AI News Today

A critical flaw in GitLab's AI Gateway could let a logged-in user with Duo Agent Platform access run commands on the gateway under certain conditions, GitLab said in an advisory. The gateway is the service that connects a GitLab instance to AI models, and only organizations that host their own gateway need to act. The flaw is fixed in gateway versions 19.2.4, 19.3.2, and 19.4.1. The flaw

OpenAI has parted ways with three members of its safety team after they leaked private information in violation of company policies, The Wall Street Journal reported. "We have parted ways with three individuals for violating our policies on accessing and handling sensitive company information," a spokesperson for the company was quoted as saying. "Our investigation confirmed that these

This week, the useful words are boring ones: inspect, cache, compile, store, trust. Each sounds harmless. Each can become an attack path when a system does a little more than people expect. A model check can run code. A cache can mix up requests. A public secret can stay useful for years. That is the lesson running through the list. Attackers do not always need a brilliant new trick. They can

OpenAI on Wednesday said it identified and disrupted a coordinated distillation campaign that was designed to illicitly extract protected reasoning from its artificial intelligence (AI) models. A "core cluster of the activity," going back to the first week of July, has been attributed to individuals associated with Moonshot AI, a Chinese AI company based in Beijing. It did not cite any

Google on Wednesday announced its latest frontier artificial intelligence (AI) model, Gemini 4 Argon, that it said is being rolled out to a set of trusted cyber defenders through its Fairwind Program. "It delivers frontier performance in complex workflows across real-world software engineering, enterprise knowledge work like legal and finance, and cybersecurity defense," Koray Kavukcuoglu,

MetaMask on Thursday said it's responding to what it described as an "ongoing security incident" impacting part of its infrastructure. "We are actively addressing and remediating the issue internally, in coordination with external partners and security advisors," the software cryptocurrency wallet maker said. "At this time, we have identified no immediate threat to MetaMask wallets." MetaMask

Microsoft has warned of phishing campaigns distributing an installer for the MSP360 Remote Monitoring and Management (RMM) software under the guise of meeting invitations, PDF-themed lures, software update prompts, and other social-engineering content. "Once executed, the legitimate MSP360 installer, distributed under a deceptive file name established remote management access on affected

Threat actors are abusing ChatGPT Custom GPTs to disguise them as legitimate product offerings and direct unsuspecting victims to malicious sites that employ ClickFix lures to deliver malware. Huntress, which observed the activity in late September 2026, said it marks the abuse of yet another feature in trusted artificial intelligence (AI) platforms. Prior campaigns have weaponized shared

AI coding agents asked to share screenshots of code changes for review have put internal company images in public GitHub repositories, security company Glow said. Its researchers found more than 13,000 internal images from developers at over 300 organizations, including customer billing records and screens of features not yet released. In most cases, they sat under developers' personal accounts

OpenAI on Monday shelved plans to release GPT-6.1 Astra, a next-generation artificial intelligence (AI) model that was planned for an October launch, after it failed internal safety and alignment audits. The development was first reported by The Wall Street Journal. The move "marks a rare case of a major AI developer ditching a new release because of safety concerns," the news publication said.

OpenAI said it has made the decision to pause training of its most powerful models after one of its agents during reinforcement learning (RL) training contacted an external chatbot by exploiting a loophole in its internet-access restrictions. "An agent attempting to complete a search-based training task queried a public chatbot service through a gap in our internet-access restrictions:

What is IAM for AI agents? AI agents authenticate, invoke tools, and act across enterprise systems with delegated authority. IAM for AI Agents is the identity-control architecture that governs those actors. This guide covers the limits of conventional provisioning, the components that matter, how to evaluate framework choices, and what runtime evidence proves an agent behaved as intended.

