Exploitation Report
Executive Summary
Critical exploitation activity is ongoing across multiple fronts, with two high-severity vulnerabilities confirmed as actively exploited in the wild. Fortinet FortiMail CVE-2026-104286, a zero-day allowing unauthenticated arbitrary file writes, has been added to CISA's Known Exploited Vulnerabilities catalog following confirmed attacks. Simultaneously, Dell Container Storage Modules (CSM) CVE-2026-63688 carries a maximum CVSS 10.0 score and enables unauthenticated administrative access and root compromise on Kubernetes nodes.
China-nexus threat actors are conducting sustained campaigns leveraging both known and zero-day vulnerabilities. The Warlock ransomware group continues weaponizing Microsoft SharePoint flaws—described as both old and new—to breach critical infrastructure, government, water utilities, telecom providers, and educational institutions across Portuguese- and Spanish-speaking regions. Separately, TA419 employs adversary-in-the-middle (AitM) phishing impersonating AI policymakers and Anthropic employees to harvest credentials from U.S. think tanks and universities, while the newly documented Antino backdoor uses Outlook and OneDrive for command-and-control in espionage targeting government entities across seven Asian nations.
Active Exploitation Details
CriticalActive exploitationPatchCVE-2026-104286#
- Description: An improper neutralization vulnerability in Fortinet FortiMail that allows unauthenticated attackers to write arbitrary files on the underlying system. The flaw carries a CVSS score of 9.8 and was exploited as a zero-day before disclosure.
- Impact: Attackers can achieve unauthenticated arbitrary file writes on FortiMail appliances, potentially leading to remote code execution, persistent access, and full system compromise without requiring credentials.
- Status: Actively exploited in the wild; added to CISA Known Exploited Vulnerabilities catalog. Fortinet has released patches.
CriticalNot observedPatchCVE-2026-63688#
- Description: Multiple critical vulnerabilities in Dell Container Storage Modules (CSM) that connect Dell enterprise storage arrays to Kubernetes environments. CVE-2026-63688 is a missing authentication for critical function vulnerability in the csm-authorization-storage gRPC server with a CVSS score of 10.0.
- Impact: Unauthenticated attackers can gain administrative access and achieve root compromise on Kubernetes nodes, enabling full control over storage infrastructure and containerized workloads.
- Status: Patches released; Dell urges immediate patching. No public exploitation reported at time of advisory.
HighActive exploitationPatch#
HighActive exploitationMitigate#
- Description: China-nexus espionage group TA419 conducts credential phishing campaigns using adversary-in-the-middle (AitM) techniques targeting AI policy experts at U.S. think tanks, universities, and legal sector organizations. Campaigns impersonate prominent economists, AI policymakers, and an Anthropic employee.
- Impact: Credential theft and account compromise of high-value targets in AI policy and research sectors; potential follow-on espionage and intellectual property theft.
- Status: Active campaigns observed; no software vulnerability exploited—relies on social engineering and AitM phishing infrastructure.
HighActive exploitationInvestigate#
- Description: A previously undocumented backdoor codenamed Antino is deployed by a China-nexus threat actor against government and policy organizations across Taiwan, India, the Philippines, Cambodia, Pakistan, Thailand, and Myanmar. The backdoor uses Microsoft Outlook and OneDrive for command-and-control communications, blending with legitimate traffic.
- Impact: Persistent remote access, data exfiltration, and espionage against government and policy entities in Asia; stealthy C2 via trusted cloud services.
- Status: Active campaign tracked by Cisco Talos; no specific vulnerability exploited—malware delivery vector not specified in reporting.
CriticalNot observedPatch#
- Description: A critical remote code execution vulnerability in GitLab's AI Gateway service (CVSS 9.9) that allows a logged-in user with Duo Agent Platform access to execute arbitrary commands on the gateway under certain conditions. Affects self-hosted GitLab instances running their own AI Gateway.
- Impact: Arbitrary command execution on the AI Gateway server, potentially leading to lateral movement, data access, and supply chain compromise.
- Status: Patched in gateway versions 19.2.4, 19.3.2, and 19.4.1. No active exploitation reported at time of advisory.
HighObservedInvestigate#
- Description: Attackers exploited a vulnerability in third-party software used by Frontline Education to gain unauthorized access to systems and steal employee information including Social Security numbers for school district employees.
- Impact: Exposure of sensitive personally identifiable information (PII) including Social Security numbers for education sector employees across multiple districts.
- Status: Breach confirmed and notifications underway; specific third-party software and CVE not identified in reporting.
HighObservedInvestigate#
- Description: Hackers accessed DTU's identity and access management system and downloaded a large volume of data potentially affecting up to 200,000 users.
- Impact: Large-scale exposure of user data from a major European university's identity infrastructure.
- Status: Breach confirmed; attack vector and specific vulnerability not disclosed in reporting.
CriticalNot observedPatch#
- Description: Remote code execution vulnerabilities in middleware used in SWIFT banking and government environments that could enable hardware-based MFA bypass in ultra-sensitive environments.
- Impact: Potential RCE in financial messaging infrastructure; possible bypass of hardware-based multi-factor authentication.
- Status: Patches available; urgent patching recommended. No active exploitation confirmed in reporting.
CriticalActive exploitationPatch#
- Description: Zero-day vulnerabilities in Kiteworks data-protection platform and Citrix products that demonstrated challenges in zero-day response. One vendor instructed customers to power down the platform during a nine-hour window; the other remained silent on reported attacks prior to patch release.
