Vulnerability feed

Executive summary

Red Hat's advisory RHSA-2026:89742 covers a security update for the python3.9 package on Red Hat Enterprise Linux 9 and several related editions and support tracks. Red Hat rates the update as Important and says it fixes three flaws, including a certificate verification bypass and a use-after-free issue. The vendor recommends applying the update, and the advisory points to Red Hat's standard update instructions.

Technical summary

The advisory fixes three python3.9 flaws: CVE-2026-12345, an arbitrary file deletion caused by a race condition during temporary directory cleanup; CVE-2026-19553, a certificate verification bypass from missing server_hostname validation in SSLContext.wrap_bio(); and CVE-2026-19445, a use-after-free of a server-side SSLContext when sni_callback switches contexts. The advisory does not state the attack vector or preconditions beyond these descriptions, and it does not give CVSS scores, though it says they are available on the linked CVE pages. Fixed package version is python3.9-3.9.25-7.el9_8.6 for the listed RHEL 9 and related channels. No workarounds are stated, and no detection guidance or exploitation status is given.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from access.redhat.com/errata/RHSA-2026:89742

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

High

Privileges required

None

User interaction

None

Scope

Unchanged

CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:H/I:H/A:H

Affects

Red HatPythonRed Hat Enterprise LinuxRed Hat CodeReady Linux BuilderRed Hat Enterprise Linux for x86_64 9Red Hat Enterprise Linux for IBM z Systems 9Red Hat Enterprise Linux for Power, little endian 9Red Hat Enterprise Linux for ARM 64 9python3.9

Executive summary

This Red Hat advisory covers a flaw in the BIND DNS software shipped with Red Hat Enterprise Linux 9 and related products, including the Extended Update Support 9.8 releases. Red Hat rates the update as Moderate severity, and the flaw can cause a denial of service through crafted DNSSEC responses. Red Hat recommends applying the updated bind9.18 packages.

Technical summary

The flaw is tracked as CVE-2026-19662 and is classed by Red Hat as a denial of service caused by crafted DNSSEC responses. The advisory does not state the attack vector details, required preconditions, or a CVSS score beyond referring to the CVE page. The fixed package version is bind9.18-9.18.29-14.el9_8.11, and the advisory lists it for each affected Red Hat Enterprise Linux 9 and Extended Update Support 9.8 variant. No workaround is stated in the advisory. No detection guidance is stated, and the advisory does not say whether exploitation in the wild has been observed.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Dropped extracted product not named in the source: ISC BIND

Summarized from access.redhat.com/errata/RHSA-2026:89746

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

High

Privileges required

None

User interaction

None

Scope

Unchanged

CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:N/I:N/A:H

Affects

Red HatRed Hat Enterprise Linux

Executive summary

Red Hat's advisory RHSA-2026:89760 updates the container-tools:rhel8 module on Red Hat Enterprise Linux 8, which provides tools such as podman, buildah, skopeo, and runc. Red Hat rates it Important and describes two Go standard library flaws, one causing denial of service through crafted HTTP Range headers and one enabling man-in-the-middle attacks through untrusted intermediate certificates. Red Hat recommends applying the updated packages described in the advisory.

Technical summary

The advisory lists two security fixes: CVE-2026-78667, a denial of service in net/http via crafted HTTP Range headers, and CVE-2026-111943, a man-in-the-middle issue in crypto/x509 involving untrusted intermediate certificates, both attributed to golang. The advisory does not state CVSS scores, attack preconditions, or exploitation status beyond the Important rating. Updated package versions are listed for the container-tools components, such as podman-4.9.4-38, buildah-1.33.15-2, runc-1.2.9-9, and skopeo-1.14.6-5, and the advisory's references point to a Red Hat solution article for application steps. Workarounds and detection guidance are not stated in the advisory.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Dropped extracted product not named in the source: Go golang

Summarized from access.redhat.com/errata/RHSA-2026:89760

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Privileges required

None

User interaction

None

Scope

Unchanged

CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H

Affects

Red HatRed Hat Enterprise LinuxRed Hat container-tools

Executive summary

A smart camera sold under the Yoosee name, model GM8135S with firmware 21.00.01.36 and possibly other versions, sends live video over the network without encryption, so anyone positioned on the network path can watch the stream. The advisory describes this as a zero-day with no vendor fix yet, and it recommends isolating the camera on a separate network with no outbound internet access, or replacing it with a device that encrypts traffic by default.

