November 20, 2026

From Zero-Days to Forgotten Patches: TSGA Backdoor, C3RB3R Ransomware and Exploits Within Hours

Executive Summary

Attackers did not wait for proof-of-concept code this week. Within hours of disclosure, scanners were already knocking on vulnerable controllers, while forgotten NAS devices and legacy libraries quietly handed over access through flaws fixed long ago. The targets ranged from firewalls and AI gateways to web servers and refrigeration controllers, a reminder that every internet-facing system, however ordinary, now sits on an attacker’s checklist.  

CISA added seven vulnerabilities to its Known Exploited Vulnerabilities (KEV) catalog, one each in Synology DiskStation Manager, BerriAI LiteLLM, Vanderbilt University REDCap, Apache Axis, Perforce Puppet Enterprise, UnrealIRCd and multiple Check Point products. The additions covered remote code execution, sandbox escape, privilege escalation and command injection flaws in widely deployed enterprise, research and network management software.

Beyond the KEV additions, active exploitation was observed against a race condition flaw in Apache HTTP Server, two vulnerabilities in Copeland XWEB EVO refrigeration controllers, an information disclosure flaw in the TYPO3 Form Framework and an unsafe deserialization flaw in Apache Roller. Several of these flaws were exploited within hours of public disclosure, and in some cases before any public proof-of-concept existed.

On the malware front, threat actors chained PHP2Shell with a Check Point Gaia OS privilege escalation flaw to deploy the TSGA backdoor on Check Point gateways, enabling persistent remote access, configuration theft and lateral movement. Separately, the C3RB3R ransomware group exploited flaws in outdated Synology DiskStation Manager versions to encrypt NAS devices, appending the. L0CK3D extension to victim files.

Key highlights of the week:

  • CISA added seven vulnerabilities to its KEV catalog, affecting Synology DSM, BerriAI LiteLLM, REDCap, Apache Axis, Puppet Enterprise, UnrealIRCd and Check Point products.
  • Attackers chained PHP2Shell with a Check Point Gaia OS privilege escalation flaw to deploy the TSGA backdoor on Check Point gateways.
  • The C3RB3R ransomware group exploited outdated Synology DSM versions to encrypt NAS devices with the .L0CK3D extension.
  • Attackers exploited an upload path leak in the TYPO3 Form Framework and a pre-authentication deserialization flaw in Apache Roller.
  • End-of-life products, including XWEB EVO and Apache Axis 1.x, left some flaws without official fixes.

What are the top trending or critical vulnerabilities observed this week?

Several high-impact vulnerabilities are currently trending across the cybersecurity community, demanding immediate attention and patch prioritization. Monitoring these emerging and widely discussed threats provides valuable insights, enabling organizations to make informed security decisions and strengthen their overall defense posture.

CVE-2026-40217 - Unprotected Alternate Channel vulnerability in BerriAI LiteLLM

An Unprotected Alternate Channel vulnerability affected BerriAI LiteLLM, an open-source AI gateway that gives applications a single interface to multiple large language model (LLM) providers, in builds released through April 8, 2026. The flaw resided in the /guardrails/test_custom_code endpoint, which ran administrator-supplied Python guardrail code inside a sandbox protected only by regex-based source-code filtering. According to X41 D-Sec, an authenticated attacker could bypass the filter by building blocked names at runtime and rewriting generator bytecode to recover the unrestricted builtins and import function. That access let the attacker import the os module and execute arbitrary commands as root in the default Docker deployment, and a public proof-of-concept demonstrated the escape. BerriAI fixed the vulnerability in LiteLLM version 1.83.10, and after confirming exploitation in the wild, CISA included the flaw in its KEV catalog.

