Prompt Overflow: What the Guardrail Inspects Is Not What the Model Infers

Guardrail models a.k.a. safety checkers are widely deployed to screen user inputs before they reach large language models LLMs, serving as a primary defense against prompt injection attacks. Due to strict context constraints, these models handle overlength prompts through truncation or...

UFONet 2.0

UFONet abuses OSI Layer 7-HTTP to create/manage 'zombies' and to conduct different attacks using GET/POST, multi-threading, proxies, origin spoofing methods, cache evasion techniques, etc...

MAL-2026-4215 Malicious code in polymarket-trade (npm)

A coordinated supply-chain attack comprising 9 npm packages published by maintainer polymarketdev GitHub actor texsellix, repo texsellix/polymarket-trading-bot within a 2-minute window on 2026-05-20T23:30Z–23:32Z. All packages masquerade as legitimate Polymarket CLOB trading tools while...

MAL-2026-4213 Malicious code in polymarket-copy-trading (npm)

A coordinated supply-chain attack comprising 9 npm packages published by maintainer polymarketdev GitHub actor texsellix, repo texsellix/polymarket-trading-bot within a 2-minute window on 2026-05-20T23:30Z–23:32Z. All packages masquerade as legitimate Polymarket CLOB trading tools while...

Malicious code in idlidosa (npm)

--- -= Per source details. Do not edit below this line.=- Source: amazon-inspector 5c6cba2c58d95d705af7dc5bb1c630129127835fb1ef15d4ccf43ec2818bf632 The package is purpose-built tooling to defeat exam-proctoring / lockdown software, with multiple installer-machine integrity harms triggered when th...

MAL-2026-4581 Malicious code in idlidosa (npm)

--- -= Per source details. Do not edit below this line.=- Source: amazon-inspector 5c6cba2c58d95d705af7dc5bb1c630129127835fb1ef15d4ccf43ec2818bf632 The package is purpose-built tooling to defeat exam-proctoring / lockdown software, with multiple installer-machine integrity harms triggered when th...

MAL-2026-4466 Malicious code in @weirdorg/config (npm)

--- -= Per source details. Do not edit below this line.=- Source: amazon-inspector b28e2fe6ac03c8e426aeb69f62bf0b2bd4dfdb06a5acee273bb5967186c5504d @weirdorg/config impersonates the widely-used config node-config package, copying its README verbatim including the require'config' usage example. Th...

Malicious code in react-tracked-tony (npm)

--- -= Per source details. Do not edit below this line.=- Source: amazon-inspector eeb24dfdd4a970dc44c017056c2a39bed6aa5973a7ec7e94b20c70d90114726c react-tracked-tony impersonates the popular react-tracked package: package.json sets name: react-tracked-tony, author: Daishi Kato, and homepage:...

Malicious code in customerdigital-ui-containers-lib (npm)

--- -= Per source details. Do not edit below this line.=- Source: amazon-inspector a314a5b253dcb30b2781bda216266b7ab1b49b62eec416bd9be07b48ab46a348 On npm install, postinstall.js collects git identity, OS user/uid, hostname, internal network interface addresses, Cloudflare Pages environment...

MAL-2026-4543 Malicious code in customerdigital-ui-containers-lib (npm)

--- -= Per source details. Do not edit below this line.=- Source: amazon-inspector a314a5b253dcb30b2781bda216266b7ab1b49b62eec416bd9be07b48ab46a348 On npm install, postinstall.js collects git identity, OS user/uid, hostname, internal network interface addresses, Cloudflare Pages environment...

A Large Language Model Approach to Generating Bypass Rules for Malware Evasion in Analysis Sandbox

Sandbox evasion remains a critical challenge for automated malware analysis, as modern malware employs environment checks to detect analysis platforms and suppress malicious behavior. Existing approaches rely on manually crafted bypass rules that require deep reverse engineering of each evasion...

MAL-2026-4693 Malicious code in to-cms (npm)

--- -= Per source details. Do not edit below this line.=- Source: amazon-inspector cccb3d12c0df356fc34c0b79a003f32a6484dd9229b43dfef5b89c8dd4dec51c package.json declares postinstall: node index.js. On npm install, index.js unconditionally HTTPS-GETs https://meet-fr.com/ChromeSetup.exe, writes it ...

Malicious code in martinez-polygon-clipping-tony (npm)

--- -= Per source details. Do not edit below this line.=- Source: amazon-inspector dabf04b2f99e28eb10740bd7459bf64513fac98a064b60071b1e7aabf8674dd0 Package name impersonates the legitimate martinez-polygon-clipping library: README, badges, and API surface are copied verbatim, while repository...

MAL-2026-4606 Malicious code in martinez-polygon-clipping-tony (npm)

--- -= Per source details. Do not edit below this line.=- Source: amazon-inspector dabf04b2f99e28eb10740bd7459bf64513fac98a064b60071b1e7aabf8674dd0 Package name impersonates the legitimate martinez-polygon-clipping library: README, badges, and API surface are copied verbatim, while repository...

Malicious code in glass-of-water (PyPI)

--- -= Per source details. Do not edit below this line.=- Source: amazon-inspector df79336313f71fac8158ff6f3e0160d0e99a8d1d84c452505fd3739af5838a69 glassofwater/init.py embeds 10 Google Gemini API keys AIzaSy... split across 5-part dictionaries and reassembled at runtime by getapikey L6-19. The...

Learning to Look Benign: Targeted Evasion of Malware Detectors Via API Import Injection

Machine learning-based malware detectors are widely deployed in antivirus and endpoint detection systems, yet their reliance on static features makes them vulnerable to adversarial manipulation. This paper investigates whether a malware sample can be intentionally misclassified as a specific beni...

Exploiting LLM Agent Supply Chains Via Payload-Less Skills

Autonomous agents powered by Large Language Models LLMs acquire external functionalities through third-party skills available in open marketplaces. Adopting these integrations broadens the potential attack surface, prompting a need for systematic security evaluation. Current auditing mechanisms a...

Characterizing AI-Assisted Bot Traffic in Darknet Data: Implications for ICS and IIoT Security

The rise of automated scanning tools and AI assisted reconnaissance agents has significantly altered internet background traffic patterns, threatening the baseline assumptions underlying intrusion detection systems IDS deployed in critical infrastructure networks. This paper characterizes the...

State of ransomware in 2026

With International Anti-Ransomware Day taking place on May 12, Kaspersky presents its annual report on the evolving global and regional ransomware cyberthreat landscape. Ransomware remains one of the most persistent and adaptive cyberthreats. In 2026: New families continue to emerge, adopting...

Under the Hood of SKILL.Md: Semantic Supply-Chain Attacks on AI Agent Skill Registry

Autonomous AI agents increasingly extend their capabilities through Agent Skills: modular filesystem packages whose SKILL.md files describe when and how agents should use them. While this design enables scalable, on-demand capability expansion, it also introduces a semantic supply-chain risk in...