Technical summary

The advisory describes cleartext transmission of sensitive information (CWE-319), insecure default initialization (CWE-1188), and insufficient verification of data authenticity (CWE-345). An attacker on the network path can capture UDP video packets during an active live stream and reconstruct H.264 video, and the device auto-connects to open Wi-Fi SSIDs named Free-AP0 through Free-AP3 without a password, which enables proximity attacks through deauthentication. Firmware in the SPI flash is not verified, so it can be modified and reflashed, and the advisory states no authentication is needed. Affected versions are stated as up to and including 21.00.01.36, possibly others. No fixed version is stated, and the vendor has not released a firmware update. The advisory gives no detection guidance and no CVSS score is stated. Mitigations are network isolation, blocking outbound internet access, using a self-hosted NVR or TLS reverse proxy for local RTSP, and connecting the device by Ethernet.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from akosibrylle.me/writeups/yoosee-cve

Impact if exploited

How it is exploited

Attack vector

Local

Attack complexity

Low

Attack requirements

None

Privileges required

High

User interaction

None

CVSS:4.0/AV:L/AC:L/AT:N/PR:H/UI:N/VC:H/VI:H/VA:H/SC:H/SI:H/SA:H/E:P/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

YooseeShenzhen Gwell Times TechnologyGwell TimesGrain MediaYoosee GM8135S CameraShenzhen Gwell Times Technology Yoosee GM8135S CameraGwell Times Yoosee Smart CameraGrain Media GM8135S

Executive summary

A smart camera sold as the Yoosee GM8135S, with firmware version 21.00.01.36 and possibly other versions, sends live video over the network in cleartext, so a nearby or network-positioned attacker can intercept and rebuild the stream. The camera also auto-connects to open WiFi networks without a password, which makes interception easier. The advisory states that the vendor has not released a firmware update, and its recommended mitigations are to isolate the camera on a separate network, use a self-hosted secure recorder or proxy, or switch to devices that encrypt traffic by default.

Technical summary

The advisory classifies the issues as cleartext transmission of sensitive information (CWE-319), insecure default initialization of resources (CWE-1188), and insufficient verification of data authenticity (CWE-345). An attacker positioned on the network path or within WiFi range can capture unencrypted UDP video frames during an active live stream without authenticating to the camera. The camera automatically connects to open SSIDs named Free-AP0 through Free-AP3 with no password, and the advisory describes forcing a reconnection by deauthentication. The advisory also says the SPI flash firmware is not verified, so an attacker could modify and reflash it. Affected versions are stated as up to and including 21.00.01.36, possibly others. A fixed version is not stated, and the advisory says no firmware update has been released as of publication. Recommended workarounds are network isolation, using a secure RTSP proxy or NVR, and wiring the device by Ethernet to avoid WiFi auto-connection. No CVSS score or in-the-wild exploitation is stated, and no specific detection guidance is given beyond the sniffing setup described in the write-up.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from akosibrylle.me/writeups/yoosee-cve

Impact if exploited

How it is exploited

Attack vector

Adjacent network

Attack complexity

High

Attack requirements

None

Privileges required

None

User interaction

None

CVSS:4.0/AV:A/AC:H/AT:N/PR:N/UI:N/VC:N/VI:L/VA:N/SC:N/SI:N/SA:N/E:P/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

YooseeGwell TimesGrain MediaYoosee GM8135S CameraGrain Media GM8135SGwell Times Yoosee GM8135S Camera

Executive summary

A smart camera sold under the Yoosee name, the GM8135S Camera running firmware 21.00.01.36 and possibly other versions, sends live video over the network without encryption, so someone positioned on the network path or nearby could watch the stream. The advisory describes this as a serious privacy risk, and the vendor has not released a firmware fix; the advisory recommends isolating the camera on a separate network with no outbound internet access, using a local stream routed through a secured recorder or proxy, or replacing the device.