CVE-2026-90436 - Unsafe Reflection vulnerability in Apache Axis

An Unsafe Reflection vulnerability affected Apache Axis 1.x through 1.4, an open-source SOAP engine for Java web services commonly bundled in enterprise software. A crafted SOAP fault message could make Axis load and instantiate an attacker-specified class from the application classpath, allowing unauthenticated remote code execution. According to Rapid7, the flaw was exploited in the wild through downstream products that bundled the library: N-central through the “Legacy API” and Confluence Data Center through the “Remote API (XML-RPC & SOAP)”. In the Confluence intrusion, attackers executed shell commands, accessed the database, harvested credentials and moved on to a cloud data warehouse and Microsoft 365. Apache stated that no fixed Apache Axis 1.x release would be published, as the product had reached end of life, though a source-level patch was made available, and CISA added the flaw to its KEV catalog.

CVE-2026-90817 - Code Injection vulnerability in Vanderbilt University Electronic Data Capture (REDCap)

A Code Injection vulnerability affected Vanderbilt University’s Research Electronic Data Capture (REDCap), a web application used by academic and clinical institutions to build online surveys and research databases, in versions 13.3.0 through 16.0.48 LTS, 17.3.9 LTS and 17.4.3. The flaw chained a routing weakness in the public survey passthrough handler with unsafe handling of user-controlled file-path and stream parameters in the Data Import processing logic. An unauthenticated attacker with a valid public survey hash could send crafted HTTP requests to reach the import controller and supply a crafted file-path or stream identifier, which caused attacker-controlled content to be processed in a code execution context on the REDCap server. Vanderbilt University addressed the vulnerability in REDCap versions 16.0.49 LTS, 17.3.10 LTS and 17.4.4 Standard. Evidence of active exploitation led CISA to include the flaw in its KEV catalog.

CVE-2026-114713 - Improper Neutralization of Delimiters vulnerability in UnrealIRCd

An Improper Neutralization of Delimiters vulnerability affected UnrealIRCd versions 6.0.0 before 6.2.7.1. UnrealIRCd is an open-source Internet Relay Chat (IRC) server used to run real-time text chat networks, from small communities to large public ones. A network can link multiple servers, which exchange commands to keep users, channels and bans in sync, and IRC operators rely on powerful commands such as KILL and GLINE to manage users and enforce network-wide bans. According to the vendor, the flaw resided in the labeled-response module, which failed to escape the IRCv3 “label” message tag when echoing the tag into replies. A user could therefore embed extra lines in those replies, and linked servers interpreted the injected lines as server-to-server commands and executed them with server authority. As a result, any ordinary user on a network with two or more linked servers could run commands with server or IRC operator privileges, including KILL and GLINE, and could disconnect users, ban them from the network or perform other operator actions. The project described the flaw as the most critical issue found in UnrealIRCd in the past 15 years. The development team discovered it during an internal audit with the help of AI. The project warned that the vulnerability was being exploited on the internet and later confirmed live attacks, which left unpatched networks at real risk. The vendor did not disclose who carried out the attacks, whom they targeted, or how large the activity was. UnrealIRCd addressed the vulnerability in version 6.2.7.1, and CISA added the flaw to its KEV catalog.

CVE-2026-115731 - Deserialization of Untrusted Data vulnerability in Perforce Puppet Enterprise

A Deserialization of Untrusted Data vulnerability affected the report-processing code of Perforce Puppet, a configuration management platform in which a central Puppet Server distributes configuration to agents installed on managed endpoints. The affected versions were Puppet Enterprise 2023.8.0 before 2023.8.13, 2025.0.0 before 2025.12.1 and 2026.0.0 before 2026.0.1, Puppet Core 8.x before 8.22.2 and 9.0.0 before 9.1.3, and the community-maintained OpenVox fork. Any client holding a certificate signed by the Puppet certificate authority (CA) could submit a report containing a crafted rich-data type through a PUT request, causing the Puppet Server to execute arbitrary code with the privileges of its service user. Where the CA signed certificate requests without strong verification, such as with autosigning enabled, an unauthenticated network attacker could obtain a valid certificate and exploit the flaw. Perforce fixed the vulnerability in security updates released on November 3, 2026, and CISA added the flaw to its KEV catalog after exploitation came to light.

