CoopGuard: Stateful Cooperative Agents Safeguarding LLMs against Evolving Multi-Round Attacks

As Large Language Models LLMs are increasingly deployed in complex applications, their vulnerability to adversarial attacks raises urgent safety concerns, especially those evolving over multi-round interactions. Existing defenses are largely reactive and struggle to adapt as adversaries refine...

Towards Unveiling Vulnerabilities of Large Reasoning Models in Machine Unlearning

Large language models LLMs possess strong semantic understanding, driving significant progress in data mining applications. This is further enhanced by large reasoning models LRMs, which provide explicit multi-step reasoning traces. On the other hand, the growing need for the right to be forgotte...

LLM-Enabled Open-Source Systems in the Wild: An Empirical Study of Vulnerabilities in GitHub Security Advisories

Large language models LLMs are increasingly embedded in open-source software OSS ecosystems, creating complex interactions among natural language prompts, probabilistic model outputs, and execution-capable components. However, it remains unclear whether traditional vulnerability disclosure...

Your Agent Is More Brittle Than You Think: Uncovering Indirect Injection Vulnerabilities in Agentic LLMs

The rapid deployment of open-source frameworks has significantly advanced the development of modern multi-agent systems. However, expanded action spaces, including uncontrolled privilege exposure and hidden inter-system interactions, pose severe security challenges. Specifically, Indirect Prompt...

SecPI: Secure Code Generation with Reasoning Models Via Security Reasoning Internalization

Reasoning language models RLMs are increasingly used in programming. Yet, even state-of-the-art RLMs frequently introduce critical security vulnerabilities in generated code. Prior training-based approaches for secure code generation face a critical limitation that prevents their direct applicati...

Perceptual Gaps: ASCII Art and Overlapping Audio As CAPTCHA

As multimodal large language models LLMs advance, traditional CAPTCHAs have become obsolete at distinguishing humans from bots. To address this shift, this paper aims to investigate the possibility of using tasks for which humans have evolved highly specialised neural processing. We introduce two...

CVE-2025-68152

Juju is an open source application orchestration engine that enables any application operation on any infrastructure at any scale through special operators called ‘charms’. From versions 2.9 to before 2.9.56 and 3.6 to before 3.6.19, it is possible that a compromised workload machine under a Juju...

CVE-2025-68152 Juju: Read All Controller Logs From Compromised Workload

Juju is an open source application orchestration engine that enables any application operation on any infrastructure at any scale through special operators called ‘charms’. From versions 2.9 to before 2.9.56 and 3.6 to before 3.6.19, it is possible that a compromised workload machine under a Juju...

CVE-2026-34760 vLLM: Downmix Implementation Differences as Attack Vectors Against Audio AI Models

vLLM is an inference and serving engine for large language models LLMs. From version 0.5.5 to before version 0.18.0, Librosa defaults to using numpy.mean for mono downmixing tomono, while the international standard ITU-R BS.775-4 specifies a weighted downmixing algorithm. This discrepancy results...

Combating Data Laundering in LLM Training

Data rights owners can detect unauthorized data use in large language model LLM training by querying with proprietary samples. Often, superior performance e.g., higher confidence or lower loss on a sample relative to the untrained data implies it was part of the training corpus, as LLMs tend to...

PT-2026-29877

vLLM is an inference and serving engine for large language models LLMs. From version 0.5.5 to before version 0.18.0, Librosa defaults to using numpy.mean for mono downmixing to mono, while the international standard ITU-R BS.775-4 specifies a weighted downmixing algorithm. This discrepancy result...

Automated Malware Family Classification Using Weighted Hierarchical Ensembles of Large Language Models

Malware family classification remains a challenging task in automated malware analysis, particularly in real-world settings characterized by obfuscation, packing, and rapidly evolving threats. Existing machine learning and deep learning approaches typically depend on labeled datasets, handcrafted...

Insufficient Session Expiration

Overview Affected versions of this package are vulnerable to Insufficient Session Expiration due to the failure to revoke existing authenticated sessions after a password reset or password change process. An attacker can maintain unauthorized access to an account by reusing a previously obtained...

CVE-2026-5212

A vulnerability has been found in D-Link DNS-120, DNR-202L, DNS-315L, DNS-320, DNS-320L, DNS-320LW, DNS-321, DNR-322L, DNS-323, DNS-325, DNS-326, DNS-327L, DNR-326, DNS-340L, DNS-343, DNS-345, DNS-726-4, DNS-1100-4, DNS-1200-05 and DNS-1550-04 up to 20260205. This issue affects the function...

CVE-2026-5312

A weakness has been identified in D-Link DNS-120, DNR-202L, DNS-315L, DNS-320, DNS-320L, DNS-320LW, DNS-321, DNR-322L, DNS-323, DNS-325, DNS-326, DNS-327L, DNR-326, DNS-340L, DNS-343, DNS-345, DNS-726-4, DNS-1100-4, DNS-1200-05 and DNS-1550-04 up to 20260205. Affected by this vulnerability is the...

acetone-nnet (>=0.1.0 <=0.4.0.dev1), acuity (=6.18.0) +370 more potentially affected by CVE-2026-34446 via onnx (>=0.2.0 <=1.20.1)

onnx PYPI version =0.2.0, =0.1.0, =0.1.0, =0.0.0, =0.0.157, =0.1.0, =0.1.8, =1.7.0, =1.3.0, =0.10.0, =0.3.1, =1.0.2 and more Source cves: CVE-2026-34446 Source advisory: OSV:GHSA-CMW6-HCPP-C6JP...

CVE-2026-5312

A weakness has been identified in D-Link DNS-120, DNR-202L, DNS-315L, DNS-320, DNS-320L, DNS-320LW, DNS-321, DNR-322L, DNS-323, DNS-325, DNS-326, DNS-327L, DNR-326, DNS-340L, DNS-343, DNS-345, DNS-726-4, DNS-1100-4, DNS-1200-05 and DNS-1550-04 up to 20260205. Affected by this vulnerability is the...

UBUNTU-CVE-2026-34445

Open Neural Network Exchange ONNX is an open standard for machine learning interoperability. Prior to version 1.21.0, the ExternalDataInfo class in ONNX was using Python’s setattr function to load metadata like file paths or data lengths directly from an ONNX model file. It didn’t check if the...

CVE-2026-34445

CVE-2026-34445 affects ONNX prior to version 1.21.0, where ExternalDataInfo used Python setattr() to load metadata directly from model files without validating keys, enabling a malicious model to overwrite internal object properties. Impact is mainly availability (HIGH) with confidentiality and i...

CVE-2026-34445 ONNX: Malicious ONNX models can crash servers by exploiting unprotected object settings.

Open Neural Network Exchange ONNX is an open standard for machine learning interoperability. Prior to version 1.21.0, the ExternalDataInfo class in ONNX was using Python’s setattr function to load metadata like file paths or data lengths directly from an ONNX model file. It didn’t check if the...