A seller on a restricted criminal forum is marketing an autonomous hacking tool that runs on stolen OpenAI, Anthropic and Google API keys, so the key's owner pays for the scans. Bitsight's threat intelligence team tracked the shift across thousands of posts on 16 underground forums from February through August 2026. Over that window, mentions of AI agents, indirect prompt injection and the Model Context Protocol (MCP) rose 67 percent. READ MORE...
A Chinese APT has been exploiting a BeyondTrust vulnerability as a zero-day since August as part of its state-sponsored espionage operations. The exploited flaw, tracked as CVE-2026-102764 (CVSS score of 9.3), is described as an improper validation of message origins that could allow attackers to perform file operations on Remote Support (RS) appliances. According to BeyondTrust, all RS versions up to and including 26.2.3 are affected. READ MORE...
Hackers could break into businesses by going after internet-exposed AI tools that often contain serious vulnerabilities, the internet monitoring firm Censys warned in a report published on Wednesday. The growing complexity of the AI ecosystem - in which public-facing services rely on a sprawling web of back-end products, many of them neglected by security teams - creates additional security risks, researchers said. READ MORE...
Arch Linux users who mistyped yay or paru while installing the package helpers from the Arch User Repository (AUR) may have given a stranger root access to their machine. AUR is a popular community package source for Arch Linux and related distributions, hosting user-submitted build scripts for proprietary programs, beta builds and older software versions that the official repositories do not carry. READ MORE...
OpenAI has disclosed two incidents in which internal AI models, asked only to look up public figures, instead broke into the live systems hosting that data. The reports, published on October 15, are the latest in a run of misalignment disclosures OpenAI started after its models compromised Hugging Face in July. OpenAI has since disclosed several other incidents, including models probing government websites for vulnerabilities and searching GitHub for leaked API keys during training. READ MORE...
Election influence campaigns, particularly ones playing out over social media, are a dime a dozen. Such campaigns - sometimes devised by foreign powers, other times by domestic groups - are often designed to influence the U.S. electorate. Fake accounts and online personas are created, networks of propaganda are disseminated and the campaign seeks to nudge American voters into thinking a certain way. Mostly, these efforts are known for getting little to no traction with U.S. audiences. READ MORE...
Despite the hype around AI agents interacting with websites, few sites bother to supply dedicated reading material for visiting bots. And occasionally, the message directed at visiting bots is an attempt at prompt injection. While about 10 percent of sites use robots.txt rules to restrict AI crawlers, only about three to 10 percent of sites in separate surveys provide agent-oriented content in the form of llms.txt, a proposed format for exposing website information to AI systems. READ MORE...
Researchers say that mercenary hackers laced ads with zero-click exploits that installed sophisticated malware on the devices of targeted individuals when they did nothing other than load a website or open an installed app. The finding, published Thursday by the University of Toronto's Citizen Lab, said the incident is likely to be the first confirmation of zero-click mercenary spyware delivered through targeted ads. READ MORE...