEPACKAGE objects If a badly constructed firmware includes multiple ACPITYPEPACKAGE objects while evaluating the AMD LPS0 DSM, there will be a memory leak. Explicitly guard against this...

Ask What Your Country Can Do for You: Towards a Public Red Teaming Model

AI systems have the potential to produce both benefits and harms, but without rigorous and ongoing adversarial evaluation, AI actors will struggle to assess the breadth and magnitude of the AI risk surface. Researchers from the field of systems design have developed several effective sociotechnic...

OSV-2025-846 Use-of-uninitialized-value in js_create_function

OSS-Fuzz report: https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=453198741 Crash type: Use-of-uninitialized-value Crash state: jscreatefunction JSEvalInternal JSEvalObject...

Multimodal Safety Is Asymmetric: Cross-Modal Exploits Unlock Black-Box MLLMs Jailbreaks

Multimodal large language models MLLMs have demonstrated significant utility across diverse real-world applications. But MLLMs remain vulnerable to jailbreaks, where adversarial inputs can collapse their safety constraints and trigger unethical responses. In this work, we investigate jailbreaks i...

GHSA-PCXQ-FJP3-R752 Ash has authorization bypass when bypass policy condition evaluates to true

Summary Bypass policies incorrectly authorize requests when their condition evaluates to true but their authorization checks fail and no other policies apply. Impact Resources with bypass policies can be accessed without proper authorization when: - Bypass condition evaluates to true - Bypass...

CVE-2025-48044

CVE-2025-48044 concerns an authorization bypass in ash-project ash due to a bug in lib/ash/policy/policy.ex (Elixir.Ash.Policy.Policy):expression/2. Affected versions are ash 3.6.3 up to, but not including, 3.7.1 (commit range 79749c... to 8b83efa...). The vulnerability can cause bypass when a by...

SoK: Taxonomy and Evaluation of Prompt Security in Large Language Models

Large Language Models LLMs have rapidly become integral to real-world applications, powering services across diverse sectors. However, their widespread deployment has exposed critical security risks, particularly through jailbreak prompts that can bypass model alignment and induce harmful outputs...

PT-2025-46066

Uncontrolled recursion in XPath evaluation in libxml2 up to and including version 2.9.14 allows a local attacker to cause a stack overflow via crafted expressions. XPath processing functions xmlXPathRunEval, xmlXPathCtxtCompile, and xmlXPathEvalExpr were resetting recursion depth to zero before...

CVE-2025-57164: Improper Neutralization of Special Elements used in a Command

Flowise through v3.0.4 is vulnerable to remote code execution via unsanitized evaluation of user input in the "Supabase RPC Filter" field...

Architectures, Risks, and Adoption: How to Assess and Choose the Right AI-SOC Platform

Scaling the SOC with AI - Why now? Security Operations Centers SOCs are under unprecedented pressure. According to SACR's AI-SOC Market Landscape 2025 , the average organization now faces around 960 alerts per day , while large enterprises manage more than 3,000 alerts daily from an average of 28...

Architectures, Risks, and Adoption: How to Assess and Choose the Right AI-SOC Platform

Scaling the SOC with AI - Why now? Security Operations Centers SOCs are under unprecedented pressure. According to SACR’s AI-SOC Market Landscape 2025 , the average organization now faces around 960 alerts per day , while large enterprises manage more than 3,000 alerts daily from an average of 28...