- Impact: Active exploitation of zero-days in file transfer and virtual desktop infrastructure; operational disruption from emergency mitigation measures.
- Status: Patches eventually released; active exploitation confirmed for at least one zero-day during the response window.
Affected Systems and Products
- Fortinet FortiMail: All versions vulnerable to CVE-2026-104286 prior to patched releases; email security appliances in enterprise and government environments
- Dell Container Storage Modules (CSM): Versions containing the csm-authorization-storage gRPC server; Dell enterprise storage arrays connected to Kubernetes clusters
- Microsoft SharePoint: On-premises and cloud deployments; specific vulnerable versions not disclosed—both older patched flaws and potential zero-days exploited by Warlock
- GitLab AI Gateway: Self-hosted GitLab instances running AI Gateway versions prior to 19.2.4, 19.3.2, and 19.4.1; organizations using Duo Agent Platform
- Frontline Education Platform: School district deployments using vulnerable third-party software component; specific software not identified
- DTU Identity and Access Management System: Technical University of Denmark's central identity infrastructure; platform details not disclosed
- SWIFT Middleware: Banking and government middleware implementations in financial messaging infrastructure; specific products not named
- Kiteworks Data-Protection Platform: Enterprise file transfer and content governance deployments; versions affected by zero-day not specified
- Citrix Products: Virtual desktop and application delivery solutions; specific products and versions affected by zero-day not specified
- Microsoft Outlook and OneDrive: Leveraged as C2 channels by Antino backdoor; legitimate cloud services abused for command-and-control
Attack Vectors and Techniques
- Adversary-in-the-Middle (AitM) Phishing: TA419 uses AitM frameworks to intercept credentials and session cookies in real-time, impersonating trusted contacts including AI policymakers and Anthropic employees to target specific high-value individuals
- SharePoint Vulnerability Exploitation: Warlock leverages both known and suspected zero-day SharePoint flaws for initial access, followed by security tool disablement (likely via living-off-the-land techniques) and ransomware deployment
- Unauthenticated Arbitrary File Write: FortiMail CVE-2026-104286 allows remote unauthenticated attackers to write arbitrary files, a primitive that typically enables RCE through web shell deployment or configuration manipulation
- Missing Authentication in gRPC Service: Dell CSM CVE-2026-63688 exposes a critical gRPC endpoint without authentication, granting direct administrative access and root privileges on Kubernetes nodes
- Cloud Service C2 Tunneling: Antino backdoor uses Microsoft Outlook and OneDrive APIs for command-and-control, blending malicious traffic with legitimate enterprise cloud communications to evade detection
- Third-Party Software Supply Chain Exploitation: Frontline Education breach originated from a vulnerability in a third-party component, highlighting supply chain risk in edtech ecosystems
- Identity System Compromise: DTU breach targeted the central identity and access management system, suggesting credential theft, privilege escalation, or identity provider vulnerabilities
- AI Gateway Command Injection: GitLab AI Gateway flaw allows authenticated users with specific permissions to execute arbitrary commands, indicating insufficient input validation or sandboxing in the AI model interface
- Middleware RCE in Financial Infrastructure: SWIFT middleware vulnerabilities enable remote code execution in high-assurance environments, with potential to bypass hardware MFA through software-layer compromise
- Zero-Day Exploitation with Disclosure Gaps: Kiteworks and Citrix incidents reveal operational challenges when vendors delay communication about active zero-day exploitation, forcing emergency mitigations without full context
Threat Actor Activities
- Warlock (China-linked): Active ransomware campaigns exploiting SharePoint vulnerabilities against critical infrastructure, government, education, water utilities, and telecom providers in Portuguese- and Spanish-speaking countries. Combines vulnerability exploitation with security tool disablement and data encryption. Tracked by Symantec and Carbon Black Threat Hunter Team.
- TA419 (China-nexus): Espionage-focused credential phishing campaigns using AitM techniques targeting U.S. AI policy experts at think tanks, universities, and legal organizations. High-fidelity impersonation of economists, policymakers, and industry figures (including Anthropic personnel) for targeted social engineering.
- Antino Operator (China-nexus): Previously undocumented backdoor campaign targeting government and policy organizations across seven Asian nations (Taiwan, India, Philippines, Cambodia, Pakistan, Thailand, Myanmar). Uses Outlook and OneDrive for stealthy C2. Tracked by Cisco Talos.
- ShinyHunters (Rey/Saif al-Din Khader): Extortion group member reportedly detained in Jordan on September 29, 2026, cooperating with FBI to identify other members. Law enforcement action against data theft and extortion operations.
- Tren de Aragua (TdA): Venezuelan gang members sanctioned by U.S. Treasury for ATM jackpotting attacks stealing millions across the United States. Financial crime operations targeting physical banking infrastructure.
- Unknown Operators (Kiteworks/Citrix Zero-Days): Unattributed threat actors exploiting zero-days in Kiteworks and Citrix products prior to patch availability. Vendor response gaps noted in both cases.
- Unknown Operators (Frontline Education): Unidentified attackers exploiting third-party software vulnerability to breach education sector HR platform and exfiltrate employee PII including SSNs.
- Unknown Operators (DTU Breach): Unidentified hackers compromising university identity management system and exfiltrating data on up to 200,000 users.
- Unknown Operators (Microsoft X Account Hijack): Attackers compromised Microsoft's official X (Twitter) account (13M+ followers) for crypto pump-and-dump scheme. Social media account takeover, not software vulnerability exploitation.