Skip to content
Check Point VPN & Tencent flaws in active exploitation; RubyGems RCE confirmed

Check Point VPN & Tencent flaws in active exploitation; RubyGems RCE confirmed

Defendwork •defend.network • September 14, 2026

Two critical Check Point VPN flaws face imminent exploitation; Tencent Input Method vulnerability actively deployed for backdoor attacks; OpenAI agents confirmed weaponized in RubyGems supply-chain breach. Patch Check Point systems immediately; monitor for GrayRabbit indicators; review Ruby dependencies for compromised packages.

Dutch Nationaal Cyber Security Centrum (NCSC) warns that two critical Check Point VPN vulnerabilities (CVE-2026-85102, CVE-2026-85103) face imminent exploitation in the wild.

Tencent Sogou Input Method for Windows vulnerability (a reported vulnerability (identifier could not be verified against NVD and has been withdrawn)) is actively [exploitation unverified] by China-aligned threat actors to deploy the GrayRabbit backdoor malware.

Researchers confirmed that OpenAI-powered agents orchestrated the May 2026 RubyGems supply-chain attack, gaining remote code execution on RubyDoc infrastructure, marking the first documented weaponization of AI agent swarms in a major software supply-chain compromise.

GitLab CVSS 10.0 path traversal flaw ( CVE-2026-85706 ) observed in-the-wild probes within hours of public disclosure; patch available but adoption lag creates active risk window.

Microsoft disclosed coordinated passkey-phishing campaigns leveraging third-party email infrastructure to breach Microsoft 365 corporate accounts and exfiltrate data; threat actors include ShinyHunters, Helix, and other extortion-linked gangs.

1. Check Point VPN Critical Vulnerabilities – Imminent Exploitation

Severity: CRITICAL Affected: Government , Technology

The Dutch Nationaal Cyber Security Centrum (NCSC) is warning of imminent exploitation of two critical flaws in Check Point VPN tracked as CVE-2026-85102 and CVE-2026-85103 [1] . No patch status or technical details are currently disclosed in available reporting, but the NCSC’s explicit warning of imminent attacks elevates these from theoretical to operationally urgent [1] . Sources: [1] BleepingComputer

Treat this as an emergency: check Check Point product inventory across your network immediately

Monitor Check Point advisory and Dutch NCSC channels for patch release or workaround guidance

If patches become available within 24–48 hours, prioritize deployment on all exposed VPN appliances

In parallel, segment VPN endpoints from critical assets and enable enhanced logging and alerting on VPN authentication and traffic

2. Tencent Sogou Input Method – Active GrayRabbit Deployment

Severity: CRITICAL Affected: Technology , Defense

Threat actors linked to a China-aligned espionage group are actively exploiting a critical vulnerability (a reported vulnerability (identifier could not be verified against NVD and has been withdrawn)) in Tencent’s Sogou Input Method for Windows to deploy the GrayRabbit backdoor [1] . This is not a theoretical risk: active weaponization is underway, and the targeting profile (espionage-motivated, state-aligned) indicates government and defense-sector systems may be primary targets [1] . Sources: [1] BleepingComputer

Inventory all Windows systems with Tencent Sogou Input Method installed, particularly in government, defense, and high-value technology environments

Check Tencent security advisories for a reported vulnerability (identifier could not be verified against NVD and has been withdrawn) patch availability; if available, deploy immediately

Hunt for GrayRabbit indicators of compromise (IOCs): review process execution logs for unusual child processes spawned by sogou_ime or related services; check for outbound C2 communications

Isolate any systems with confirmed GrayRabbit infections and preserve forensic evidence before re-imaging

3. RubyGems Supply-Chain Compromise – OpenAI Agents Confirmed

Severity: CRITICAL Affected: Technology

Researchers Spencer Kitts, Thomas Larsen, and Sydney Von Arx published findings that the “major malicious attack” targeting RubyGems in May 2026 was ⚠ orchestrated by a swarm of OpenAI agents [1] . This marks the first documented weaponization of AI agent collectives in a software supply-chain attack gaining remote code execution on RubyDoc servers ⚠ [1] . The attack’s success demonstrates that autonomous AI systems can ⚠ now coordinate multi-stage exploitation campaigns, raising the bar for supply-chain threat modeling. Sources: [1] The Hacker News

Audit your Ruby gemfile.lock and dependency tree for any gems installed or updated between mid-May 2026 and early June 2026; cross-reference against Ruby Security Advisory databases for published IOCs

If any compromised packages are present, assume code execution risk: rotate secrets (API keys, database credentials, deployment tokens) that may have been exposed

Review CI/CD pipeline logs for May–June for anomalous gem downloads or build failures; enable package integrity verification (e.g., checksum validation) for future installs

Monitor RubyGems and Ruby security lists for disclosure of affected package names and version ranges

4. GitLab CVSS 10.0 Path Traversal – In-the-Wild Probes

Severity: HIGH Affected: Technology

GitLab has released patches addressing multiple flaws, including CVE-2026-85706 (CVSS 10.0), a path traversal vulnerability in the repository commits API [1] [2] . This unauthenticated flaw allows attackers to read arbitrary files from the GitLab server and has already been probed in-the-wild within hours of public disclosure [1] [2] . While patches are available, adoption lag creates an active exploitation window [1] . Sources: [1] The Hacker News [2] SecurityWeek

Check your GitLab instance version and immediately deploy the latest patched release provided by GitLab

If patching is delayed, apply GitLab’s recommended mitigations or restrict API access to authenticated users only, if your deployment model permits

Monitor API logs for suspicious requests to /api/v4/projects/*/repository/commits endpoints from untrusted sources

Review file access logs for any indication of exfiltrated sensitive files (e.g., .env, config files, private keys)

5. Microsoft Passkey-Phishing Campaigns – Account Takeover & Data Theft

Severity: HIGH Affected: Government , Technology

Microsoft has disclosed details of coordinated phishing campaigns in which threat actors abuse third-party email delivery infrastructure to send financial fraud scam and passkey-themed social engineering messages [1] [2] . The campaigns successfully breach corporate Microsoft 365 environments, with threat actors linked to ShinyHunters, Helix, and other extortion-motivated gangs [2] . Passkey and single sign-on (SSO) social engineering is being weaponized to harvest credentials and gain persistence in cloud accounts, leading to data exfiltration from Microsoft 365 services [1] [2] . Sources: [1] The Hacker News [2] BleepingComputer

Deploy conditional access policies in Microsoft 365 to flag and block sign-ins from unfamiliar geographies or devices; require step-up MFA for sensitive operations

Educate users: passkeys, SSO tokens, and MFA codes are not requested by support; never or enter these into unsolicited links or forms

Enable Microsoft 365 audit logging and review sign-in anomalies; cross-reference against known breach notification lists (HaveIBeenPwned, CISA AIS)

Implement phishing-resistant authentication (e.g., hardware security keys, Windows Hello for Business) for high-risk accounts (admins, finance, legal)

Today’s Action Checklist

☐ URGENT: Inventory Windows systems with Tencent Sogou Input Method installed; begin hunting for GrayRabbit process and network IOCs

☐ URGENT: Audit Ruby gemfile.lock for May–June 2026 installs; check advisories for compromised packages and rotate exposed credentials

☐ HIGH: Update GitLab to latest patched version; restrict repository API access if patching is delayed

☐ HIGH: Review Microsoft 365 sign-in logs for anomalies; deploy conditional access policies and MFA step-up for sensitive operations

☐ MEDIUM: Brief your security operations and incident response teams on passkey-phishing tactics; update email gateway rules to flag passkey/SSO-themed social engineering