Heap buffer overflow in php-src under specific OPcache optimization settings poses remote code execution risk
When SCCP is enabled and the compact-literals optimization path is disabled, if both the class and method names of a dynamic static call are optimized by constant propagation, the ZEND_INIT_STATIC_METHOD_CALL_SPEC_CONST_CONST_HANDLER handler incorrectly reads memory outside the function runtime cache. An attacker can craft a malicious PHP script to trigger this logic flaw, causing a heap buffer overflow that may crash the service or enable unauthorized code execution.
Component
php-src is the core source repository of the PHP interpreter. It builds the PHP language runtime and Zend Engine, parses and executes PHP scripts, and integrates Opcode caching to improve performance. It aims to provide efficient and stable dynamic content processing for Web applications across server-side development scenarios.
Review runtime cache allocation during OPcache optimization so that static method calls after constant propagation correctly compute and allocate cache slots, and verify memory access boundaries under different optimization levels to eliminate out-of-bounds reads caused by inconsistent cache sizes.
CSSA Exclusive Early Warning
10.0 Critical
Rule flaw in Coreruleset allows SQL injection bypass and data disclosure risk
The vulnerability stems from missing logic in SQL injection detection rules under certain combined scenarios. When an attacker crafts a payload containing a LIKE (SELECT ...) structure, the PL1-level rule 942100 fails to recognize this variant and incorrectly allows the request. An attacker can build statements such as ' OR (8459 LIKE (SELECT 8459)) * 4906 -- to bypass protection and execute arbitrary SQL, leading to sensitive data disclosure or integrity damage. All protection nodes that use the default configuration and have not upgraded to PL2 mode are affected.
Component
Coreruleset is an open-source Web Application Firewall (WAF) rule set based on the ModSecurity core. It intercepts malicious HTTP requests through deep packet inspection and semantic analysis. Its architecture relies on regular-expression matching and anomaly scoring. Deployed on reverse proxies or front-end servers, it provides defense in depth for backend applications by identifying and blocking common OWASP Top 10 attack vectors.
Immediately update the Coreruleset rule library to fix the detection logic flaw, or temporarily raise the Paranoia Level to PL2 for stricter detection. Review existing logs for potential historical intrusions, and regularly verify that the WAF continues to block new attack variants.
CVE Intelligence
2
CVE-2026-94503
CVSS
10.0 Critical
2026-10-09
Unrestricted upload of dangerous file types in PX-lab Zombify enables Web Shell upload
In PX-lab Zombify version 1.7.7 and earlier, the file upload module lacks strict validation of dangerous file types and contains an unrestricted upload of dangerous file type (CWE-434) flaw. An attacker can upload files containing malicious code (such as a Web Shell) to the Web server and fully control it. Zombify does not effectively filter file extensions or content types during upload, allowing executable scripts to be uploaded. All Zombify users who have not updated to a secure version are affected. The attack can be launched remotely over the network without complex conditions or user interaction.
Component
PX-lab Zombify is a software component for image processing and effects generation, commonly used for image editing features in Web applications.
Risks
From ordinary user to administrator: If the Web server runs with high privileges, an uploaded Web Shell may obtain the highest server privileges
Complete system control: An attacker can use the uploaded Web Shell to execute arbitrary commands, read/modify/delete sensitive data, or move laterally to other hosts on the internal network
No user interaction: An attacker can upload malicious files directly via HTTP requests without inducing any user action
Configure a strict file-upload allowlist at the Web server layer so that only static resources such as images are accepted
Deploy a Web Application Firewall (WAF) to intercept suspicious upload requests and Web Shell signatures
CVE-2026-107824
CVSS
10.0 Critical
2026-10-10
Unauthorized access to x64dbg-MCP Server leads to remote code execution
Before x64dbg-MCP Server version 1.1, this native Model Context Protocol (MCP) plugin for x64dbg exposes the debugger’s full functionality over HTTP. Because it listens on 0.0.0.0 by default and lacks authentication (CWE-306), any unauthenticated network client that can reach the default ports (9094 for x64, 9095 for x32) can execute arbitrary x64dbg commands, attach to processes by PID, read and write the debuggee’s memory, and write files to arbitrary paths. The issue is fixed in version 1.1.
Component
x64dbg-MCP Server is a native plugin for the x64dbg debugger that exposes the debugger’s complete functionality over HTTP and SSE for remote debugging operations.
Risks
Complete system control: An attacker can execute arbitrary code on the victim system and, depending on privileges, install programs, view/change/delete data, or create new fully privileged accounts
Remote unauthorized access: An attacker can control the debugger over the network without authentication and execute arbitrary commands
Memory and file manipulation: An attacker can read and write the debuggee’s memory and write files to arbitrary paths, which may lead to data disclosure or persistent backdoors