Technical summary

The advisory classifies the issues as cleartext transmission of sensitive information (CWE-319), insecure default initialization of resources (CWE-1188), and insufficient verification of data authenticity (CWE-345). The /npc/npc binary sends live UDP video frames unencrypted during an active stream, and the attacker must be able to observe traffic on the network path, which the advisory says requires no authentication on the camera. The device is also configured to auto-connect to open WiFi SSIDs Free-AP0 through Free-AP3 without a password, which the advisory says enables proximity-based interception via deauthentication. The advisory states that firmware is not verified on the SPI flash and that a telnet backdoor was set up during testing. Affected versions are stated as up to and including 21.00.01.36, possibly others. The fixed version is not stated because the vendor has not released a firmware update. Recommended workarounds are network isolation, local RTSP access with a secured proxy, and wired Ethernet connection. No detection guidance or CVSS score is stated.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from akosibrylle.me/writeups/yoosee-cve

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

High

Attack requirements

None

Privileges required

None

User interaction

None

CVSS:4.0/AV:N/AC:H/AT:N/PR:N/UI:N/VC:L/VI:N/VA:N/SC:N/SI:N/SA:N/E:P/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

YooseeShenzhen Gwell Times TechnologyGwell Times (HK) Co., LimitedGrain MediaYoosee GM8135S CameraShenzhen Gwell Times Technology Yoosee GM8135S CameraGrain Media GM8135S

Executive summary

A security flaw in GouguOA, reported in version 6.1.2 and described as affecting versions up to 6.1.2, lets any logged-in employee upload a file that the server then runs as code, which can lead to full control of the server. The advisory rates the issue 8.8 High using CVSS 3.1. The advisory's remediation calls for enforcing a file type allowlist on merged uploads, moving storage out of the web root, and fixing the related code errors.

Technical summary

The advisory describes an unrestricted file upload (CWE-434) in the chunked upload endpoint POST /disk/api/chunkUpload, which leads to remote code execution. Any authenticated employee with a default role can exploit it, and no special privileges are required according to the advisory. The extension allowlist is applied only to chunk parts, while the merged file takes its extension from an unchecked POST parameter, so a php extension can be written under public/storage, which Apache executes as PHP. The advisory states the affected version as v6.1.2 and says earlier versions were not tested, with versions containing the same implementation presumed affected. No fixed version is stated, and the advisory gives remediation steps rather than a patched release. The advisory gives no detection beyond hunting for public/storage/**/*.php and public/backup/*.sql files after patching, and it does not state a workaround other than those remediation steps.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/Gum29/rrrreeeeeeee-readme/bl…rd-Vuln-Report-EN.md

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

Low

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:P/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

GouguOAGouguOA

Executive summary

Red Hat has released an update for xorg-x11-server, the X Window System server, on Red Hat Enterprise Linux 8. The advisory rates the update as Important and describes multiple flaws that could allow arbitrary code execution or privilege escalation.

Technical summary

The advisory lists eight CVEs (CVE-2026-93516, CVE-2026-93521, CVE-2026-88812, CVE-2026-93523, CVE-2026-93517, CVE-2026-93518, CVE-2026-93519, CVE-2026-93520) covering use-after-free, double-free, heap and buffer overflows, out-of-bounds writes, integer truncation, and privilege escalation via crafted RandR requests. Attack vector, required privileges, and user interaction are not stated in the advisory text. Affected package versions and fixed versions are not stated in the text provided. No workarounds or detection guidance are stated. CVSS base scores are referenced via the CVE pages but are not given in the advisory text.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Impact if exploited

How it is exploited

Attack vector

Local

Attack complexity

Low

Privileges required

Low

User interaction

None

Scope

Unchanged

CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H

Affects

Red Hatx.orgRed Hat Enterprise LinuxRed Hat Enterprise Linux 8x.org xorg-x11-server

Executive summary

Red Hat has released a security update for the freeipmi package on Red Hat Enterprise Linux 8, which provides tools for managing servers through the Intelligent Platform Management Interface. The advisory rates the update as Moderate severity and fixes six vulnerabilities that can lead to arbitrary code execution or denial of service. The vendor recommends applying the updated freeipmi packages as described in its linked guidance.