RatHat's operators build and publish the Android banking trojan and control infected phones from a web console, according to security company Cleafy. Cleafy has traced nearly 100 deployments of that console since April 2026. It said this fits a malware-as-a-service model, in which each customer runs a separate copy. The console stores what the malware collects from each phone,

A domain used as harmless placeholder text showed up in roughly 1,700 repositories. Then somebody registered it and started serving malicious lures. That is the kind of week this was: forgotten assumptions turning into live attack surface. Elsewhere, weak service accounts, old bugs, exposed systems, phishing kits, and strangely easy exploit paths kept doing useful work for attackers. Nothing

AI agents are moving into production faster than security teams can govern them. They are connecting to apps, handling data, calling APIs, and acting across business systems—often without the same controls applied to human users. According to Okta’s Global CISO Insights 2026 report, only 47% of CISOs are confident they can identify every AI agent in their environment. Even among those who feel

Cybersecurity researchers have disclosed details of a new botnet malware called Carbonato that's targeting exposed Docker daemons to deploy an open-source artificial intelligence (AI) agent framework called Hermes Agent. "The implant installs the framework unchanged, then overwrites its SOUL.md persona file," ThreatDown said. "The 39-line prompt directs it to execute tasks received through

The way we talk about AI agents is shifting, and the way we implement them requires an even more fundamental shift. While earlier discourse focused on how quickly organizations could stand up agents and how much productivity they could promise, a string of recent incidents, including a widely discussed intrusion at Hugging Face during an evaluation of OpenAI agents, has spurred organizations to

Security leaders keep debating whether AI will produce an entirely new class of cyberattack. The nearer change is quieter and already visible: AI has made a failed attack cheap to retry. The routine version looks like this. An attacker lands on a low-privilege cloud account, and the first try at privilege escalation goes nowhere. That dead end used to cost hours of documentation reading,

The U.S. Cybersecurity and Infrastructure Security Agency (CISA), on Thursday, added two critical security flaws impacting WSO2 and Adobe Commerce and Magento to its Known Exploited Vulnerabilities (KEV) catalog, based on evidence of active exploitation. The vulnerabilities are listed below - CVE-2026-5430 (CVS score: 9.8) - A path traversal vulnerability in WSO2 API Control Plane,

This week, the dangerous stuff keeps arriving dressed as something boring. An update. A login box. A search answer. A coding tool. A link you have clicked a hundred times before. That is the thread running through the pile. Trusted paths get poisoned. Old bugs find new jobs. AI tools leak more than expected. Fake prompts look real enough. And some attacks barely need an exploit at all — just

AI coding agents are changing how quickly developers can build and ship software as well as how quickly credentials can become exposed. According to GitGuardian’s 2026 State of Secrets Sprawl Report, commits identified as AI-assisted are leaking secrets at approximately twice the rate of human-written ones. Most of the fastest-growing categories of leaked credentials are now connected to AI

An AI agent on an internal OpenAI research task bypassed access controls on an Australian government Medicare statistics portal in June, Prime Minister Anthony Albanese said. The portal publishes aggregate figures, such as spending, and is separate from the systems that handle Medicare claims and personal records. The agent reached files on it that were not public, but no personal

A Windows malware called CLOSEDQUORUM is built to take orders from a vote of up to four AI models instead of an attacker's server, Cisco Talos said on September 22. The models can choose to steal Windows credentials, saved browser passwords, and crypto wallet data. Talos has not seen this setup work from start to finish, and the public version of the malware does not work as it is.

Autonomous security agents are getting good at finding bugs. Nobody has a good way to measure how good. Point one at a realistic target and what comes back is a report the agent wrote about itself: confident prose, a list of findings, and no way to tell which of them happened. Someone with a security background then sits down and checks every claim against the target. Which findings are real,

Anthropic and OpenAI on Tuesday announced new models, with both artificial intelligence (AI) companies noting that they are continuing to invest in improving alignment to combat risky behavior. Opus 5.5, per Anthropic, is a "major step up from Opus 5," and "achieves the best scores of any model to date on our automated behavioral audit, our alignment suite that tests Claude across thousands

The cyber extortion group known as ShinyHunters on Tuesday claimed it had breached the U.S. Federal Bureau of Investigation and stolen data belonging to current and former employees at the agency. "We have compromised the FBI. We hold very sensitive data on almost ALL FBI Agents and individuals who filed an application with the FBI for a job," the group said in a statement posted on their dark

Microsoft on Tuesday announced the takedown of the EvilTokens device code phishing service that it said used artificial intelligence (AI) "at every step of the attack chain." The action, carried out with authorization from the U.S. District Court for the Eastern District of Virginia, involved the efforts of Health-ISAC, alongside Cloudflare, Coinbase, OpenAI, Railway, SpyCloud, The Shadowserver

