CVE-2025-62593
Ray-Project Ray is affected by a code injection and CSRF vulnerability, allowing unauthenticated attackers to execute arbitrary code.
Critical vulnerabilities, curated daily for security professionals
A cluster of Google Chrome flaws scoring CVSS 9.6 dominates the day's critical disclosures, alongside remote code execution issues in IBM AIX and a maximum-severity flaw in the W3 Total Cache WordPress plugin. Critical CVEs totaled 30, down 44% from the prior day's 54, while high-priority CVEs fell 40% to 61. CVE-2026-18051 (CVSS 10) affects W3 Total Cache, CVE-2026-16816 and CVE-2026-15068 (both CVSS 9.9) affect IBM AIX, and CVE-2026-64696 (CVSS 9.8) affects Apple macOS. Enterprise virtualization, collaboration, and machine learning stacks also feature: six CVEs carry confirmed exploitation, including issues in VMware Cloud Foundation, Microsoft SharePoint, Ray, and MLflow. No vendor patch data was recorded for these entries at publication, so verify fix availability directly with each vendor advisory before scheduling remediation.
Immediate action: Prioritize browser fleets running Google Chrome, IBM AIX servers, Apple macOS endpoints, and VMware Cloud Foundation or vCenter deployments, followed by Microsoft SharePoint and any exposed Ray or MLflow instances. Patch availability is unconfirmed in this data set, so check vendor advisories directly and apply documented mitigations or network restrictions where no fix is yet published.
CVSS score (e.g. 9.1) — severity from 0–10. Red marks critical (9+), orange high (7–8.9).
Exploitability — how hard the flaw is to attack, read from the CVSS vector:
The lower the bar on all three, the easier to exploit at scale — “Network · No privileges · No interaction” is the worst case: hit from anywhere, no credentials, no victim action.
🔴 Actively exploited — confirmed under attack in the wild (CISA’s Known Exploited Vulnerabilities catalog). Prioritize these regardless of score.
EPSS · Nth percentile — FIRST.org’s estimated chance a flaw is exploited within 30 days. We flag it only in the top 10% — a statistical signal it’s unusually likely to be targeted, separate from whether attacks are confirmed.
Ray-Project Ray is affected by a code injection and CSRF vulnerability, allowing unauthenticated attackers to execute arbitrary code.
A double-free vulnerability in the Windows IKE Extension allows unauthenticated attackers to execute arbitrary code over a network.
VMware vCenter contains a directory traversal vulnerability in the Syslog server, which may allow an unauthenticated attacker with network access to execute arbitrary code on the system.
A weak authentication vulnerability in Microsoft SharePoint allows unauthenticated remote attackers to bypass security controls and access sensitive information.
An improper authentication flaw in Apple macOS Screen Sharing allows unauthenticated network attackers to bypass credentials and gain remote access.
MLflow contains a Server-Side Request Forgery (SSRF) vulnerability that allows unauthenticated attackers to reach internal services due to improper validation of redirected URLs.
The W3 Total Cache WordPress plugin contains a path traversal vulnerability that allows unauthenticated attackers to write or overwrite arbitrary files on the host server.
A command injection vulnerability in IBM AIX and PowerVM VIOS allows authenticated remote attackers to execute arbitrary system commands via improper neutralization of special input elements.
A command injection vulnerability exists within the NIM component of IBM AIX and PowerVM VIOS, allowing authenticated remote attackers to execute arbitrary commands.
Use after free in Input in Google Chrome prior to 151.0.7922.72 allowed a remote attacker who had compromised the renderer process to potentially perform a sandbox escape via a crafted HTML page. (Chromium security severity: High)
Use after free in DataTransfer in Google Chrome on Windows prior to 151.0.7922.72 allowed a remote attacker who had compromised the renderer process to potentially perform a sandbox escape via a crafted HTML page. (Chromium security severity: High)
The issue was addressed with improved memory handling. This issue is fixed in macOS Sequoia 15.7.8, macOS Sonoma 14.8.8, macOS Tahoe 26.6. A remote user may be able to cause unexpected system termination or corrupt kernel memory.
