Five vendors, one bug: the trust gap spreading through Model Context Protocol
A single security researcher just got five different organizations, including Google and JPMorgan Chase, to confirm the same mistake in servers they built independently. The punchline is that the mistake isn’t exotic: it’s classic server-side request forgery in AI agents connected through the Model Context Protocol (MCP), the open standard that lets AI applications call […]
GitLab Patches a Second Critical AI Gateway RCE, and This One Hits Self-Hosted Installs
GitLab shipped an emergency patch release on Friday for a critical remote code execution flaw in the GitLab AI Gateway, the standalone service that backs the company’s Duo AI features. The bug, tracked as CVE-2026-90970, scores 9.9 out of 10 on the CVSS scale. If your AI Gateway runs in your own infrastructure, you need […]
Meta’s Muse AI assistant had a zero-day that turned it into the perfect Mac backdoor
Muse, Meta’s new AI assistant for macOS, shipped with a zero-day that let any locally running app or terminal command seize complete control of the agent: read its session token, inject its own prompts, and quietly exfiltrate whatever the assistant could reach, which given its permissions is a lot. Mac security researcher Patrick Wardle found […]
MetaMask pulls thousands of Ethereum validators offline after infrastructure breach
MetaMask disclosed an ongoing security incident affecting part of its infrastructure on September 30, and the response is unusual in scale: the company is deliberately exiting thousands of Ethereum validators it operates through its staking service, accepting lost rewards and possible penalties to eliminate a risk it won’t fully describe. The company says it has […]
NetScaler zero-days: check for compromise before you patch
Two zero-days, already exploited, patch available Citrix NetScaler appliances are under active attack through two critical vulnerabilities that were exploited in the wild before any patch existed. Citrix disclosed eight vulnerabilities on September 27 in bulletin CTX697096. The two headline flaws, CVE-2026-88771 and CVE-2026-88772, both score 9.5 on the CVSS v4 scale and both lead […]
80,000 organizations had AI logins stolen, and session cookies are the real problem
SOCRadar’s AI Identity Exposure Report, covered by BleepingComputer, puts a number on something security teams have been quietly dreading: infostealer logs now contain employee logins for AI platforms at serious scale. More than a million records tied to AI services span over 80,000 corporate domains. The researchers narrowed that set to 482 major enterprises and […]
Serving a static site over Tor from your own server
A hidden service in ten lines of config The “dark web” sounds exotic, but the mechanics of putting a site on it are closer to a reverse proxy config than anything exotic. David Alvarez Rosa published a walkthrough of self-hosting a personal site as a Tor hidden service, and the whole setup is small enough […]
Jury finds Facebook liable for deceiving users in the Cambridge Analytica case
A jury finally said it out loud Eight years after the Cambridge Analytica story broke, a New Mexico state court jury has found Facebook liable for deceiving users about privacy protections on the platform. The verdict came down on September 25 in Santa Fe, after a two-week trial in the First Judicial District Court. This […]
CISA flags four actively exploited flaws in WSO2, Adobe Commerce, SharePoint and MikroTik
Four more flaws join the exploit catalog The U.S. Cybersecurity and Infrastructure Security Agency added four vulnerabilities to its Known Exploited Vulnerabilities catalog this week, and all four carry evidence of attacks in the wild rather than a mere proof of concept. Two entries landed on September 24: a critical JWT authentication bypass in WSO2 […]
Researchers forge 1024-bit RSA signatures without factoring the key
A team from UC San Diego and Inria has done something cryptographers long assumed was impossible or at least pointless: forged 1024-bit RSA signatures without ever factoring the modulus. The paper, posted September 20 by Laura Shea, Miro Haller, Adam Suhl, Nadia Heninger, and Emmanuel Thomé, ran a 2007 algorithm at real scale and found […]
SourceHut build logs hid a wormable XSS that could take over accounts
A vulnerability researcher known as Arusekk published a writeup on September 23 describing CVE-2026-92973, a cross-site scripting flaw in the ansi2html library that let anyone who could inject text into a SourceHut build log take over the accounts of people who viewed it. The bug had been sitting in the wild for close to four […]
Why Trail of Bits calls SAML a fractal of bad design
A retrospective on the protocol nobody can replace Security firm Trail of Bits published a historical analysis of SAML this week, and the title is not subtle: they call it a fractal of bad design, straightforward on the surface but built on what they describe as a foundation of sand, bone dust, and ash. The […]