TL;DR
Issabel Framework vulnerability (CVE-2026-89026, CVSS 9.8) is under active exploitation enabling unauthenticated remote code execution. Microsoft released its largest patch batch ever—974 CVEs—but emergency fixes were required after breaking domain trust on some systems. Browser AI assistants are vulnerable to extension-based hijacking across five products.
Executive Summary
- Issabel Framework flaw CVE-2026-89026 enables unauthenticated remote attackers to execute OS commands with CVSS 9.8/9.3; exploitation is active in the wild.
- Microsoft issued its largest single patch batch ever—at least 974 security holes—but the Windows 11 KB5124008 update broke domain trust relationships on some enterprise systems, requiring investigation and likely emergency fixes.
- A single malicious browser extension can hijack AI assistants built into Chrome (Gemini Live), Microsoft Edge, Perplexity Comet, Opera Neon, and the Claude extension, potentially exposing sensitive user interactions.
- Threat groups NightEagle, Hacking Cat, and Toy Ghouls are targeting Russian enterprises with backdoors, ransomware, and wipers; Iranian state-linked hackers deploy CHOSEN BRICK malware targeting dissidents and journalists globally.
- Spanish regulators received the first report of an AI agent independently conducting a full attack chain—login, vulnerability discovery, and data exfiltration—marking a potential escalation in autonomous cyber threats.
Top Threats Today
1. Issabel Framework Unauthenticated RCE Under Active Exploitation
Severity: CRITICAL Affected: Technology
A critical vulnerability in Issabel Framework, a web-based framework for open-source unified communications PBX software, is being actively exploited in the wild. CVE-2026-89026 allows an unauthenticated remote attacker to execute arbitrary OS commands [1]. The vulnerability carries a CVSS v3.1 score of 9.8 and CVSS v4.0 score of 9.3 [1]. No public information indicates a patch is currently available, making this an immediate priority for organizations running Issabel deployments.
Sources:[1] The Hacker News
Recommended Action
- Immediately inventory all systems running Issabel Framework and take them offline or isolate network access pending a vendor security patch.
- Monitor for indicators of compromise such as unexpected process execution or network connections from Issabel servers.
- Subscribe to Issabel security advisories and check for patches on an urgent basis.
- If patching is delayed, implement network segmentation to restrict direct internet exposure to Issabel instances.
2. Microsoft Patches 974 CVEs; Emergency Fixes Required for Domain Trust Regression
Severity: HIGH Affected: Technology, Government
Microsoft issued updates to plug at least 974 security holes in Windows operating systems and other software—the largest single patch batch in the company's history [1]. Among the CVEs identified are CVE-2026-69730, CVE-2026-69829, CVE-2026-81963, and CVE-2026-85880 [1]. However, the Windows 11 KB5124008 security update is breaking domain trust relationships on some enterprise systems, preventing users from logging in with valid ⚠ domain credentials . Microsoft is investigating these reports, and emergency fixes are likely forthcoming [2]. This regression underscores the risk of large patch batches: while AI is helping Microsoft discover vulnerabilities faster, the sheer volume and complexity increase the likelihood of unintended side effects [1].
Sources:[1] Krebs on Security[2] Dark Reading
Recommended Action
- Before deploying KB5124008 broadly, test it in a staging environment with your domain infrastructure to confirm domain trust is preserved.
- If KB5124008 is already deployed and domain trust is broken, isolate affected systems and monitor Microsoft's emergency fix announcements daily.
- Stagger patch deployment across departments rather than applying all 974 fixes simultaneously to allow time for regression identification.
- Document baseline domain trust health (NTLM/Kerberos status) before applying patches so you can quickly detect and isolate problems.
3. Browser AI Assistant Hijacking via Single Malicious Extension
Severity: HIGH Affected: Technology
Security researchers at Forever Security have demonstrated that a single ordinary browser extension can take control of AI assistants built into five Chromium-based products: Gemini Live in Chrome, Perplexity Comet, Microsoft Edge, Opera Neon, and the Claude in Chrome extension [1]. Once installed, the extension can hijack user interactions with these assistants, potentially accessing sensitive information, executing malicious actions, and exfiltrating ⚠ data [1]. This attack does not require a zero-day—it exploits the inherent privilege model of browser extensions to intercept and manipulate AI-mediated communications that users may assume are private.
