Cross-LLM Generalization of Behavioral Backdoor Detection in AI Agent Supply Chains

As AI agents become integral to enterprise workflows, their reliance on shared tool libraries and pre-trained components creates significant supply chain vulnerabilities. While previous work has demonstrated behavioral backdoor detection within individual LLM architectures, the critical question ...

Exploit for OS Command Injection in Vsftpd_Project Vsftpd

metasploitable-Security-Assessment comprehensive penetration t...

APT24 Deploys BADAUDIO in Years-Long Espionage Hitting Taiwan and 1,000+ Domains

A China-nexus threat actor known as APT24 has been observed using a previously undocumented malware dubbed BADAUDIO to establish persistent remote access to compromised networks as part of a nearly three-year campaign. "While earlier operations relied on broad strategic web compromises to...

EdgeStepper Implant Reroutes DNS Queries to Deploy Malware via Hijacked Software Updates

The threat actor known as PlushDaemon has been observed using a previously undocumented Go-based network backdoor codenamed EdgeStepper to facilitate adversary-in-the-middle AitM attacks. EdgeStepper "redirects all DNS queries to an external, malicious hijacking node, effectively rerouting the...

MAL-2025-190564 Malicious code in tailwindcss-breezium (npm)

--- -= Per source details. Do not edit below this line.=- Source: amazon-inspector f92d0ba07d104ef809e0b5b17fe93670e80af4cceecde2f9afb14f8357febcdc The package tailwindcss-breezium was found to contain malicious code. Source: ghsa-malware...

Steganographic Backdoor Attacks in NLP: Ultra-Low Poisoning and Defense Evasion

Transformer models are foundational to natural language processing NLP applications, yet remain vulnerable to backdoor attacks introduced through poisoned data, which implant hidden behaviors during training. To strengthen the ability to prevent such compromises, recent research has focused on...

BackWeak: Backdooring Knowledge Distillation Simply with Weak Triggers and Fine-Tuning

Knowledge Distillation KD is essential for compressing large models, yet relying on pre-trained "teacher" models downloaded from third-party repositories introduces serious security risks -- most notably backdoor attacks. Existing KD backdoor methods are typically complex and computationally...

EUVD-2021-34716

Denver SHO-110 IP cameras expose a secondary HTTP service on TCP port 8001 that provides access to a '/snapshot' endpoint without authentication. While the primary web interface on port 80 enforces authentication, the backdoor service allows any remote attacker to retrieve image snapshots by...

CVE-2021-4469

CVE-2021-4469 : Denver SHO-110 IP cameras expose a secondary HTTP service on port 8001 with an unauthenticated /snapshot endpoint. While port 80 requires authentication, the backdoor service allows remote attackers to fetch snapshots directly, enabling repeated collection and potential reconstruc...

CVE-2021-4469 Denver SHO-110 IP Camera Unauthenticated Snapshot Access

Denver SHO-110 IP cameras expose a secondary HTTP service on TCP port 8001 that provides access to a '/snapshot' endpoint without authentication. While the primary web interface on port 80 enforces authentication, the backdoor service allows any remote attacker to retrieve image snapshots by...

North Korean Hackers Turn JSON Services into Covert Malware Delivery Channels

The North Korean threat actors behind the Contagious Interview campaign have once again tweaked their tactics by using JSON storage services to stage malicious payloads. "The threat actors have recently resorted to utilizing JSON storage services like JSON Keeper, JSONsilo, and npoint.io to host...

We opened a fake invoice and fell down a retro XWorm-shaped wormhole

Somebody forwarded an “invoice” email and asked me to check the attachment because it looked suspicious. Good instinct—it was, and what we found inside was a surprisingly old trick hiding a modern threat. What it does If the recipient had opened the attached Visual Basic Script .vbs file, it woul...

MTAttack: Multi-Target Backdoor Attacks against Large Vision-Language Models

Recent advances in Large Visual Language Models LVLMs have demonstrated impressive performance across various vision-language tasks by leveraging large-scale image-text pretraining and instruction tuning. However, the security vulnerabilities of LVLMs have become increasingly concerning,...

GootLoader Is Back, Using a New Font Trick to Hide Malware on WordPress Sites

The malware known as GootLoader has resurfaced yet again after a brief spike in activity earlier this March, according to new findings from Huntress. The cybersecurity company said it observed three GootLoader infections since October 27, 2025, out of which two resulted in hands-on keyboard...

Hundred Plus EIP Plus 代码问题漏洞

Hundred Plus EIP Plus is an enterprise management software from Hundred Plus Ares Hundred Plus of Taiwan, China. Hundred Plus EIP Plus suffers from a code issue vulnerability that originates from allowing a privileged remote attacker to upload and execute a web backdoor that could lead to the...

Trojanized ESET Installers Drop Kalambur Backdoor in Phishing Attacks on Ukraine

A previously unknown threat activity cluster has been observed impersonating Slovak cybersecurity company ESET as part of phishing attacks targeting Ukrainian entities. The campaign, detected in May 2025, is tracked by the security outfit under the moniker InedibleOchotense , describing it as...

Cavalry Werewolf Hit Russian Government with New ShellNET Backdoor

Doctor Web uncovers a targeted cyberattack on a Russian government body by the Cavalry Werewolf group using a new ShellNET backdoor and Telegram-based control...

SesameOp Backdoor Abused OpenAI Assistants API for Remote Access

Microsoft researchers found the SesameOp backdoor using OpenAI’s Assistants API for remote access, data theft, and command communication...

Operation SkyCloak Deploys Tor-Enabled OpenSSH Backdoor Targeting Defense Sectors

Threat actors are leveraging weaponized attachments distributed via phishing emails to deliver malware likely targeting the defense sector in Russia and Belarus. According to multiple reports from Cyble and Seqrite Labs, the campaign is designed to deploy a persistent backdoor on compromised host...

Microsoft Detects "SesameOp" Backdoor Using OpenAI's API as a Stealth Command Channel

Microsoft has disclosed details of a novel backdoor dubbed SesameOp that uses OpenAI Assistants Application Programming Interface API for command-and-control C2 communications. "Instead of relying on more traditional methods, the threat actor behind this backdoor abuses OpenAI as a C2 channel as ...