Technical summary

The advisory describes stack-based buffer overflows and buffer over-reads in freeipmi, including in the ipmi-oem component, and one flaw triggered by malformed Fujitsu System Event Log responses. The impact listed includes arbitrary code execution for most issues and denial of service for CVE-2026-85505. Affected packages are freeipmi versions before 1.6.20-1.el8_10, and the fixed version is freeipmi-1.6.20-1.el8_10 across the listed RHEL 8 and CodeReady Linux Builder channels. The advisory does not state attack vector, required preconditions, or any workaround. It does not state whether exploitation in the wild has occurred, and it does not give a CVSS score in the text, referring readers to the CVE pages instead. Detection guidance is not stated beyond the update instructions at the linked Red Hat article.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from access.redhat.com/errata/RHSA-2026:89754

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

High

Privileges required

None

User interaction

Required

Scope

Unchanged

CVSS:3.1/AV:N/AC:H/PR:N/UI:R/S:U/C:H/I:H/A:H

Affects

Red HatFreeIPMIRed Hat Enterprise LinuxRed Hat Enterprise Linux for x86_64 8Red Hat Enterprise Linux for IBM z Systems 8Red Hat Enterprise Linux for Power, little endian 8Red Hat Enterprise Linux for ARM 64 8Red Hat CodeReady Linux BuilderFreeIPMIRed Hat freeipmi

Executive summary

Red Hat has released an openssh security update for Red Hat Enterprise Linux 9.2 Update Services for SAP Solutions. Red Hat rates the update as Important and says it fixes six security issues in the OpenSSH client and server.

Technical summary

The fixes cover several classes of flaw: a local man-in-the-middle of X11 forwarding via abstract UNIX socket pre-binding in the OpenSSH client, a double free in the DH-GEX client path during FIPS known-group validation that causes client-side denial of service, a sftp client issue that lets an attacker control the downloaded file location, a use-after-free during client-side host key re-exchange, a sshd security bypass from incorrect handling of forwarding and tunneling options, and an scp file misplacement during remote copy. The advisory does not state affected or fixed package versions, attack preconditions beyond what the titles imply, workarounds, or detection guidance. CVSS scores are referenced via the CVE pages but are not given in this text.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Dropped extracted product not named in the source: OpenBSD OpenSSH

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

High

Privileges required

None

User interaction

None

Scope

Unchanged

CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:H/I:H/A:L

Affects

Red HatRed Hat Enterprise Linux

Executive summary

The advisory says the edoc doctor appointment system, a PHP and MySQL web application, has an unauthenticated SQL injection flaw in its admin add-new endpoint. An outside attacker could read database contents and create arbitrary doctor accounts without logging in. The advisory does not state a severity score, and it does not state a fixed version or a vendor recommendation beyond general remediation advice in the write-up.

Technical summary

The flaw is an unauthenticated time-based blind SQL injection in POST /admin/add-new.php, via the spec parameter, which is concatenated unquoted into an INSERT statement. The endpoint lacks effective authentication because header() redirects do not terminate script execution, so the SQL runs regardless of session state. The advisory describes the same pattern in 60+ other handlers. It notes that stacked queries are not possible through mysqli::query(), so UPDATE and DELETE are not reachable, and that the webuser INSERT hardcodes usertype 'd'. The affected version is given as commit cc258425b915d1af9e298777ccd26d160493e68d, and the advisory states that the affected version range is up to that commit. A fixed version is not stated. A workaround is not stated beyond the remediation advice to add exit after redirects and use prepared statements. Detection guidance such as log signatures or WAF rules is not stated.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from gist.github.com/asdfghjkl12345678910111…2441cca911ecbfaa9acb

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

None

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:P/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

HashenUdaraHashenUdara edoc-doctor-appointment-system

Executive summary

A proof-of-concept exploit script has been published for 2K Games Civilization V version 1.0.3.144. The advisory text provided describes a Lua bytecode exploit that would run a command on the host, and it does not state how serious the flaw is by any official rating or what the vendor recommends.

Technical summary

The advisory is a copied Lua exploit script (exploit.lua) that manipulates string.dump bytecode to gain control of the instruction pointer and then calls os.execute with a command (calc.exe in the sample), which suggests a memory corruption class of flaw. The advisory does not describe the attack vector, whether authentication or user interaction is required, or the precondition that a user must load or run a crafted Lua payload. Affected versions are stated only as 1.0.3.144, and no fixed version, workaround, CVSS score, or detection guidance is stated in the text provided.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/drbloop2000/Sid-Meier-s-Civi…lob/main/exploit.lua

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

None

User interaction

Passive

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:P/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

2K Games2K Games Civilization V

Executive summary

A vulnerability was reported in GiantappMan LiveWallPaper version 2.6.8.0 on Windows, a desktop wallpaper application. The advisory text does not describe the flaw's type or how serious it is, and it does not give a vendor recommendation. Defenders should check whether this application is installed and watch for a vendor fix or further details.

