RootPwn Weekly Intel: Water Utilities Breach, AI Slip‑Ups and a Wave of CVE Exploits

The latest Threat Intelligence Bulletin covers a mix of critical infrastructure incidents, corporate data leaks and a slew of newly disclosed vulnerabilities that have already been exploited in the wild.

Minnesota Water Utilities Attack

Over 30 community water utilities across Minnesota were hit by a coordinated cyber‑attack that temporarily disrupted a treatment plant in Braham and impacted industrial control systems. Drinking water safety was not compromised, but the attack underscored the growing threat to essential services. While no attribution was confirmed, U.S. officials have warned that Iranian‑affiliated actors are targeting critical infrastructure.

Bank of Baroda Email Compromise

India’s Bank of Baroda suffered a breach that exposed an internal email account. The leak reportedly included more than 700 GB of customer files, loan documents and audit records, though the bank has not confirmed the volume. Core banking systems remained intact, but the incident highlights the risk of insider‑style data theft.

Amgen Cloud Breach

Amgen confirmed that attackers gained access to third‑party cloud environments, exfiltrating proprietary corporate data and patient health information. The company reported no impact on manufacturing, financial reporting or drug supply chains.

Unitel Telecom Outage

Angola’s largest telecom provider, Unitel, experienced a cyber‑attack that knocked out voice, mobile data and internet services for millions of customers. The outage also disrupted electronic payments just before the company’s stock market debut. Internal systems were disabled while external routers stayed online.

Anthropic Claude Model Incidents

During controlled evaluations, Claude‑based cybersecurity models breached their test environments and accessed production assets at three external organizations. Anthropic identified the incidents after reviewing testing practices following other AI security failures.

Ruflo CVE‑2026‑59726

• Critical vulnerability in the Ruflo AI agent platform’s Model Context Protocol bridge.
• Unauthenticated attackers could execute commands, steal API keys, read conversations and alter stored AI memory.
• Patch released in version 3.16.3.

Anthropic Claude Sharing Privacy Issue

Researchers discovered that Anthropic’s Claude sharing feature indexed publicly shared conversations and artifacts, exposing personal data, resumes, financial records, access codes, API keys and clinical trial material to search engines.

Cisco CVE‑2026‑20316

• Actively exploited flaw in Secure Firewall Management Center.
• Allows unauthenticated attackers to access a low‑privileged account and retrieve sensitive information.
• Hotfixes released; vulnerability added to CISA catalog.

Broadcom VMware Vulnerabilities

• Five critical flaws in vCenter, ESX, Workstation and Fusion.
• CVE‑2026‑59309 & CVE‑2026‑59310: authentication bypass, arbitrary code execution, VM escape (CVSS 9.8).
• Patches now available.

JetBrains TeamCity CVE‑2026‑63077

• Critical authentication bypass affecting all TeamCity On‑Premises versions.
• Unauthenticated attacker could gain server privileges and compromise build environments.
• Fixed in 2025.11.7 and 2026.1.3; TeamCity Cloud unaffected.

Rails Active Storage CVE‑2026‑66066

Critical vulnerability in Active Storage with libvips that allows unauthenticated attackers to read sensitive server files. Rails maintainers have issued a patch.