CVE-2026-119618 - Race Condition vulnerability in Apache HTTP Server

A Race Condition vulnerability affected Apache HTTP Server versions 2.4.55 through 2.4.70 with the mod_http2 module enabled, along with standalone mod_h2 versions 2.0.8 through 2.0.42 and backported packages on Ubuntu 22.04 LTS, Ubuntu 20.04 LTS and Red Hat Enterprise Linux 9 and 10. The flaw resided in the assign_headers() function of mod_http2: concurrent HTTP/2 request threads shared a lock-free memory pool without proper locking, so one thread could overwrite memory owned by another. An unauthenticated remote attacker could exploit this to corrupt heap memory and run OS commands with the privileges of the web server process. Singapore’s Cyber Security Agency (CSA) warned on November 16, 2026 of active exploitation, with full technical details and a public proof-of-concept exploit available. The Apache Software Foundation fixed the vulnerability in Apache HTTP Server version 2.4.71, and HTTP/2 could be disabled with “Protocols http/1.1” in httpd.conf as an interim mitigation.

CVE-2026-123039 - Incorrect Permission Assignment for Critical Resource vulnerability in Check Point multiple products

An Incorrect Permission Assignment for Critical Resource vulnerability affected Check Point Firewall, Multi-Domain Security Management Server, Scalable Platforms, Security Gateway and Security Management Server running Gaia OS R81 through R82.20. The Gaia OS DynamicCLI daemon ran as root and listened on a world-writable local socket without authenticating connecting processes, so an authenticated local user could have commands executed as root. According to Check Point, the flaw was exploited as a zero-day starting October 30, 2026. Attackers first gained initial access through PHP2Shell on exposed Mobile Access portals, then used this flaw to escalate to root, deploy webshells and a persistent implant, move laterally to VMware ESXi hosts and, in some Israeli organizations, run a destructive wiper. Check Point fixed the vulnerability in Jumbo Hotfix Accumulator R82.20 Take 14, R82.10 Take 52, R82 Take 132 and R81.20 Take 174, and CISA added the flaw to its KEV catalog.

CVE-2026-124010 - Command Injection vulnerability in XWEB EVO

A Command Injection vulnerability affected Copeland XWEB EVO, a Dixell-developed refrigeration supervisory controller used to manage store refrigeration systems, in versions prior to 5.6.2. The flaw resided in xwconsole.cgi, a setuid-root compiled CGI with no session check. The sh request parameter was passed directly to popen(), so a single unauthenticated HTTP GET request could execute arbitrary commands as root. According to VulnCheck, active exploitation began on November 13, 2026, shortly after public disclosure, with probes sending /cgi-bin/xwconsole.cgi?sh=id requests from a Hong Kong-based source. Copeland fixed the vulnerability in XWEB EVO version 5.6.2 and recommended keeping devices off the public internet, as the product line reached end of service in March 2026.

CVE-2026-124014 - Unauthenticated Arbitrary File Write vulnerability in XWEB EVO

An Unauthenticated Arbitrary File Write vulnerability affected Copeland XWEB EVO, a Dixell-developed refrigeration supervisory controller used to manage store refrigeration systems, in versions 4.3.0 through 5.8.8. The flaw resided in config.lua, which performed no session validation. A crafted POST request with action=set wrote the raw, unvalidated request body to the system configuration file /etc/xwebconfig.json, allowing remote attackers to tamper with device configuration. According to VulnCheck, active exploitation began on November 12, 2026, less than two hours after public disclosure and without any public proof-of-concept, with 125 requests sent from a single Tokyo-based scanner. Copeland did not release a fix and will not release one, as XWEB EVO reached end of service in March 2026, and recommended keeping devices off the public internet and migrating to the supported XWEB PRO platform.