Sources:[1] The Hacker News
Recommended Action
- Audit installed browser extensions across your organization and remove any from untrusted sources or with unusual permissions.
- Educate users that browser extensions can intercept AI assistant interactions; recommend installing extensions only from official stores and verifying publisher identity.
- Monitor browser extension installation logs and alert on any suspicious extension installs, particularly those requesting broad content-script permissions.
- For sensitive transactions (banking, credentials, confidential decisions), avoid using in-browser AI assistants and use native client applications or offline tools instead.
4. Multiple Threat Groups Target Russian Enterprises; CHOSEN BRICK Malware Targets Global Dissidents
Severity: HIGH Affected: Government, Technology
Kaspersky reports that three threat activity clusters—NightEagle (aka APT-Q-95), Hacking Cat, and Toy Ghouls—are targeting enterprises in Russia with backdoors, ransomware, and wipers [1]. Separately, government agencies warn that Iranian state-linked hackers are using a Windows malware strain named CHOSEN BRICK to target dissidents, activists, and journalists worldwide [2]. These campaigns represent ongoing nation-state and cybercriminal activity with clear intent to disrupt critical infrastructure and silence opposition voices globally.
Sources:[1] The Hacker News[2] BleepingComputer
Recommended Action
- Monitor for CHOSEN BRICK indicators of compromise (IOCs) published by U.S. government agencies and integrate them into endpoint detection and response (EDR) tools.
- Implement application whitelisting on high-value systems to prevent execution of unsigned or suspicious Windows executables.
- For organizations in Russia or with Russian operations, review network logs and EDR telemetry for lateral movement, credential dumping, and backup deletion—hallmarks of NightEagle, Hacking Cat, and Toy Ghouls activity.
- Strengthen incident response plans for ransomware and wiper attacks; ensure offline backup copies exist and are regularly verified.
5. First Agentic AI Data Breach: Autonomous Agent Chains Attack Steps
Severity: HIGH Affected: Technology
Spain's Data Protection Agency (AEPD) received the first report of an attack allegedly carried out with an AI agent powered by a known large language model [1]. The AI agent independently chained together a successful login, vulnerability discovery, and access to personal data ⚠[2]. This represents a milestone in autonomous cyberattacks—rather than a human attacker executing steps, an AI agent performed reconnaissance, exploitation, and data exfiltration without manual intervention between steps. While details remain limited, this incident illustrates the escalating threat of AI-powered autonomous attack chains.
Sources:[1] BleepingComputer[2] SecurityWeek
Recommended Action
- Review authentication and authorization logs for unusual patterns of API calls, privilege escalation, or data access that could indicate autonomous AI agent activity.
- Implement behavioral analytics and anomaly detection tuned to detect rapid, sequential exploitation steps that humans would unlikely execute in quick succession.
- Segment databases and sensitive systems to limit the blast radius if an attacker gains initial access via AI-assisted reconnaissance.
- Establish incident response procedures specific to AI-driven attacks, including chain-of-custody for AI model inspection if an attack is suspected.
Today’s Action Checklist
- ☐ URGENT: If running Issabel Framework, immediately isolate instances from direct internet access and verify no active exploitation has occurred; monitor vendor advisories for patched versions.
- ☐ URGENT: Before rolling out Windows 11 KB5124008 enterprise-wide, test in staging to confirm domain trust is preserved; plan rollback steps if regression occurs.
- ☐ HIGH: Audit browser extensions in use across the organization; remove untrusted or overly-privileged extensions and educate users on extension security risks.
- ☐ HIGH: Integrate CHOSEN BRICK IOCs into EDR/detection tools and review logs for Iranian state-linked malware indicators.
- ☐ MEDIUM: Review AI agent/LLM usage policies in your organization; implement controls and monitoring for AI tools with access to sensitive systems or data.