Integer overflow in ANGLE in Google Chrome prior to 151.0.7922.72 allowed a remote attacker who had compromised the renderer process to potentially perform a sandbox escape via a crafted HTML page. (Chromium security severity: High)
Insufficient validation of untrusted input in Chrome for iOS in Google Chrome on iOS prior to 151.0.7922.72 allowed a remote attacker who had compromised the renderer process to potentially perform a sandbox escape via a crafted HTML page. (Chromium security severity: High)
Type Confusion in ANGLE in Google Chrome prior to 151.0.7922.72 allowed a remote attacker who had compromised the renderer process to potentially perform a sandbox escape via a crafted HTML page. (Chromium security severity: High)
Out of bounds write in ANGLE in Google Chrome on Windows prior to 151.0.7922.72 allowed a remote attacker to potentially perform a sandbox escape via a crafted HTML page. (Chromium security severity: High)
Inappropriate implementation in ANGLE in Google Chrome on Mac prior to 151.0.7922.72 allowed a remote attacker to potentially perform a sandbox escape via a crafted HTML page. (Chromium security severity: High)
The OZOLS software is vulnerable to code execution via an abandoned auto-update domain, allowing for insecure code download and cleartext transmission of sensitive data.
The issue was addressed with improved memory handling. This issue is fixed in macOS Sequoia 15.7.8, macOS Sonoma 14.8.8, macOS Tahoe 26.6. An app may be able to cause unexpected system termination or corrupt kernel memory.
A flaw in the search-v2-operator allows for privilege escalation due to a ClusterRole granting excessive permissions, including the ability to manage RBAC configurations and approve CSRs.
The Balbooa Forms extension for Joomla is vulnerable to pre-authentication PHP code injection via an unescaped query parameter, allowing unauthenticated attackers to execute arbitrary code.
Cisco Crosswork Planning contains a SQL injection vulnerability due to improper neutralization of special elements in SQL commands, which may allow unauthenticated remote code execution.
Cisco Secure Workload contains an improper access control vulnerability that may allow unauthenticated remote attackers to bypass security restrictions and compromise the system.
Cisco Secure Workload contains an improper authentication vulnerability that may allow unauthenticated remote attackers to bypass authentication and execute unauthorized actions.
A missing authentication vulnerability in Cisco Crosswork Planning allows unauthenticated remote attackers to execute critical functions.
An external control of file system path vulnerability in Cisco Crosswork Planning allows remote attackers to manipulate file operations.
An injection vulnerability exists in Cisco Secure Workload due to improper neutralization of special elements in output.
Cisco Crosswork Planning contains multiple vulnerabilities related to insufficiently protected credentials, which may allow for unauthorized access or system compromise.
The TRENDnet TEW-755AP file /cgi-bin/wan.cgi is vulnerable to a stack-based buffer overflow via the cameo.wan.wan_pppoe_password_00 argument, allowing remote code execution.
The TRENDnet TEW-755AP file /sbin/mycli contains a stack-based buffer overflow vulnerability in the function FUN_401000, which can be triggered remotely via the ssid argument.
Etherpad 2.1.0 through 3.0.x incorrectly authorizes OAuth requests to /api/2/*, allowing non-admin users with valid tokens to perform administrative actions.
A stack-based buffer overflow vulnerability in the TRENDnet TV-IP751WIC camera allows remote attackers to execute arbitrary code via the Currenttime argument in set_time.cgi.
J-BusinessDirectory for Joomla contains an unauthenticated arbitrary file upload and deletion vulnerability due to path traversal and improper input validation.
The Zoo extension for Joomla contains an unauthenticated arbitrary file upload vulnerability within the image element, allowing attackers to bypass file type restrictions.
A buffer overflow vulnerability in the Linux kernel MCTP serial driver allows local attackers to trigger arbitrary memory corruption via crafted zero-length frames.
A command injection vulnerability in Advantech ADAM-3600 EdgeLink firmware allows unauthorized execution of arbitrary commands during the verification process.
A privilege management vulnerability in Splunk Enterprise for Windows allows authenticated local users to gain unintended control over system resources.