CVE-2026-125160 - Information Disclosure vulnerability in TYPO3

An Information Disclosure vulnerability affected the Form Framework of TYPO3, an open-source content management system, in versions 10.0.0 through 10.4.62, 11.0.0 through 11.5.56, 12.0.0 through 12.4.51, 13.0.0 through 13.4.39 and 14.0.0 through 14.3.11. The Form Framework stored uploaded files in an unguessable sub-folder, but when a save operation failed, the raw error message revealed the folder name to the visitor. Attackers could use the leaked path to locate and access files uploaded through TYPO3 forms, and forms that accepted HTML files could enable cross-site scripting on servers without TYPO3’s recommended Content-Security-Policy. TYPO3 confirmed active exploitation of the vulnerability and released the fix on November 17, 2026, alongside patches for nine other flaws that could be chained into a larger compromise. TYPO3 fixed the vulnerability in versions 10.4.63, 11.5.57, 12.4.52, 13.4.40 and 14.3.12, and recommended restricting allowed MIME types on form upload fields and applying the Content-Security-Policy to the upload folder.

CVE-2026-129448 - Unsafe Deserialization vulnerability in Apache Roller

An Unsafe Deserialization vulnerability affected Apache Roller, an open-source Java-based blog server, in versions 4.0.0 through 6.1.4. The flaw resided in the XML-RPC endpoint, which accepted vendor extension types and deserialized them before authenticating the requester. An unauthenticated remote attacker could therefore send crafted data to execute arbitrary code on the server, and disabling XML-RPC in the admin panel did not protect older versions, as those versions deserialized the request before checking that setting. Apache confirmed active exploitation of the vulnerability. The vulnerability extended an earlier disclosure from September 2026 that listed only Roller 6.1.5 as affected, so scanners relying on that first record could have missed older installations. Apache fixed the vulnerability in Roller version 6.1.6, and recommended setting “enabledForExtensions” to “false” in WEB-INF/web.xml or removing the “/roller-services/xmlrpc” servlet mapping as an interim mitigation.

CVE-2024-10441 - Improper Encoding or Escaping of Output vulnerability in Synology DiskStation Manager (DSM)

An Improper Encoding or Escaping of Output vulnerability affected the system plugin daemon in Synology DiskStation Manager (DSM) and Synology BeeStation OS (BSM), operating systems for Synology NAS and personal cloud devices. The affected versions were BSM before 1.1-65374 and DSM before 7.2-64570-4, 7.2.1-69057-6 and 7.2.2-72806-1, and a remote attacker could execute arbitrary code through unspecified vectors. Although Synology fixed the vulnerability in December 2024, threat actors actively abused the flaw in October 2026 to compromise NAS devices still running outdated DSM versions. The attackers deployed C3RB3R, a variant of the Cerber ransomware, which encrypted victim files with the .L0CK3D extension and dropped ransom notes in affected directories. Synology urged users to update DSM immediately and enable automatic updates, and CISA added the flaw to its KEV catalog in light of the continued exploitation.

What did Cytellite sensors detect this week?

Cytellite telemetry captured active exploit attempts and mass scanning campaigns against exposed services globally. The data highlights which vulnerabilities are under attack and provides source IPs and payloads to authorized teams for detailed threat analysis and validation.