Technical summary

The advisory states only that a vulnerability was found in GiantappMan LiveWallPaper 2.6.8.0 on Windows. The vulnerability class, attack vector, and required preconditions are not stated. No fixed version is stated, and no workaround is stated. No CVSS score, exploitation status, or detection guidance is stated. The rest of the text is repository page content, including branch names such as v3.x, which is not identified as affected or fixed.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/GiantappMan/livewallpaper

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

None

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:P/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

GiantappManGiantappMan LiveWallPaper

Executive summary

A vulnerability has been found in Continue 2.0.0, an AI coding assistant made by continuedev. The advisory says it allows operating system command injection that can be triggered remotely, and that exploit details are public. The advisory does not state a severity score, and the vendor did not respond to early disclosure contact.

Technical summary

The flaw is an OS command injection in the function MCPManagerSingleton.setConnections in core/config/ConfigHandler.ts, within the MCP Server Transport component. The advisory says the attack can be executed remotely, but it does not describe the required preconditions or user interaction. The advisory states that the exploit has been disclosed and may be used. No fixed version, workaround, CVSS score, or detection guidance is stated, and the vendor did not respond to early disclosure contact.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

None

User interaction

Passive

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:P/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

continuedevcontinuedev continue

Executive summary

Red Hat published security update RHSA-2026:89738 for the compat-openssl11 package on Red Hat Enterprise Linux 9.6 Extended Update Support and related subscriptions. Red Hat rates the update as Important and describes two vulnerabilities in the OpenSSL compatibility libraries, one a heap buffer overflow and one a DTLS handshake issue. Red Hat recommends installing the updated compat-openssl11 packages.

Technical summary

The advisory lists two flaws: CVE-2026-63072, a heap buffer overflow in CMS key unwrapping, and CVE-2026-84782, a DTLS issue where retransmitted handshake messages are taken from a stale buffer offset. The advisory does not state attack vectors, preconditions, or CVSS scores, and it refers readers to the CVE pages for those details. Fixed package version is compat-openssl11-1.1.1k-5.el9_6.4 as listed in the package tables. No workarounds or detection guidance are stated, and no exploitation in the wild is mentioned. Update instructions point to https://access.redhat.com/articles/11258.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from access.redhat.com/errata/RHSA-2026:89738

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Privileges required

None

User interaction

None

Scope

Unchanged

CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H

Affects

Red HatOpenSSLRed Hat Enterprise LinuxOpenSSLRed Hat compat-openssl11

Executive summary

Red Hat has released a security update for the rhc client tool on Red Hat Enterprise Linux 8, rated Important by the publisher. The flaws are denial-of-service and man-in-the-middle issues in the Go components that rhc is built with, and Red Hat recommends applying the update described in its advisory.

Technical summary

The advisory lists four CVEs in Go standard library components: CVE-2026-78667 (denial of service via crafted HTTP Range headers), CVE-2026-97031 (denial of service via multiple ECH outer extension references in crypto/tls), CVE-2026-78669 (denial of service via excessive HTTP/2 SETTINGS frames), and CVE-2026-111943 (man-in-the-middle via untrusted intermediate certificates in crypto/x509). The fixed package is rhc-0.2.5-13.el8_10 for the listed RHEL 8 and RHEL 8.10 Extended Life Cycle architectures. Attack vector and preconditions beyond the descriptions are not stated, and no CVSS scores are given in the text (they are referenced via CVE pages not included here). No workarounds or detection guidance are stated; the advisory points to https://access.redhat.com/articles/11258 for update instructions.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Dropped extracted product not named in the source: Go Go standard library

