“World’s Best AI Hacker” Just Uncovered A Massive Web Vulnerability
A critical NGINX HTTP/2 vulnerability discovered by an AI researcher enables unauthenticated remote code execution on millions of web servers worldwide.
Velocity
How fast coverage is spreading — measured hourly from article rate × source diversity. How this works →
The brief
Security researchers have disclosed multiple critical vulnerabilities in NGINX, the world's most widely deployed open-source web server, with the most severe being CVE-2026-123007, an HTTP/2 integer underflow that allows unauthenticated remote code execution. The vulnerability affects NGINX's HTTP/2 implementation, specifically an integer underflow condition that can be triggered by unauthenticated attackers sending specially crafted HTTP/2 frames to vulnerable servers.
Coverage emphasizes the extraordinary nature of the discovery, with Forbes highlighting that the vulnerability was uncovered by an individual described as the "World's Best AI Hacker," suggesting artificial intelligence played a significant role in identifying the decade-old flaw. Infosecurity Magazine reports that five critical vulnerabilities were found in total across the NGINX codebase, with CVE-2026-123007 and CVE-2026-123633 both enabling unauthenticated remote code execution according to wiz.io's analysis. The Hacker News notes the ten-year latency between the flaw's introduction and its discovery, underscoring the difficulty of detecting such subtle logic errors in mature, widely audited codebases.
The context for this disclosure is significant: NGINX powers a substantial portion of the world's web infrastructure, serving as a reverse proxy, load balancer, and HTTP cache for countless high-traffic websites, cloud platforms, and enterprise applications. An unauthenticated remote code execution vulnerability in such ubiquitous software represents a worst-case scenario for defenders, as it allows attackers to compromise servers without any credentials or prior access. The involvement of AI-assisted vulnerability discovery marks a potential inflection point in security research, where machine learning systems may uncover classes of flaws that have eluded human auditors for years.
Organizations running NGINX should immediately consult Rapid7's advisory and the official NGINX security announcements to determine affected versions and apply patches. The coordinated disclosure at 15:00 UTC suggests patches are available, but coverage does not specify which NGINX branches receive fixes or whether long-term support releases are included. Administrators should also review wiz.io's analysis of CVE-2026-123633 alongside CVE-2026-123007, as both enable unauthenticated RCE. The Forbes profile of the AI researcher may reveal additional details about the discovery methodology, which could signal how future vulnerability research will evolve. Monitoring for exploitation attempts in the wild will be critical in the coming weeks, as threat actors race to weaponize the disclosed flaws against unpatched infrastructure.
Synthesized by PULSE from the headlines below under a strict no-invention contract. ✓ fact-checked: unsupported claims removed (76% supported) Updated 7m ago.
Quick answers
What is CVE-2026-123007?
CVE-2026-123007 is a critical integer underflow vulnerability in NGINX's HTTP/2 implementation that allows unauthenticated attackers to achieve remote code execution on affected servers.
How long has this vulnerability existed?
According to The Hacker News coverage, the flaw has been present in the NGINX codebase for approximately ten years before its discovery and disclosure on November 18, 2026.
Who discovered the vulnerability?
Forbes reports the vulnerability was uncovered by an individual described as the "World's Best AI Hacker," suggesting artificial intelligence played a key role in the discovery, though coverage does not provide the researcher's name.
Coverage (5)
- CVE-2026-123007: Critical NGINX HTTP/2 Integer Underflow Leads to RCE (FIXED) Rapid7 · 5h ago
- 10-Year-Old NGINX HTTP/2 Flaw Lets Unauthenticated Attackers Execute Code on Servers The Hacker News · 5h ago
- Five Critical Vulnerabilities Found in World's Most Widely Used Open-Source Web Server Infosecurity Magazine · 5h ago
- Unauthenticated Remote Code Execution in NGINX: CVE-2026-123633 and CVE-2026-123007 wiz.io · 5h ago
- “World’s Best AI Hacker” Just Uncovered A Massive Web Vulnerability Forbes · 5h ago
Topics
Related trends
A.I. Helps Find Major Security Bug in Java, a Software Platform Vital to the Internet
A critical security flaw has been discovered in the Java programming language with the help of artificial intelligence.
CISA Confirms Active Exploitation of Two PHP2Shell Flaws Enabling Remote Code Execution
CISA confirms active attacks utilizing critical PHP2Shell vulnerabilities as federal agencies face strict remediation deadlines.
Anthropic's super bug-hunting model Mythos is hardcore good at math, as latest vuln under attack shows
Anthropic's bug-hunting AI model Mythos uncovers a critical vulnerability in an HTTP file server as active exploitation hits the wild.
Warning: Two Unpatched Citrix NetScaler RCE Zero-Days Under Active Exploitation
Global alerts sound as two critical Citrix NetScaler zero-day vulnerabilities face active exploitation.
Microsoft Entra ID Flaw (CVSS 10.0) Exploited in Wild, Allows Remote Code Execution
A maximum severity vulnerability in Microsoft Entra ID is being actively exploited to enable remote code execution, prompting an urgent patch from Microsoft.
Unpatched GeoServer Zero-Day Targeted in Active Exploitation Attempts, Can Lead to RCE
Active exploitation of an unpatched GeoServer zero‑day threatens exposed geospatial servers with remote code execution.