MAL-2026-3679 Malicious code in @2oolkit/hyperliquid-cli (npm)

--- -= Per source details. Do not edit below this line.=- Source: amazon-inspector 1c3af30011dcf54950f270463028270d732fce20b5cd5da44342a0748922e6df The package is advertised as a neutral CLI/MCP wrapper for Hyperliquid, but its distributed code silently routes value from the installer to an...

Malicious code in dcchbot (PyPI)

--- -= Per source details. Do not edit below this line.=- Source: amazon-inspector df79831d1b486c8ca704295b410cec7b66be85aa87c3244d97ff1e87f643183a The package performs multiple installer-hostile behaviors. 1 dcchbot/init.py auto-invokes run on import, which triggers interactive input prompts and...

ContextualJailbreak: Evolutionary Red-Teaming Via Simulated Conversational Priming

Large language models LLMs remain vulnerable to jailbreak attacks that bypass safety alignment and elicit harmful responses. A growing body of work shows that contextual priming, where earlier turns covertly bias later replies, constitutes a powerful attack surface, with hand-crafted multi-turn...

Owner-Harm: A Missing Threat Model for AI Agent Safety

Existing AI agent safety benchmarks focus on generic criminal harm cybercrime, harassment, weapon synthesis, leaving a systematic blind spot for a distinct and commercially consequential threat category: agents harming their own deployers. Real-world incidents illustrate the gap: Slack AI...

Incident response for AI: Same fire, different fuel

In this article 1. The fundamentals still hold 2. Where AI changes the equation 3. Closing the gaps in telemetry, tooling, and response 4. The human dimension 5. Looking ahead When a traditional security incident hits, responders replay what happened. They trace a known code path, find the defect...

Risk-Adjusted Harm Scoring for Automated Red Teaming for LLMs in Financial Services

The rapid adoption of large language models LLMs in financial services introduces new operational, regulatory, and security risks. Yet most red-teaming benchmarks remain domain-agnostic and fail to capture failure modes specific to regulated BFSI settings, where harmful behavior can be elicited...

Systematic Scaling Analysis of Jailbreak Attacks in Large Language Models

Large language models remain vulnerable to jailbreak attacks, yet we still lack a systematic understanding of how jailbreak success scales with attacker effort across methods, model families, and harm types. We initiate a scaling-law framework for jailbreaks by treating each attack as a...

Instagram flagged explicit messages to minors in 2018. Image-blurring arrived six years later

Meta took six years to blur explicit images on Instagram, even though internal emails show executives were aware in 2018 that minors were receiving them, according to newly unsealed court documents. In a deposition given last year, Adam Mosseri now the head of Instagram discusses an email thread...

Deepfake ‘Nudify’ Technology Is Getting Darker—and More Dangerous

Sexual deepfakes continue to get more sophisticated, capable, easy to access, and perilous for millions of women who are abused with the technology...

Enshittification is ruining everything online (Lock and Code S07E01)

This week on the Lock and Code podcast … There's a bizarre thing happening online right now where everything is getting worse. Your Google results have become so bad that you’ve likely typed what you’re looking for, plus the word “Reddit,” so you can find discussion from actual humans. If you...

The Most Dangerous People on the Internet in 2025

From Donald Trump to DOGE to Chinese hackers, this year the internet’s chaos caused outsize real-world harm...

AIs Exploiting Smart Contracts

I have long maintained that smart contracts are a dumb idea: that a human process is actually a security feature. Here's some interesting research on training AIs to automatically exploit smart contracts: AI models are increasingly good at cyber tasks, as we've written about before. But what is t...

PT-2025-48405

opening the vault on a 3-year-old vulnerability: CVE-2022-35420 back in 2022, I got bored and decided to hunt for a zero-day instead of writing the red-team report. I found an unauthenticated admin takeover in a SCADA system used to control real-world physical machinery. I waited until now to...

TASO: Jailbreak LLMs Via Alternative Template and Suffix Optimization

Many recent studies showed that LLMs are vulnerable to jailbreak attacks, where an attacker can perturb the input of an LLM to induce it to generate an output for a harmful question. In general, existing jailbreak techniques either optimize a semantic template intended to induce the LLM to produc...

Scaling Patterns in Adversarial Alignment: Evidence from Multi-LLM Jailbreak Experiments

Large language models LLMs increasingly operate in multi-agent and safety-critical settings, raising open questions about how their vulnerabilities scale when models interact adversarially. This study examines whether larger models can systematically jailbreak smaller ones - eliciting harmful or...

Malicious code in surya-bubursumsum5-sluey (npm)

--- -= Per source details. Do not edit below this line.=- Source: amazon-inspector 2a3c2ac2da93b636fcc97e37abce1fe9b443d5b7a79e0072ce0d27ba62414d89 This package appears to be part of the tea.xyz token reward campaign that flooded npm. These packages typically contain autopublish scripts auto.js,...

AI-driven scams are preying on Gen Z’s digital lives​

Gone are the days when extortion was only the plot line of crime dramas—today, these threatening tactics target anyone with a smartphone. As AI makes fake voices and videos sound and look real, high-pressure plays like sextortion, deepfakes, and virtual kidnapping feel more believable than ever...

Pattern Enhanced Multi-Turn Jailbreaking: Exploiting Structural Vulnerabilities in Large Language Models

Large language models LLMs remain vulnerable to multi-turn jailbreaking attacks that exploit conversational context to bypass safety constraints gradually. These attacks target different harm categories like malware generation, harassment, or fraud through distinct conversational approaches...

Age verification and parental controls coming to ChatGPT to protect teens

OpenAI is going to try and predict the ages of its users to protect them better, as stories of AI-induced harms in children mount. The company, which runs the popular ChatGPT AI, is working on what it calls a long-term system to determine whether users are over 18. If it can't verify that a user ...

GOP Cries Censorship Over Spam Filters That Work

The chairman of the Federal Trade Commission FTC last week sent a letter to Google's CEO demanding to know why Gmail was blocking messages from Republican senders while allegedly failing to block similar missives supporting Democrats. The letter followed media reports accusing Gmail of...