Summarized from access.redhat.com/errata/RHSA-2026:89741

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Privileges required

None

User interaction

None

Scope

Unchanged

CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H

Affects

Red HatRed Hat Enterprise LinuxRed Hat Enterprise Linux for x86_64 8Red Hat Enterprise Linux for IBM z Systems 8Red Hat Enterprise Linux for Power, little endian 8Red Hat Enterprise Linux for ARM 64 8Red Hat Enterprise Linux for x86_64 - Extended Life Cycle 8.10Red Hat Enterprise Linux for ARM 64 - Extended Life Cycle 8.10Red Hat Enterprise Linux for Power, little endian - Extended Life Cycle 8.10Red Hat Enterprise Linux for IBM z Systems - Extended Life Cycle 8.10Red Hat rhc

Executive summary

A flaw has been found in OpenBB, an investment research platform from openbq-org, affecting versions up to 4.7.2 and 5.0.0. The advisory says a public exploit has been published and may be used, and it notes the vendor did not respond to early disclosure. The advisory does not state a severity score or a fix.

Technical summary

The weakness is classified as improper access controls in the Local API Server component, specifically in the file openbb_platform/core/openbb_core/app/model/api_settings.py. The advisory states remote exploitation is possible and that a published exploit may be used. Affected versions are stated as up to 4.7.2/5.0.0, and no fixed version is stated. No workaround is stated, no CVSS score is stated, and no detection guidance is stated. The vendor did not respond to early disclosure, so no vendor patch or guidance is referenced.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

None

User interaction

Passive

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:L/VI:N/VA:N/SC:N/SI:N/SA:N/E:P/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

openbq-orgopenbq-org OpenBB

Executive summary

A flaw in the Datafeedr Product Sets Plugin for WordPress, in versions up to 1.3.24, lets someone perform an action they should not be allowed to perform. The advisory says a public exploit exists and recommends upgrading to version 1.3.25.

Technical summary

The weakness is missing authorization in the dfrps_ajax_delete_saved_search function in functions/ajax.php, part of the AJAX Handlers component. The advisory says the attack may be launched remotely and that an exploit is public. It does not state what privileges or user interaction are required, what impact beyond the missing check is possible, or any CVSS score. Affected versions are up to and including 1.3.24, and the fixed version is 1.3.25. The advisory does not state any workaround or detection guidance. It states the vendor was contacted early about the disclosure.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

Low

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:N/VI:L/VA:L/SC:N/SI:N/SA:N/E:P/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

DatafeedrDatafeedr Product Sets Plugin

Executive summary

A security flaw has been reported in the HeyCyan App version 1.0.103 on Android. The flaw involves improper access controls in a file-sharing component, and it can reportedly be triggered remotely. The vendor was contacted before disclosure but did not respond.

Technical summary

The advisory describes an improper access control weakness in the FileProvider component of HeyCyan App 1.0.103 on Android, with the specific functionality left unspecified. The attack may be initiated remotely. Severity score, CVSS vector, and exploitation status are not stated. Preconditions beyond remote initiation are not stated. Affected versions beyond 1.0.103 are not stated. A fixed version is not stated, and the vendor did not respond to disclosure. No workaround or detection guidance is stated.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

None

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

HeyCyanHeyCyan App

Executive summary

A weakness has been identified in the HeyCyan App version 1.0.103 for Android, which could allow sensitive information to be sent over the network without encryption. The advisory describes the attack as difficult to carry out, and the vendor did not respond to early disclosure contact. No vendor fix or recommended action is stated in the advisory.

Technical summary

The weakness is in the Network Security Configuration component, specifically the AndroidManifest.xml file, and results in cleartext transmission of sensitive information (CWE class cleartext transmission). The attack can be launched remotely but the advisory says it requires high complexity and is difficult to exploit. The advisory does not state any fixed version, workaround, or detection guidance. The vendor was contacted early but did not respond, so no vendor patch or statement is stated.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

High

Attack requirements

None

Privileges required

None

User interaction

None

CVSS:4.0/AV:N/AC:H/AT:N/PR:N/UI:N/VC:L/VI:N/VA:N/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

HeyCyanHeyCyan App

Executive summary

Red Hat published advisory RHSA-2026:89733 for the flatpak package, which is used to build, distribute, and run sandboxed desktop applications on Red Hat Enterprise Linux 9.6 Extended Update Support and related subscriptions. Red Hat rates the update Important and lists four security flaws that can lead to code execution or privilege escalation. Red Hat recommends applying the flatpak update described in the advisory.

