The Daily Zero-Day

Executive Summary

Today's threat landscape highlights a dangerous escalation in zero-day exploitation, autonomous malware architectures, and advanced evasion techniques. Critical infrastructure and enterprise appliances remain under immediate pressure, led by unpatched remote code execution vulnerabilities in Citrix NetScaler and widespread infostealer campaigns abusing legitimate drivers and tools like Tailscale's tailcat. Simultaneously, autonomous AI C2 implants (CLOSEDQUORUM), AI platform vulnerabilities in ServiceNow, and sandbox escapes underscore how threat actors are weaponizing next-generation technologies. Organizations must prioritize robust asset discovery, rigorous identity hygiene, and enhanced forensic readiness across both traditional and AI-driven environments.

Top Intelligence Briefings

VULNERABILITIES & EXPLOITS

Warning: Two Unpatched Citrix NetScaler RCE Zero-Days Under Active Exploitation

Two new unpatched zero-day vulnerabilities affecting Citrix NetScaler ADC and NetScaler Gateway appliances are currently under active exploitation in the wild, permitting remote code execution. Because these edge appliances sit at the network perimeter, compromise can grant attackers immediate footholds into internal corporate networks.

Actionable Takeaway: Immediately implement strict network segmentation, restrict management interface access to trusted jump hosts, and monitor for abnormal child processes spawned by NetScaler binaries pending official patches from Citrix.

MALWARE ANALYSIS

The Closed Quorum: Inside the First Reported Autonomous AI C2 Implant

Discovered through Cisco Talos’ CAIRN project, the CLOSEDQUORUM malware binary exhibits fully autonomous command and control (C2) capabilities. This breakthrough represents a major tactical shift toward self-governing threat infrastructure that can adapt tactics without real-time human operator intervention.

Actionable Takeaway: Update behavioral monitoring heuristics to detect localized decision-making loops and automated asset discovery patterns indicative of autonomous implants within internal subnets.

THREAT INTELLIGENCE

Lunex Stealer Abuses AMD Driver to Disable Security Monitoring and Steal Browser Credentials

Distributed via compromised Ukrainian websites leveraging ClickFix-style Cloudflare verification checks, the Psychedelic/Lunex Stealer family bypasses host-level security controls by exploiting legitimate AMD drivers to disable security monitoring tools before harvesting sensitive browser credentials.

Actionable Takeaway: Audit vulnerable kernel drivers across endpoints using Windows Defender Application Control (WDAC) or endpoint detection policies to block abused vulnerable driver blocklists.

CLOUD & AI SECURITY

AI Sandbox Escapes: Why Forensic Readiness Matters More Than Containment

Recent analyses of autonomous AI agent sandbox escapes—alongside parallel issues in Google Gemini models—reveal that the root problem is not rogue technology, but persistent, legacy access-control and identity architecture failures that allow lateral movement.

Actionable Takeaway: Implement strict least-privilege guardrails for AI agents and ensure centralized logging platforms capture granular API interactions to preserve forensic fidelity after a container breakout.

Quick Hits & Advisories

Insider Threat: U.S. Soldier Gets 70 Months in Prison for AT&T, Verizon Extortions involving call and text metadata theft.
Data Privacy: Consumer data broker Radaris Loses Domains in Privacy Fight following widespread pressure over ignored opt-out requests.
Research & Community: Read Schneier's notes on Friday Squid Blogging: Participatory Squid Dissection in October in Tennessee.
AI Policy: Commentary published On Anthropic’s AI Misuse Report regarding early-month detection frameworks.
Data Breaches: Massive incidents disclosed, including a Student Loan Breach Exposes 2.5M Records.
Reconnaissance: Watering hole attacks by APT TA423 push the ScanBox Keylogger for targeted espionage.
Mobile & App Security: Essential tips on whether that vibe coded app is safe with 5 pre-download checks.
Financial Fraud: Guidance on what to do if Been told to pay at a Bitcoin ATM? Read this first.
Threat Advisory: Talos explores social engineering tactics in Trust and the enticing consultancy offer.
Platform Security: LinkedIn adds new checks for fake profiles and work histories to combat AI-generated fraud.
Evasion Tactics: Kothamine malware uses Tailscale’s tailcat to receive operator commands over encrypted streams.
Credential Leaks: Have I Been Pwned indexes historical exposures including LimeLeads (17.8M accounts) and Burger King Russia (3.1M accounts).
SANS Intelligence: Review the latest threat breakdowns in A Closer Look at Malware From the Macfinger ClickFix Campaign and the ISC Stormcast For Friday, September 25th, 2026.
Cloud Digest: Explore the ICYMI: August 2026 @AWS Security summary and guidance Supporting ASD’s MFA Campaign.
Phishing & Behavior: Cofense Expands AI-Driven Phishing Defense Platform and insights on Beyond Human Risk.
Dark Web & Infostealers: AhnLab releases reports on Ransom & Dark Web Issues Week 4, September 2026 and the August 2026 Infostealer Trend Report.
Gaming Malware: Fake Xeno Roblox Cheats Deliver Powerful Java Stealer via Discord and community forums.
European Scams: Bitdefender uncovers Fake shops targeting shoppers across Europe with fraudulent Samsung and World Cup deals.
Policy Submissions: UN Reports Citing Citizen Lab Submissions Published alongside a Submission to the Immigration and Refugee Board of Canada.
Vulnerabilities & Advisories: CERT-PL documents a Vulnerability in WEBCON BPS software (CVE-2026-92419) and analyzes a multi-stage toll fraud operation targeting Poland.
Frameworks & Baselines: Center for Internet Security releases the CIS Community Defense Model v3.0 and the CIS Benchmarks September 2026 Update.
Vendor Advisories: Critical advisories issued for Multiple Vulnerabilities in ServiceNow's AI Platform and Multiple Vulnerabilities in Google Chrome.