A buffer overflow vulnerability in the Zephyr HL7800 cellular modem driver allows for potential memory corruption during the processing of AT command responses.
A critical authentication bypass and information exposure vulnerability in phpMyFAQ allows unauthenticated remote attackers to gain unauthorized access to sensitive data.
Gotenberg, a Docker-powered API for PDF files, contains a path traversal vulnerability that could allow unauthorized access to sensitive files on the host system.
The Booking calendar, Appointment Booking System plugin for WordPress is vulnerable to Cross-Site Scripting (XSS) due to insufficient input validation.
The Simple File List WordPress plugin is susceptible to Cross-Site Scripting (XSS) via improper sanitization of user-supplied input.
The SAML Single Sign On WordPress plugin contains an improper authentication vulnerability that may allow unauthorized access to the application.
Splunk Enterprise contains an improper privilege management vulnerability that allows authenticated users to gain unintended control over system resources.
Splunk Enterprise contains an access control vulnerability that allows authenticated attackers to bypass intended restrictions on system resources.
Splunk Enterprise contains a code injection vulnerability where improper input neutralization allows attackers to execute arbitrary code.
Splunk Enterprise is susceptible to improper neutralization of special elements during code construction, which may allow authenticated users to execute arbitrary code.
Splunk Enterprise contains a flaw in query generation that fails to neutralize special elements, potentially allowing unauthorized data access or manipulation via an injection attack.
Splunk Enterprise is vulnerable to path traversal due to improper neutralization of directory sequences, potentially allowing authenticated users to access files outside of restricted directories.
Splunk Enterprise contains an authorization bypass vulnerability that allows authenticated users to perform unauthorized actions or access restricted resources.
Splunk Enterprise is susceptible to code injection due to improper neutralization of externally influenced input, allowing authenticated users to modify intended code execution.
Splunk Enterprise suffers from improper privilege management, allowing authenticated users to obtain unauthorized control over system resources.
A Server Side Request Forgery (SSRF) vulnerability exists in Splunk Enterprise and Splunk Secure Gateway, allowing an authenticated user to perform unauthorized requests to internal resources.
An authorization bypass vulnerability in Splunk Enterprise allows authenticated users to perform unauthorized actions due to improper implementation of access control checks.
A Server-Side Request Forgery vulnerability in the Cisco Talos Intelligence for Enterprise Security Cloud app for Splunk allows unauthorized destination requests.
An insecure deserialization vulnerability in the Splunk AI Toolkit allows authenticated users to execute arbitrary code by supplying malicious serialized data.
A missing authorization vulnerability in Dell Command Update (DCU) allows locally authenticated users to gain elevated privileges and execute malicious actions.
Dell OpenManage Enterprise is vulnerable to SQL injection, allowing an authenticated user to execute arbitrary database commands and compromise sensitive system data.
Dell OpenManage Enterprise contains an OS command injection vulnerability, enabling authenticated users to execute arbitrary commands on the underlying host operating system.
Agno's PythonTools contains a path traversal vulnerability in the joinpath file name argument, allowing potential unauthorized file access.
Termix is vulnerable to OS command injection, which allows an authenticated user to execute arbitrary commands on the underlying server.
Termix suffers from a missing authorization vulnerability that allows authenticated users to perform unauthorized actions within the platform.
ISPConfig contains an authenticated SQL injection vulnerability within its Remote API, allowing attackers to manipulate database queries via the primary ID parameter.
The electerm client contains an OS command injection vulnerability that can be triggered by a malicious user, leading to arbitrary code execution on the host machine.
GitPython contains an argument injection vulnerability that can be exploited to execute arbitrary commands when processing untrusted inputs.
GitPython is vulnerable to a configuration injection attack, allowing a low privileged attacker to execute unauthorized commands by manipulating option names.
The Frauscher FDS 102 system is vulnerable to session hijacking, as administrative session identifiers are inadvertently stored in plaintext within downloadable error log archives.
The Grav API Plugin for Grav CMS is susceptible to an authorization bypass, allowing authenticated users to access or modify content beyond their intended permissions.
A vulnerability in NetScaler ADC and NetScaler Gateway allows for remote service disruption. The flaw is exploitable by unauthenticated attackers over the network.
