CVE-2026-34760 vLLM: Downmix Implementation Differences as Attack Vectors Against Audio AI Models

vLLM is an inference and serving engine for large language models LLMs. From version 0.5.5 to before version 0.18.0, Librosa defaults to using numpy.mean for mono downmixing tomono, while the international standard ITU-R BS.775-4 specifies a weighted downmixing algorithm. This discrepancy results...

Threat actor abuse of AI accelerates from tool to cyberattack surface

For the last year, one word has represented the conversation living at the intersection of AI and cybersecurity: speed. Speed matters, but it’s not the most important shift we are observing across the threat landscape today. Now, threat actors from nation states to cybercrime groups are embedding...

Threat actor abuse of AI accelerates from tool to cyberattack surface

For the last year, one word has represented the conversation living at the intersection of AI and cybersecurity: speed. Speed matters, but it’s not the most important shift we are observing across the threat landscape today. Now, threat actors from nation states to cybercrime groups are embedding...

New Whitepaper: Stealthy BPFDoor Variants are a Needle That Looks Like Hay

Executive Overview Advanced persistent threats APTs are constantly and consistently changing tactics as network defenders plug holes in defenses. Static indicators of compromise IoCs for the BPFDoor have been widely deployed, forcing threat actors to get creative in their use of this particular...

The State of Trusted Open Source Report

In December 2025, we shared the first-ever The State of Trusted Open Source report, featuring insights from our product data and customer base on open source consumption across our catalog of container image projects, versions, images, language libraries, and builds. These insights shed light on...

Vanna 安全漏洞

Vanna is a personalized AI SQL proxy from Vanna Corporation. Versions of vanna 2.0.2 and earlier contained security vulnerabilities. These vulnerabilities were caused by overly lax cross-domain policies implemented in the FastAPI/Flask Server component, which could lead to remote attacks...

Synthetic Trust Attacks: Modeling How Generative AI Manipulates Human Decisions in Social Engineering Fraud

Imagine receiving a video call from your CFO, surrounded by colleagues, asking you to urgently authorise a confidential transfer. You comply. Every person on that call was fake, and you just lost $25 million. This is not a hypothetical. It happened in Hong Kong in January 2024, and it is becoming...

Wikipedia’s AI agent row likely just the beginning of the bot-ocalypse

The Internet is filled with people who insist on being right. In the past, at least they could be reasonably sure that they were arguing with other humans. Those days are gone, apparently. Wikipedia just had to ban an AI that was making edits on its own. Apparently, the AI took it personally. The...

Exploit for Path Traversal in Apache Http_Server

🚀 AutoReconAI – Intelligent Vulnerability Scanner & Exploit Re...

Why we’re still not doing April Fools’ Day

People lost an estimated $442 billion to scams last year worldwide, according to the Global Anti-Scam Alliance. The scale of that is hard to picture, but people's day-to-day scam experience is easier to recognize: Our research found that 44% of people say they encounter mobile scams every single...

Asking AI for personal advice is a bad idea, Stanford study shows

Stanford computer scientists just proved what therapists already suspected: AI chatbots will agree with almost anything you say to keep you happy. The researchers caught these systems validating dangerous decisions just to maintain user engagement. That's a worrying development, especially given...

Wordfence Bug Bounty Program Monthly Report – February 2026

Last month in February 2026, the Wordfence Bug Bounty Program received 1078 vulnerability submissions from our growing community of security researchers working to improve the overall security posture of the WordPress ecosystem. These submissions are reviewed, triaged, and processed by the...

CVE-2026-32243 Discourse: Stored XSS in discourse-ai shared conversations onebox

Discourse is an open-source discussion platform. From versions 2026.1.0-latest to before 2026.1.3, 2026.2.0-latest to before 2026.2.2, and 2026.3.0-latest to before 2026.3.0, an attacker with the ability to create shared AI conversations could inject arbitrary HTML and JavaScript via crafted...

Applying security fundamentals to AI: Practical advice for CISOs

What to know about the era of AI The first thing to know is that AI isn’t magic The best way to think about how to effectively use and secure a modern AI system is to imagine it like a very new, very junior person. It’s very smart and eager to help but can also be extremely unintelligent. Like a...

Applying security fundamentals to AI: Practical advice for CISOs

What to know about the era of AI The first thing to know is that AI isn’t magic The best way to think about how to effectively use and secure a modern AI system is to imagine it like a very new, very junior person. It’s very smart and eager to help but can also be extremely unintelligent. Like a...

The Real Risk of Vibecoding

This blog looks at how AI‑driven vibecoding speeds up software development while increasing security risk by outpacing traditional review and ownership. It explains why security needs to move earlier and be built into modern development workflows...

Architecting Secure AI Agents: Perspectives on System-Level Defenses against Indirect Prompt Injection Attacks

AI agents, predominantly powered by large language models LLMs, are vulnerable to indirect prompt injection, in which malicious instructions embedded in untrusted data can trigger dangerous agent actions. This position paper discusses our vision for system-level defenses against indirect prompt...

24/7 Payments for 24/7 Agents: The Case for Crypto in the Machine Economy

Crypto enables 24/7 payments for AI agents, replacing fiat limits with scalable machine-to-machine transactions and powering the emerging machine economy...

⚡ Weekly Recap: Telecom Sleeper Cells, LLM Jailbreaks, Apple Forces U.K. Age Checks and More

Some weeks are loud. This one was quieter but not in a good way. Long-running operations are finally hitting courtrooms, old attack methods are showing up in new places, and research that stopped being theoretical right around the time defenders stopped paying attention. There's a bit of everythi...

CISO Spotlight: Dimitris Georgiou on Building Security that Serves People First

Dimitris Georgiou has been a self-professed computer geek since the early 80s. At university, he studied the convergence of educational technology with computer science as part of his psychology MA – finding, to his disbelief, that systems were perilously insecure. Since then, he’s always worked ...