Vulnerabilities Product Severity Title Exploited in the-wild CISA KEV
CVE-2026-63030 WordPress Core Critical Interpretation Conflict Vulnerability in WordPress Core No True
CVE-2025-55182 Meta React Server Components Critical Remote Code Execution vulnerability in Meta React Server Components Yes True
CVE-2025-3248 Langflow Critical Missing Authentication vulnerability in Langflow Yes True
CVE-2025-31324 SAP NetWeaver Critical Unrestricted File Upload vulnerability in SAP NetWeaver Yes True
CVE-2025-30208 Vite High Sensitive Information Disclosure vulnerability in Vite No False
CVE-2025-22457 Ivanti Connect Secure, Policy Secure, and ZTA Gateways Critical Stack-Based Buffer Overflow vulnerability in Ivanti Connect Secure, Policy Secure and ZTA Gateways Yes True
CVE-2024-47176 OpenPrinting CUPS Medium Improper Input Validation vulnerability in OpenPrinting CUPS No False
CVE-2024-3721 TBK DVR-4104 and DVR-4216 Medium OS Command Injection vulnerability in TBK DVR-4104 and DVR-4216 Yes False
CVE-2024-36401 OSGeo GeoServer Critical Eval Injection vulnerability in OSGeo GeoServer GeoTools Yes True
CVE-2024-3400 Palo Alto Networks PAN-OS Critical Command Injection vulnerability in Palo Atlo Networks PAN-OS Yes True

Which vulnerabilities were abused by malware this week?

Active malware campaigns exploited specific vulnerabilities to deliver payloads and carry out post-exploitation actions. Each targeted vulnerability is proactively monitored, manually analysed, and mapped to MITRE ATT&CK tactics and techniques. Insights are derived from the LOVI vulnerability intelligence platform, which aggregates and curates data from multiple sources, OSINT groups, blogs, and data leak sites.

PHP2Shell and Gaia OS Flaw Chained to Deploy TSGA Backdoor on Check Point Gateways

According to Check Point, attackers chained PHP2Shell with the Gaia OS DynamicCLI privilege escalation flaw (CVE-2026-123039) to gain root access on Check Point Security Gateways. The root access was used to deploy TSGA, a standalone passive backdoor launched from /tmp/CPconfigdaemon. TSGA copied itself to the boot persistence directory as slot_CPconfigdaemon and communicated over authenticated, encrypted channels on TCP port 11072 and the Unix socket /tmp/.p_sock, listening on all IPv4 interfaces. The backdoor supported remote shell access, execution of shell commands and Python or Perl scripts, file transfer, theft of appliance configurations, backup and snapshot persistence, process injection and tampering with binary firewall logs. TSGA also dropped TSProxy, an embedded shared-library backdoor written to /lib/libcpfwcommon.so. TSProxy was injected into selected appliance processes, where it intercepted networking functions, handled specially marked attacker connections and provided encrypted SOCKS-style proxying while passing ordinary traffic through unchanged. TSGA traced back to Tiny SHell and shared operator credentials, transport secrets and the same TLS certificate and key with PhiliKit, a passive backdoor targeting Ivanti VPN appliances that ESET previously linked to the China-nexus group UNC5221.

C3RB3R Ransomware exploits outdated synology DSM flaws to encrypt NAS devices

According to Aviatrix, the C3RB3R ransomware group, which had a documented history of targeting NAS infrastructure, compromised internet-exposed Synology NAS devices running outdated DiskStation Manager (DSM) versions. Initial access came through unauthenticated remote code execution flaws in DSM, including the December 2024-patched system plugin daemon flaw and two newly disclosed 2026 DSM flaws, and gave the attackers root access on the devices directly. Once inside, the attackers scanned local file systems and encrypted data on NAS volumes, including backups, appended the .L0CK3D extension to encrypted files and dropped a ransom note named read-me3.txt in every affected directory. The campaign hit individual consumers and small-to-medium businesses globally, and followed earlier NAS-focused campaigns such as StealthWorker and eCh0raix.

Vulnerabilities Severity Title Patch Abused By Malware OSS
CVE-2026-123039 High Incorrect Permission Assignment for Critical Resource vulnerability in Check Point multiple products Yes TSGA backdoor False
CVE-2024-10441 Critical Improper Encoding or Escaping of Output Vulnerability in Synology DiskStation Manager (DSM) Yes C3RB3R Ransomware False

What were the most trending OSS vulnerabilities this week?

Open-Source Software (OSS) vulnerabilities are security weaknesses discovered in publicly available codebases that can be exploited across widely used libraries, frameworks, and tools, often impacting thousands of downstream applications.

