EUVD-2026-26053

GNU nano creates the user’s /.local directory with overly permissive permissions when the directory does not exist yet. On first use of features requiring Cross-Desktop Group XDG data storage, nano explicitly requests directory mode 0777, making the directory world‑writable in environments where...

EUVD-2026-26012

In Spring AI, having access to a shared environment can expose the ONNX model used by the application. Affected versions: Spring AI: 1.0.0 - 1.0.5 fixed in 1.0.6, 1.1.0 - 1.1.4 fixed in 1.1.5...

ONNX model cache defaults to world-writable predictable /tmp directory

In Spring AI, having access to a shared environment can expose the ONNX model used by the application. Only applications that use TransformersEmbeddingModel and have the cache enabled, using the default location, are affected...

The vulnerability of the application management tools and Flatpak environments relates to incorrect path name restrictions, allowing attackers to gain access to delete arbitrary files.

The vulnerability of the application management tool and Flatpak environments is related to an incorrect limitation on the path name to the catalog. Exploiting this vulnerability could allow a remote attacker to gain access to delete any files they desire...

The Security Cost of Intelligence: AI Capability, Cyber Risk, and Deployment Paradox

Firms are deploying more capable AI systems, but organizational controls often have not kept pace. These systems can generate greater productivity gains, but high-value uses require broader authority exposure -- data access, workflow integration, and delegated authority -- when governance control...

CVE-2026-35367

The nohup utility in uutils coreutils creates its default output file, nohup.out, without specifying explicit restricted permissions. This causes the file to inherit umask-based permissions, typically resulting in a world-readable file 0644. In multi-user environments, this allows any user on the...

GHSA-HX6P-XPX3-JVVV vulnerabilities

Vulnerabilities for packages: zed, wasmcloud, wizer, yara-x...

CVE-2026-6842

A flaw was found in nano. In environments with permissive umask settings, a local attacker can exploit incorrect directory permissions 0777 instead of 0700 for the /.local directory. This allows the attacker to inject a malicious .desktop launcher, which could lead to unintended actions or...

Embedded Malicious Code

Overview Affected versions of this package are vulnerable to Embedded Malicious Code that target Kubernetes environments by install a full LLM proxy service on the victim's machine, allowing the attacker to route LLM traffic through the compromised server. Remediation Avoid using kube-node-health...

CVE-2026-35251

Vulnerability in the Oracle VM VirtualBox product of Oracle Virtualization component: Core. The supported version that is affected is 7.2.6. Difficult to exploit vulnerability allows high privileged attacker with logon to the infrastructure where Oracle VM VirtualBox executes to compromise Oracle...

CVE-2026-35248

Vulnerability in the Oracle VM VirtualBox product of Oracle Virtualization component: Core. The supported version that is affected is 7.2.6. Difficult to exploit vulnerability allows high privileged attacker with logon to the infrastructure where Oracle VM VirtualBox executes to compromise Oracle...

CVE-2026-35250

Vulnerability in the Oracle VM VirtualBox product of Oracle Virtualization component: Core. The supported version that is affected is 7.2.6. Easily exploitable vulnerability allows high privileged attacker with logon to the infrastructure where Oracle VM VirtualBox executes to compromise Oracle V...

CVE-2026-40566

FreeScout is a free self-hosted help desk and shared mailbox. Versions prior to 1.8.213 have a Server-Side Request Forgery SSRF vulnerability in the IMAP/SMTP connection test functionality of FreeScout's MailboxesController. Three AJAX actions fetchtest line 731, sendtest line 682, and imapfolder...

Mapping Your API Ecosystem: Wiz Expands API Discovery with Apigee

See your full Apigee architecture on the Wiz Security Graph, from API gateways and environments to every endpoint and its authorization scheme...

Home Assistant 代码注入漏洞

Home Assistant is an open-source family automation management system developed by Home Assistant. This system is primarily used to control household automation devices. Versions of Home Assistant prior to 1.0.0 had a code injection vulnerability. This vulnerability stemmed from the use of unlimit...

Do Agents Dream of Root Shells? Partial-Credit Evaluation of LLM Agents in Capture the Flag Challenges

Large Language Model LLM agents are increasingly proposed for autonomous cybersecurity tasks, but their capabilities in realistic offensive settings remain poorly understood. We present DeepRed, an open-source benchmark for evaluating LLM-based agents on realistic Capture The Flag CTF challenges ...

API Security Based on Automatic OpenAPI Mapping

This paper presents Map Reduce Graph MRG, a novel unsupervised method for modeling and securing HTTP REST APIs. MRG learns API structure from real-world traffic without prior knowledge or labels, automatically generating OpenAPI-compliant documentation by reconstructing routes, methods, and...

CVE-2026-29648

In OpenXiangShan NEMU, when Smstateen is enabled, clearing mstateen0.ENVCFG does not correctly restrict access to henvcfg and senvcfg. As a result, less-privileged code may read or write these CSRs without the required exception, potentially bypassing intended state-enable based isolation control...

Zeek 8.0.7

Zeek is a powerful network analysis framework that is much different from the typical IDS you may know. While focusing on network security monitoring, Zeek provides a comprehensive platform for more general network traffic analysis as well. Well grounded in more than 15 years of research, Zeek ha...

Terminal Wrench: A Dataset of 331 Reward-Hackable Environments and 3,632 Exploit Trajectories

The authors of this paper release Terminal Wrench, a subset of 331 terminal-agent benchmark environments, copied from the popular open benchmarks that are demonstrably reward-hackable. The data set includes 3,632 hack trajectories and 2,352 legitimate baseline trajectories across three frontier...