Microsoft released its February 2026 Patch Tuesday update on Tuesday, patching more than 50 security vulnerabilities across Windows and related software, including six zero-day flaws that attackers are actively exploiting.
Six Zero-Days Under Active Exploitation
The first zero-day, CVE-2026-21510, is a security feature bypass in Windows Shell. A single click on a malicious link silently bypasses Windows protections and executes attacker-controlled content with no warning or consent dialog. It affects all currently supported versions of Windows.
CVE-2026-21513 targets MSHTML, the proprietary rendering engine of Windows’ default browser, while CVE-2026-21514 is a related bypass flaw in Microsoft Word. Both allow attackers to circumvent built-in security controls.
CVE-2026-21533 lets local attackers escalate privileges to SYSTEM-level access within Windows Remote Desktop Services. Separately, CVE-2026-21519 is an elevation-of-privilege flaw in the Desktop Window Manager, a core component that manages how application windows are rendered on screen. Microsoft patched a different zero-day in the same component just last month.
The sixth zero-day, CVE-2026-21525, is a denial-of-service vulnerability in the Windows Remote Access Connection Manager, the service responsible for maintaining corporate VPN connections.
Out-of-Band Patches Since January
Chris Goettl at Ivanti notes Microsoft issued several out-of-band updates between this and January’s Patch Tuesday. On January 17, Microsoft pushed a fix for credential prompt failures affecting remote desktop connections. On January 26, the company patched CVE-2026-21509, a security feature bypass zero-day in Microsoft Office.
More Read
AI Tools in the Crosshairs
This month’s release also includes patches for remote code execution vulnerabilities affecting GitHub Copilot and several integrated development environments, including VS Code, Visual Studio, and JetBrains products. The relevant CVEs are CVE-2026-21516, CVE-2026-21523, and CVE-2026-21256.
Kev Breen at Immersive explained that these vulnerabilities stem from a command injection flaw triggerable through prompt injection, which tricks an AI agent into executing malicious code or commands.
“Developers are high-value targets for threat actors, as they often have access to sensitive data such as API keys and secrets that function as keys to critical infrastructure, including privileged AWS or Azure API keys,” Breen said.
Breen added that organizations deploying AI agents in development pipelines should apply least-privilege principles to limit exposure if developer credentials are compromised. His position is direct: teams should understand the risks, not abandon the technology.
- Six zero-days actively exploited at time of disclosure
- Over 50 total vulnerabilities patched this cycle
- AI-related CVEs span GitHub Copilot, VS Code, Visual Studio, and JetBrains
- Two out-of-band patches issued since January’s Patch Tuesday
The SANS Internet Storm Center maintains a severity-indexed breakdown of all February fixes. Enterprise administrators managing staged rollouts can track patch behavior at askwoody.com.
Photo by Sunny Hassan on Unsplash
This article is a curated summary based on third-party sources. Source: Read the original article