An out-of-bounds write vulnerability in IBM AIX and PowerVM VIOS allows local or adjacent attackers to potentially execute arbitrary code or cause system crashes.
A wrap-around error in IBM AIX and PowerVM VIOS may allow an adjacent attacker to compromise the integrity, confidentiality, and availability of the system.
A stack-based buffer overflow in IBM AIX and PowerVM VIOS allows an authenticated low-privileged user to achieve full system compromise.
A stack-based buffer overflow in IBM AIX and PowerVM VIOS allows an authenticated attacker to compromise the system.
A critical out-of-bounds write vulnerability exists in IBM AIX and PowerVM VIOS, which may allow an unauthenticated adjacent attacker to cause system crashes or arbitrary code execution.
An OS command injection vulnerability in IBM AIX and PowerVM VIOS allows unauthenticated adjacent attackers to execute arbitrary commands with elevated privileges.
IBM AIX and PowerVM VIOS contain an OS command injection vulnerability that permits unauthenticated adjacent attackers to execute arbitrary commands on the underlying system.
IBM AIX and PowerVM VIOS are affected by an out-of-bounds write vulnerability, which could allow an unauthenticated attacker on the local network to execute arbitrary code or cause a system crash.
IBM AIX and PowerVM VIOS are susceptible to an OS command injection vulnerability, allowing an unauthenticated attacker on the local network to execute arbitrary commands with elevated privileges.
IBM AIX and PowerVM VIOS contain an improper privilege management vulnerability, allowing an unauthenticated attacker on the local network to gain unauthorized access or elevate their privileges.
IBM AIX and PowerVM VIOS are affected by an OS command injection vulnerability, potentially allowing unauthorized command execution by adjacent attackers.
Wekan contains an authorization flaw that allows authenticated users to perform unauthorized actions due to improper privilege management.
A code injection vulnerability exists in marimo before version 0.23.15, which can be exploited via malicious MCP server configuration.
A SQL injection vulnerability exists in vsDesk version 11.06.02, allowing unauthenticated attackers to execute malicious database queries via insecure processing of user-supplied parameters.
IBM AIX and PowerVM VIOS contain an out-of-bounds write vulnerability, which could lead to memory corruption or arbitrary code execution by adjacent attackers.
An OS command injection vulnerability in Cudy WR3000 2.0 routers allows authenticated attackers to execute arbitrary system commands via the mesh MQTT command handler.
An OS command injection vulnerability in Renovate versions 37.158.0 through 37.198.0 allows attackers to execute arbitrary commands via malicious Helm chart configurations.
ArcadeDB contains a code injection vulnerability due to improper control of generation of code within the Groovy fallback mechanism.
A stack-based buffer overflow in the FreeBSD libalias RTSP handler allows for memory corruption due to missing length checks when rewriting outgoing packets.
A vulnerability in the Linux kernel VXLAN Multicast Database (MDB) subsystem can lead to source list corruption during failed replace operations, potentially causing network traffic misrouting.
A memory safety vulnerability in the Linux kernel mac802154 subsystem allows unauthenticated remote attackers to trigger a denial of service via a crafted IEEE 802.15.4 frame.
An out of bounds write vulnerability in the Linux kernel ice driver allows an authenticated guest VM to trigger a kernel page fault and denial of service.
A denial of service vulnerability in the Linux kernel TCP stack allows off-path attackers to terminate half-open TCP connections by injecting in-window RST packets.
A memory handling vulnerability in the Linux kernel amdgpu driver allows for potential resource leaks or denial of service due to improper invocation of memory management functions.
A use-after-free vulnerability in the Linux kernel PPPoE implementation allows local attackers to trigger memory corruption and potential privilege escalation.
A flaw in the Linux kernel rtase driver allows a remote attacker to trigger a transmission hang and denial of service via crafted network packets with malformed headers.
A logic error in the Linux kernel Rados Block Device (rbd) component allows an unauthenticated attacker to trigger a denial of service via a corrupted object map update.
A missing facility check in the Linux kernel ptp_s390 module allows for improper physical clock registration, potentially leading to unauthorized access or system instability.