Latest cybersecurity reporting from selected sources.

Also, 23andMe Breach Settlement, GitHub AI Flaw, Ubiquiti Patches Critical BugsThis week, hidden breaches, researchers exposed a GitHub AI agent flaw, an Ubiquiti critical flaw, a ColdFusion flaw, a growing Chinese ORB, U.K. government FortiBleed exposure, 23andMe victims get $46.75 million, 3.8 million affected by Medtronic breach and Cerner's 2025 victim count went up.

Researchers Urge Organizations to Rotate Credentials and Review Audit LogsThreat actor 888 claims it stole Accenture source code and cloud credentials, prompting researchers to warn that any exposed Azure tokens, encryption keys or DevOps secrets could enable follow-on intrusions and downstream supply-chain attacks even as Accenture says operations remain unaffected.

Also: Zcash Patches Flaw, $32M Humanity Protocol HackThis week, a key player in a $97M laundering scheme got prison time, Humanity Protocol suffered $32M in losses, Zcash patched a flaw, the EU targeted crypto platforms tied to Russia, authorities froze $3.8M in illicit funds and researchers exposed a Trezor chip weakness.

AI Dependency Attack Reportedly Exposes Data and Source CodeA LiteLLM supply-chain compromise enabled attackers to harvest credentials and access internal environments at scale at Mercor. The firm was the first to confirm a LiteLLM breach, and researchers are warning about growing AI system exposure and limited visibility.

Database Misconfiguration Exposed 1.5 million API TokensA misconfigured database at Moltbook, the viral social network for AI agents, exposed 1.5 million API authentication tokens, 35,000 email addresses and private messages. Security researchers discovered unauthenticated read and write access to all platform data within days of launch.

Also, SmarterMail Flaw, Nike Breach Probe, Empire Market Co-Creator Pleads GuiltyThis week, researchers exposed an Android RAT abusing Hugging Face. Attackers exploited a SmarterMail flaw. Automakers raised cyber spending. CISA flagged a VMware bug. Microsoft patched Office. An Empire Market co-creator pleaded guilty. Nike probed a breach.

Patches Issued for MongoBleed as Ransomware Groups Target Flaw to Steal DataTens of thousands of internet-exposed MongoDB databases are at risk as attackers actively target a critical vulnerability in the software to steal sensitive data, with ransomware groups having joined the fray, researchers warn. MongoDB has issued patches and mitigation advice.

Unpatched Flaw in Open-Source Gogs Service Facilitates Remote Code ExecutionAn attacker has been exploiting a zero-day vulnerability in Gogs, an open-source and popular Git service that allows for self-hosting, warned researchers. At least 700 internet-exposed servers running Gogs shows signs of being infected with command-and-control malware; no patch is yet available.

AppOmni Finds Now Assist Agents Could Trigger Unauthorized ActionsServiceNow's Now Assist agents could be manipulated through second-order prompt injection, enabling unauthorized record changes and data exposure despite protections, shows new research from AppOmni. The issue stemmed from default configurations that allow agents to invoke each other.

Russian Intel Hackers Flexible in Face of DetectionRussia-linked threat group COLDRIVER rapidly replaced its exposed malware with a stealthier PowerShell variant, using fake CAPTCHA prompts and cryptographic key-splitting to evade detection and escalate surveillance on NGOs, dissidents and policy experts, according to new research.

Default credentials, weak passwords, misconfigurations and a variety of other security shortcomings are exposing millions of medical devices and their data on the internet, said Soufian El Yadmani, CEO and co-founder of Modat, who shared recent research findings.

Agentforce Agentic AI Tool Was Exposed to Indirect Prompt Injection AttacksSalesforce has patched a vulnerability involving its Agentforce agentic artificial intelligence tool, discovered by researchers, that attackers could have exploited using an indirect prompt injection attack to steal sensitive customer data and leads being stored in the CRM system.

Default credentials, weak passwords, misconfigurations and a variety of other security shortcomings are exposing millions of medical devices and their data on the internet, said Soufian El Yadmani, CEO and co-founder of Modat, who shared recent research findings.

'Widespread Data Theft Campaign' Compromised Many Drift OAuth Tokens, Warn ExpertsThreat researchers report that "a widespread data theft campaign" traces to attackers stealing OAuth access tokens for applications integrated with Salesloft's AI chatbot Drift, then exfiltrating data. Victims include Salesforce customer Zscaler. Google Workspace instances were also breached.

'Heightened Threat Environment' Faces Critical Infrastructure, US Government WarnsMany types of commonly used types of industrial control systems continue to be deployed in a manner that leaves them publicly exposed to the internet, often by U.S.-based critical infrastructure operators, in what amounts to a preventable security risk, researchers warn.

Thousands of MCP Servers Leave AI Apps Open to Attack SurfacesHundreds of Model Context Protocol servers designed to help AI tools access private data are insecurely exposed online, say BackSlash Security researchers. Weak configurations leave systems vulnerable to data leaks and remote code execution attacks.

Researcher Says Firm Failed to Secure Sensitive Health Data From Survey FormsAn unsecured database containing 2 terabytes of data allegedly exposed more than 1.6 million clinical research records to the internet, including sensitive personal and medical information of patients, said the security researcher who discovered the lapse. Why does this keep happening?

Many Charging Cable Interfaces Have Exposed SSH and HTTP Ports, Researchers WarnResearchers demonstrated that multiple brands of EV charging stations have vulnerabilities due to manufacturers often leaving open and unsecured SSH and HTTP ports. The risks of these vulnerabilities range from an expanded attack surface to a launching pad for assaults on the power grid.

Researchers Say Illegal Access to Private AI models Can Enable Cross-Tenant AttacksSecurity researchers have discovered two critical vulnerabilities in the Hugging Face AI platform that exposed potential gaps for attackers seeking unauthorized access and manipulation of customer data and models. The risks highlight the security concerns about AI-as-a-service offerings.

Researchers Find Unprotected Unitronics DevicesHere's one reason why Iranian state hackers may have been able to target Israeli-made pressure-monitoring controllers used by American water systems: Nearly 150 of the controllers are exposed to the internet - and some still use the default password 1111.