A critical vulnerability in Bifrost, an open-source AI gateway that routes requests to more than 20 LLM providers, allows an unauthenticated attacker to run arbitrary commands on the gateway server with a single HTTP request. The flaw, tracked as CVE-2026-90898 (CVSS score: 9.8), affects all versions of the Bifrost HTTP transport before 2.1.0 when management authentication is

Security teams have spent decades asking whether an identity has too much access. AI agents raise a harder question: how can we determine which paths an autonomous system can discover, given the access it already has? A person may try several ways to complete a task. A deterministic application follows the flow its developer wrote. But an AI agent is relentless in its pursuit of done. In May

Malware already running on a Mac can quietly take over Meta's Muse assistant and use the broad access its owner granted the app, security researcher Patrick Wardle has shown in a proof-of-concept released on September 21. It works by changing a hidden setting so that when the user taps the microphone and dictates a prompt, the words go to the attacker instead of Meta. The flaw is in

A browser. A plugin. A package. A login screen. Normal stuff. That is basically the problem this week. The trouble keeps showing up inside things people already trust: code that takes a bad turn, old payloads coming back, exposed systems, weak checks, fake fixes, and attack paths that look almost too easy. Even the research side is getting messy, with more findings, more automation, and not

Three researchers at the security firm Hacktron used Anthropic's Claude Opus 5 to chain two flaws and take over the ChatGPT and Codex accounts of several OpenAI employees, then reach an internal OpenAI code repository. The chain began with a bug in the software that runs OpenAI's public help forum and moved through a weakness in OpenAI's own login system. This was security research,

Google's Gemini model has become the latest artificial intelligence (AI) system to access the internet and break into other companies during a cybersecurity evaluation. The development was first reported by The Wall Street Journal. The incidents occurred in May 2026 as part of a test run conducted by Israeli company Irregular. The evaluation partner was also involved in similar hacks disclosed

WordPress today released patches to fix a new set of vulnerabilities in its core software, one of which could allow a crafted web link, opened by a logged-in administrator, to install a theme from the official WordPress.org directory without anyone clicking Install. The security firm pwn.ai, whose researchers reported the flaw, calls the attack chain Click2Shell. On its own the flaw only

Microsoft has released fixes for a maximum-severity security flaw in Azure AI Foundry that could be exploited to achieve privilege escalation. No customer action is required. The vulnerability, tracked as CVE-2026-85889, carries a CVSS score of 10.0. "Missing authentication for critical function in Azure AI Foundry allows an unauthorized attacker to elevate privileges over a network,"

A flaw in four widely used AI coding agents lets someone who controls a plugin's code repository swap the plugin an agent installs for a malicious one, even when the agent locked that plugin to a specific reviewed version, security firm Air Security said on Thursday. The firm said Anthropic has patched the flaw in Claude Code 2.1.179 and OpenAI in Codex 0.146.0, that GitHub Copilot has no

A financially motivated threat actor has been linked to the development and distribution of a JavaScript (JS)-based information stealer known as PhantomRaven via the npm package registry. "The developer likely wrote the malware using a large language model (LLM), an assessment made with high confidence based on verbose comments, placeholder code, and statistical token-analysis patterns,"

Cybersecurity researchers have flagged a new Android malware called RatHat that's assessed to be operated by China-based threat actors and features an artificial intelligence (AI)-powered system to navigate and control compromised devices. "Distributed primarily via targeted smishing (SMS/text phishing) and malvertising campaigns leading to deceptive third-party download portals, RatHat uses

Attackers keep finding new keys. The funny part is that defenders keep inventing where to store them. This week, those keys sit in AI tools, exposed services, old bugs, weak logins, and software sold like a monthly subscription. Some attacks use new tricks. Others just reuse what was already lying around. Both work often enough. So the threat landscape is not getting cleaner. It is just

A new CVE drops. Your scanner finds it. The severity score looks ugly. But that still does not answer the question that matters: Can it actually be exploited in your environment? Mythos-class AI is compressing the time between disclosure and working exploitation, while many security programs still validate risk on weekly or quarterly cycles. The dangerous gap is no longer just technical. It is