Technical summary

The advisory lists four flaws in flatpak: missing symlink protection leading to arbitrary code execution (CVE-2026-90616), extra-data extraction giving arbitrary write access as root (CVE-2026-96275), a buffer overflow in OCI delta stream path names on 32-bit systems (CVE-2026-96280), and local privilege escalation via symlink traversal in the revokefs writer (CVE-2026-96808). The fixed package version is flatpak-1.12.9-4.el9_6.5, as written in the advisory. The advisory does not state CVSS scores in the text provided, attack preconditions beyond the descriptions above, workarounds, or exploitation in the wild, and it refers readers to the CVE pages and to https://access.redhat.com/articles/11258 for update instructions.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from access.redhat.com/errata/RHSA-2026:89733

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Privileges required

None

User interaction

Required

Scope

Unchanged

CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H

Affects

Red HatflatpakRed Hat Enterprise LinuxflatpakRed Hat Enterprise Linux for x86_64 - Extended Update Support 9.6Red Hat Enterprise Linux Server - AUS 9.6Red Hat Enterprise Linux for IBM z Systems - Extended Update Support 9.6Red Hat Enterprise Linux for Power, little endian - Extended Update Support 9.6Red Hat Enterprise Linux for ARM 64 - Extended Update Support 9.6Red Hat Enterprise Linux Server for Power LE - Update Services for SAP Solutions 9.6Red Hat Enterprise Linux for x86_64 - Update Services for SAP Solutions 9.6Red Hat CodeReady Linux Builder for x86_64 - Extended Update Support 9.6Red Hat CodeReady Linux Builder for Power, little endian - Extended Update Support 9.6Red Hat CodeReady Linux Builder for IBM z Systems - Extended Update Support 9.6Red Hat CodeReady Linux Builder for ARM 64 - Extended Update Support 9.6Red Hat Enterprise Linux for ARM 64 - 4 years of updates 9.6Red Hat Enterprise Linux for IBM z Systems - 4 years of updates 9.6Red Hat Enterprise Linux for x86_64 - Extended Life Cycle 9.6Red Hat Enterprise Linux for ARM 64 - Extended Life Cycle 9.6Red Hat Enterprise Linux for Power, little endian - Extended Life Cycle 9.6Red Hat Enterprise Linux for IBM z Systems - Extended Life Cycle 9.6

Executive summary

Red Hat published an Important-rated security update for the compat-openssl11 package on Red Hat Enterprise Linux 9.4 systems, including Update Services for SAP Solutions, 4 years of updates, and Extended Life Cycle channels. The update fixes two OpenSSL flaws, one a heap buffer overflow in CMS key unwrapping and one a DTLS handshake retransmission issue. Red Hat recommends applying the updated packages, and the advisory points to its update instructions.

Technical summary

The advisory describes CVE-2026-63072, a heap buffer overflow in CMS key unwrapping, and CVE-2026-84782, a DTLS issue where retransmitted handshake messages are taken from a stale buffer offset. The advisory does not state the attack vector, required preconditions, or CVSS scores, and points to the CVE pages for those details. The fixed package is compat-openssl11-1.1.1k-5.el9_4.4 across the listed RHEL 9.4 channels. The advisory does not state any workaround or mitigation, and it gives no detection guidance or information on exploitation in the wild.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from access.redhat.com/errata/RHSA-2026:89736

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Privileges required

None

User interaction

None

Scope

Unchanged

CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H

Affects

Red HatOpenSSLRed Hat Enterprise LinuxOpenSSLOpenSSL compat-openssl11

Executive summary

Red Hat published an Important security update for the compat-openssl11 package on Red Hat Enterprise Linux 9.2 systems, including Update Services for SAP Solutions, Extended Life Cycle, and related channels. The advisory fixes two OpenSSL flaws, a heap buffer overflow in CMS key unwrapping and a DTLS handshake retransmission issue, and Red Hat recommends applying the updated packages. The advisory does not state a CVSS score for either flaw in the text provided.

