Agent Security Is a Systems Problem

We take the position that agent security must be approached as a systems problem: the AI model powering the agent must be treated as an untrusted component, and security invariants must be enforced at the system level. Through this lens, efforts to increase model robustness the dominant viewpoint...

Hackers or Hallucinators? A Comprehensive Analysis of LLM-Based Automated Penetration Testing

The rapid advancement of Large Language Models LLMs has created new opportunities for Automated Penetration Testing AutoPT, spawning numerous frameworks aimed at achieving end-to-end autonomous attacks. However, despite the proliferation of related studies, existing research generally lacks...

Supporting Artifact Evaluation with LLMs: A Study with Published Security Research Papers

Artifact Evaluation AE is essential for ensuring the transparency and reliability of research, closing the gap between exploratory work and real-world deployment is particularly important in cybersecurity, particularly in IoT and CPSs, where large-scale, heterogeneous, and privacy-sensitive data...

RCE-Exploit-Tools

RCE Exploit Research Tools Project Overview This repository cont...

Towards a Systematic Taxonomy of Attacks against Space Infrastructures

Space infrastructures represent an emerging domain that is critical to the global economy and society. However, this domain is vulnerable to attacks. To enhance the resilience of this domain, we must understand the attacks that can be waged against it. The status quo is that there is no systemati...

Fake ChatGPT Atlas Browser Used in ClickFix Attack to Steal Passwords

Cybersecurity researchers have uncovered a critical ChatGPT Atlas browser attack, confirming the danger of the ongoing surge in the ClickFix threat...

Researchers Uncover ECScape Flaw in Amazon ECS Enabling Cross-Task Credential Theft

Cybersecurity researchers have demonstrated an "end-to-end privilege escalation chain" in Amazon Elastic Container Service ECS that could be exploited by an attacker to conduct lateral movement, access sensitive data, and seize control of the cloud environment. The attack technique has been...

AI-Generated Malicious npm Package Drains Solana Funds from 1,500+ Before Takedown

Cybersecurity researchers have flagged a malicious npm package that was generated using artificial intelligence AI and concealed a cryptocurrency wallet drainer. The package, @kodane/patch-manager, claims to offer "advanced license validation and registry optimization utilities for high-performan...

Researchers Expose Massive Online Fake Currency Operation in India

Cybersecurity researchers at CloudSEK's STRIKE team used facial recognition and GPS data to expose a massive, over $2…...

BRB, pausing for a "Sanctuary Moon" marathon

Welcome to this week's edition of the Threat Source newsletter. Yesterday, Cisco Talos debuted the first Humans of Talos episode, where I interviewed Hazel Burton, a face and voice you're probably familiar with. In our conversation, Hazel shared not just the story of how she found her way onto th...

eSIM Vulnerability in eUICC Cards Exposes Billions of IoT Devices to Malicious Attacks

Cybersecurity researchers have discovered a new hacking technique that exploits weaknesses in the eSIM technology used in modern smartphones, exposing users to severe risks. The issues impact the Kigen eUICC card. According to the Irish company's website, more than two billion SIMs in IoT devices...

New TokenBreak Attack Bypasses AI Moderation with Single-Character Text Changes

Cybersecurity researchers have discovered a novel attack technique called TokenBreak that can be used to bypass a large language model's LLM safety and content moderation guardrails with just a single character change. "The TokenBreak attack targets a text classification model's tokenization...

CrazyHunter Campaign Targets Taiwanese Critical Sectors

This blog entry details research on emerging ransomware group CrazyHunter, which has launched a sophisticated campaign aimed at Taiwan's essential services...

New AkiraBot Abuses OpenAI API to Spam Website Contact Forms

Cybersecurity researchers have identified a new spam campaign driven by 'AkiraBot,' an AI-powered bot that targets small business…...

New Morphing Meerkat Phishing Kit Mimics 114 Brands Using Victims' DNS Email Records

Cybersecurity researchers have shed light on a new phishing-as-a-service PhaaS platform that leverages the Domain Name System DNS mail exchange MX records to serve fake login pages that impersonate about 114 brands. DNS intelligence firm Infoblox is tracking the actor behind the PhaaS, the phishi...

Analyze Mobile Threats Faster: ANY.RUN Introduces Android OS to Its Interactive Sandbox

Disclosure: This article was provided by ANY.RUN. The information and analysis presented are based on their research and findings...

Ransomware threat mailed in letters to business owners

Business owners and CEOs across the United States received customized ransomware threats this month from the most unusual of places—letters in the mail. The letters, which were first reported by multiple cybersecurity researchers, claim to come from a ransomware group called BianLian. But since...

Malicious ML Models on Hugging Face Leverage Broken Pickle Format to Evade Detection

Cybersecurity researchers have uncovered two malicious machine learning ML models on Hugging Face that leveraged an unusual technique of "broken" pickle files to evade detection. "The pickle files extracted from the mentioned PyTorch archives revealed the malicious Python content at the beginning...

Despite Catastrophic Hacks, Ransomware Payments Dropped Dramatically Last Year

Ransomware gangs continued to wreak havoc in 2024, but new research shows that the amounts victims paid these cybercriminals fell by hundreds of millions of dollars...

New AI Jailbreak Method 'Bad Likert Judge' Boosts Attack Success Rates by Over 60%

Cybersecurity researchers have shed light on a new jailbreak technique that could be used to get past a large language model's LLM safety guardrails and produce potentially harmful or malicious responses. The multi-turn aka many-shot attack strategy has been codenamed Bad Likert Judge by Palo Alt...