CVE-ID Title Ecosystem
CVE-2026-122006 Arbitrary File Write and Read vulnerability in Rails RubyGems
CVE-2026-126187 Unauthenticated Remote Code Execution vulnerability in React npm
CVE-2026-58868 Out of Bounds Write vulnerability in Google Android Android
CVE-2025-39682 Improper Check for Unusual or Exceptional Conditions vulnerability in Linux Kernel Ubuntu
CVE-2024-3271 Command Injection Vulnerability in the run-llama/llama_index repository PyPI

Were any PRE-NVD vulnerabilities identified this week?

PRE-NVD vulnerabilities refer to security flaws that are discovered, discussed, or even exploited in the wild before their official inclusion in the National Vulnerability Database (NVD). These early-stage vulnerabilities often emerge through threat actor chatter, exploit proof-of-concepts, and technical disclosures shared across social media platforms and underground forums, signaling potential exploitation risks before public awareness.

CVE-ID Type of vulnerability Product Reference
CVE-2026-19776 Link Following Local Privilege Escalation G DATA Total Security Resource
CVE-2026-63389 Improper Access Control Filament Curator Resource
CVE-2026-101318 Race Condition FreeBSD Resource
CVE-2026-122990 Improper Authentication Citrix App Layering Resource
CVE-2026-130001 Arbitrary Code Execution Mustache template engine for Ruby Resource

Conclusion

The week’s activity showed that attackers no longer waited for public exploit code. Several flaws were exploited within hours of disclosure, while older vulnerabilities in Synology DSM and end-of-life platforms such as Copeland XWEB EVO and Apache Axis continued to give attackers dependable entry points. Network gateways, AI infrastructure such as LiteLLM, and configuration management systems such as Puppet emerged as high-value targets, because a single compromise exposed credentials, configurations and downstream environments. Organizations needed to move beyond severity scores and prioritize remediation based on confirmed exploitation, internet exposure and the business role of each affected asset. Loginsoft Vulnerability Intelligence (LOVI) supported that shift by delivering timely, contextual intelligence on actively exploited vulnerabilities, threat actor activity and patch status, helping security teams act on the risks that mattered most before attackers could.

FAQs

Q1. What is the TSGA backdoor?

TSGA was a passive backdoor that provided encrypted remote access, command execution, configuration theft and firewall log tampering. It also deployed TSProxy, a SOCKS-style proxy injected into appliance processes.

Q2. Why was the LiteLLM vulnerability significant?

LiteLLM often acts as a central AI gateway holding provider API keys and database credentials, so root-level code execution on LiteLLM could expose an organization’s entire AI infrastructure.

Q3. What should organizations do when a vulnerability has no patch available?

When no patch exists, organizations should reduce exposure by restricting access to affected management interfaces through trusted IP allowlists or VPN connections. Security teams should actively hunt for signs of compromise, such as unexpected files, suspicious processes or unusual outbound traffic. Once the vendor releases a fix, it should be applied immediately and its effectiveness verified.

Q4. How does LOVI help organizations manage vulnerabilities effectively?

Loginsoft Vulnerability Intelligence empowers you to efficiently prioritize and respond to potential vulnerabilities by focusing on those actively exploited in the wild. LOVI correlates vulnerability data with real-world threat activity to reduce noise and improve decision-making. This approach enables faster remediation and stronger security posture.

Q5. What is Cytellite?

Cytellite is a Loginsoft security intelligence platform that provides real-time visibility into emerging threats through a global sensor network. It delivers actionable IP intelligence to help organizations detect, analyze, and respond to attacks quickly. By correlating threat data with live activity, Cytellite strengthens resilience across dynamic threat landscapes.

Stay Ahead

Get the Latest Cybersecurity Insights

Security research, threat intelligence, vulnerability updates, product news, and expert insights, delivered directly to your inbox. Stay informed. Stay secure.