Massive Data Breach Hits Navia Benefit Solutions
A US-based employee benefits administrator, Navia Benefit Solutions, has disclosed a data breach affecting over 2.6 million individuals. Unauthorized access and potential data exfiltration occurred between December 22, 2025, and January 15, 2026. Exposed information includes personal, health, and benefits data.
“This incident underscores the persistent threat to sensitive HR and benefits systems,” said a cybersecurity analyst at Recorded Future. “Organizations must prioritize network segmentation and multi-factor authentication to limit blast radius.”
Aura Breach: 900,000 Records Exposed After Phone Phishing
Identity protection firm Aura suffered a breach after a phone phishing attack allowed an intruder to access an employee account and a marketing platform. The attacker acquired approximately 900,000 records, primarily names and emails. Crucially, Aura’s core systems and identity protection services remained uncompromised.
“Phone-based social engineering remains a low-tech but highly effective vector,” noted a senior threat researcher at Malwarebytes. “Companies must train staff to verify identity via out-of-band channels.”
Puerto Rico Water Utility Cyberattack: Customer Data Leaked
The Puerto Rico Aqueduct and Sewer Authority, responsible for the territory’s water supply, has confirmed a cyberattack that exposed customer and employee information. The authority emphasized that critical infrastructure was not affected because network segmentation isolated operational systems from business networks. The incident was limited to administrative environments and business data.
“The good news is that operational technology was separated,” said an ICS security expert at Dragos. “But the breach still highlights the risk to public sector utilities managing sensitive citizen data.”
Intuitive Surgical Data Breach: Robotic Surgery Giant Hit
Intuitive, the US-based maker of the da Vinci and Ion robotic surgical systems, reported a data breach following a targeted phishing incident that compromised an employee account. Exposed information includes customer contact details, employee data, and corporate records. The company confirmed that its surgical platforms were not affected.
“Medical device makers are increasingly targeted for intellectual property and patient data,” commented a healthcare cybersecurity analyst. “Even non-critical systems require robust access controls.”
AI Threats Escalate: Agentic Era Attacks Emerge
Check Point Research has documented key developments in the AI threat ecosystem during January–February 2026. The report highlights a shift to the “agentic era,” where threat actors evolve from simple prompting to structured workflows, attacks become AI-led rather than human-led, and safeguard bypass techniques increasingly exploit agent mechanisms.
“We are witnessing a paradigm shift where AI agents are both the target and the weapon,” said a Check Point researcher. “Defenders must adopt AI-aware security frameworks.”
Claude.ai Vulnerabilities Enable Stealthy Data Theft
Researchers discovered three chained flaws in Anthropic’s Claude.ai, enabling invisible prompt injection, silent exfiltration of conversation history via the Files API, and redirection through an open redirect. Anthropic patched the injection issue and is addressing the remaining weaknesses. The chain allows attackers to steal data without detection.
“The combination of these vulnerabilities effectively creates a backdoor into AI conversations,” warned a vulnerability researcher at Snyk. “Enterprises should monitor AI service integrations closely.”
Langflow RCE Flaw Exploited Within 20 Hours
Attackers have weaponized CVE-2026-33017, a critical unauthenticated remote code execution vulnerability in Langflow, an open-source framework for AI agents and RAG pipelines. The bug was exploited within 20 hours of disclosure, allowing arbitrary Python execution on exposed instances via a single crafted request. Check Point IPS provides protection against this threat.
ConnectWise Patches Critical ScreenConnect Flaw
ConnectWise has patched CVE-2026-3564, a critical cryptographic signature verification flaw in ScreenConnect, its remote access platform widely used by MSPs and IT teams. The issue could let attackers use extracted machine keys to authenticate to other servers.
“This patch is critical for MSPs managing multiple client environments,” said a vulnerability management expert at Qualys. “Delaying updates could lead to supply-chain attacks.”
Background
This wave of attacks and vulnerabilities occurs amid a broader escalation in cyber threats targeting both traditional enterprise systems and emerging AI platforms. The breaches at Navia, Aura, Puerto Rico water authority, and Intuitive illustrate the diverse range of targets—from benefits administrators to critical infrastructure and medical device manufacturers. Meanwhile, AI-specific threats are accelerating, with vulnerabilities in Claude.ai and Langflow demonstrating new attack surfaces in agentic AI systems. The rapid exploitation of Langflow within hours of disclosure highlights the urgency of patching.
What This Means
Organizations face an expanding threat landscape where attackers leverage both social engineering and technical exploits. The Navia and Aura breaches underscore the need for robust identity and access management, particularly for cloud-based administrative platforms. The Puerto Rico incident reaffirms the value of network segmentation for OT/IT separation. The AI-related vulnerabilities signal that the era of AI-native attacks has arrived—defenders must integrate security into AI development and deployment workflows. Immediate actions include patching ScreenConnect and Langflow, deploying phishing-resistant MFA, and auditing AI service integrations for prompt injection risks.
For the latest discoveries in cyber research for the week of 23rd March, please download our Threat Intelligence Bulletin.