PYSEC-2026-3406 vLLM: image EXIF Rotation & PNG tRNS Transparency Not Normalized, Causing Mismatch Between Model Input and Expectations

Summary Issue 1: EXIF orientation not normalized → The image orientation processed by the model differs from how humans view it, introducing interpretation bias. Issue 2: PNG tRNS not explicitly flattened before converting to RGB → After conversion, transparent/semi-transparent pixels are rendere...

PYSEC-2026-2903 PraisonAI CLI automatically resolves @url mentions in prompt text and can read loopback URLs into model context

Summary PraisonAI's direct-prompt CLI automatically expands @url: mentions in raw prompt text before agent execution begins. If a prompt contains @url:, the CLI calls MentionsParser.process.... The @url: handler then performs a direct urllib.request.urlopen request to the attacker-controlled URL...

CVE-2026-55646

vLLM is an inference and serving engine for large language models. From 0.22.0 to 0.23.0, the /v1/audio/transcriptions and /v1/audio/translations routes call request.file.read to fully materialize an uploaded audio file into memory before vLLM checks the documented VLLMMAXAUDIOCLIPFILESIZEMB...

XXE Vulnerability in MerlinReader

Impact MPXJ used the default configuration when creating a DocumentBuilder instance, which leaves doctype declarations enabled, when parsing the XML content of the ZTIMEINTERVALS column from a Merlin project SQLite file. This would allow a carefully crafted XML payload to read an arbitrary file...

NCMD: Benign-Anchored Feature Selection for Imbalanced Network Intrusion Detection

Feature selection is critical for network intrusion detection systems NIDS operating under high-dimensional, highly imbalanced traffic, as found in operational and defense networks. Traditional filter methods rank features using global statistics computed symmetrically across classes and thus fai...

PraisonAI CLI automatically resolves @url mentions in prompt text and can read loopback URLs into model context

Summary PraisonAI's direct-prompt CLI automatically expands @url: mentions in raw prompt text before agent execution begins. If a prompt contains @url:, the CLI calls MentionsParser.process.... The @url: handler then performs a direct urllib.request.urlopen request to the attacker-controlled URL...

Astra Linux – Vulnerability in Zabbix

JavaScript preprocessing, webhooks, and global scripts can lead to uncontrolled utilization of CPU, memory, and disk I/O resources. The ability to preprocess/webhook/configure and test global scripts is only available to Administrative roles Admin and Superadmin. Administrative privileges should...

Astra Linux – Vulnerability in Zabbix

JavaScript preprocessing can be exploited by attackers to gain access to the file system read-only access on behalf of the user “zabbix” on the Zabbix Server or Zabbix Proxy, potentially leading to unauthorized access to sensitive data...

CVE-2025-47390

Memory corruption while preprocessing IOCTL request in JPEG driver...

EUVD-2025-209225

Memory corruption while preprocessing IOCTL request in JPEG driver...

CVE-2025-47390

Memory corruption while preprocessing IOCTL request in JPEG driver...

CVE-2025-47390 Buffer Over-read in Camera

Memory corruption while preprocessing IOCTL request in JPEG driver...

CVE-2025-47390 Buffer Over-read in Camera

Memory corruption while preprocessing IOCTL request in JPEG driver...

CVE-2025-47390

CVE-2025-47390 describes memory corruption during preprocessing of an IOCTL request in the JPEG driver (also referenced as a Camera-related issue). The vulnerability is categorized with high impact (CVSSv3.1: AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H) and requires local access with low privileges; ther...

PT-2026-30636

Memory corruption while preprocessing IOCTL request in JPEG driver...

Machine Learning Transferability for Malware Detection

Malware continues to be a predominant operational risk for organizations, especially when obfuscation techniques are used to evade detection. Despite the ongoing efforts in the development of Machine Learning ML detection approaches, there is still a lack of feature compatibility in public...

Shape and Substance: Dual-Layer Side-Channel Attacks on Local Vision-Language Models

On-device Vision-Language Models VLMs promise data privacy via local execution. However, we show that the architectural shift toward Dynamic High-Resolution preprocessing e.g., AnyRes introduces an inherent algorithmic side-channel. Unlike static models, dynamic preprocessing decomposes images in...

CVE-2025-33249

NVIDIA NeMo Framework for all platforms contains a vulnerability in a voice-preprocessing script, where malicious input created by an attacker could cause a code injection. A successful exploit of this vulnerability might lead to code execution, escalation of privileges, information disclosure, a...

CVE-2025-33249

NVIDIA NeMo Framework for all platforms contains a vulnerability in a voice-preprocessing script, where malicious input created by an attacker could cause a code injection. A successful exploit of this vulnerability might lead to code execution, escalation of privileges, information disclosure, a...

CVE-2025-33249

NVIDIA NeMo Framework for all platforms contains a vulnerability in a voice-preprocessing script, where malicious input created by an attacker could cause a code injection. A successful exploit of this vulnerability might lead to code execution, escalation of privileges, information disclosure, a...