Technical summary

The advisory describes CVE-2026-63072, a heap buffer overflow in CMS key unwrapping, and CVE-2026-84782, a DTLS flaw where retransmitted handshake messages are taken from a stale buffer offset. The advisory does not state the attack vector, exploitation preconditions, or whether either flaw is exploited in the wild. The fixed package version is compat-openssl11-1.1.1k-4.el9_2.3, and the vulnerable prior versions are not stated. No workaround or detection guidance is stated, and the advisory directs readers to https://access.redhat.com/articles/11258 for how to apply the update. CVSS scores are referenced to the CVE pages and are not given in the advisory text.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from access.redhat.com/errata/RHSA-2026:89737

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Privileges required

None

User interaction

None

Scope

Unchanged

CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H

Affects

Red HatOpenSSLRed Hat Enterprise LinuxOpenSSL

Executive summary

Red Hat advisory RHSA-2026:89734 fixes two denial-of-service flaws in the Jackson Java JSON libraries, shipped in Red Hat Enterprise Linux 9.6 Extended Update Support and related RHEL 9.6 subscriptions. Red Hat rates the update Important. Red Hat recommends applying the updated jackson packages listed in the advisory.

Technical summary

The advisory covers two denial-of-service flaws in jackson-databind: CVE-2026-68497, a CPU denial of service via unbounded numeric parsing, and CVE-2026-91777, a denial of service via quadratic forward-reference completion. The advisory does not state the attack vector, required privileges, or user interaction. Fixed package versions are listed as jackson-core-2.21.6-1.el9_6, jackson-databind-2.21.6-1.el9_6.1, jackson-jaxrs-providers-2.21.6-1.el9_6, and jackson-modules-base-2.21.6-1.el9_6, along with pki-jackson packages. No CVSS score is given in the text, and it refers to the CVE pages for details. No workarounds or detection guidance are stated. Administrators should apply the update using the article referenced in the advisory.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from access.redhat.com/errata/RHSA-2026:89734

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Privileges required

None

User interaction

None

Scope

Unchanged

CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H

Affects

Red HatFasterXMLRed Hat Enterprise LinuxFasterXML jackson-coreRed Hat Enterprise Linux for x86_64 - Extended Update Support 9.6FasterXML jackson-databindFasterXML jackson-jaxrs-providersFasterXML jackson-modules-base

Executive summary

A security update for libX11, the core X11 client library, is available for Red Hat Enterprise Linux 8 systems. Red Hat rates the issue as Important and says it is a heap-based buffer overflow that can be triggered through a malicious X server. Red Hat recommends applying the updated libX11 packages, version 1.6.8-10.el8_10, as described in its update instructions.

Technical summary

The flaw is tracked as CVE-2026-88806 and is classified as a heap-based buffer overflow in libX11 triggered via a malicious X server. The advisory does not state the attack preconditions beyond a malicious X server, and it does not give a CVSS score in the text provided, referring instead to the CVE page. The fixed package version is libX11-1.6.8-10.el8_10 across the listed Red Hat Enterprise Linux 8 and Extended Life Cycle 8.10 architectures. The advisory does not state any workaround or mitigation, nor any detection guidance, and it does not state whether the issue is being exploited in the wild.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Dropped extracted product not named in the source: X.Org libX11

Summarized from access.redhat.com/errata/RHSA-2026:89730

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

High

Privileges required

None

User interaction

Required

Scope

Unchanged

CVSS:3.1/AV:N/AC:H/PR:N/UI:R/S:U/C:H/I:H/A:H

Affects

Red HatRed Hat Enterprise LinuxRed Hat Enterprise Linux for x86_64 8Red Hat Enterprise Linux for IBM z Systems 8Red Hat Enterprise Linux for Power, little endian 8Red Hat Enterprise Linux for ARM 64 8Red Hat Enterprise Linux for x86_64 - Extended Life Cycle 8.10Red Hat Enterprise Linux for ARM 64 - Extended Life Cycle 8.10Red Hat Enterprise Linux for Power, little endian - Extended Life Cycle 8.10Red Hat Enterprise Linux for IBM z Systems - Extended Life Cycle 8.10

Executive summary

A code injection flaw affects Sony NNabla Neural Network Libraries up to version 1.39.0. Converting a malicious model file can run attacker-chosen commands on the machine doing the conversion, and the advisory rates it High with a CVSS v3.1 score of 8.8. The advisory states that the project is end of life and no upstream fix is expected, so it recommends not converting .nnp files from untrusted sources until a patched fork exists.

Technical summary

The flaw is CWE-94 code injection: the NNP converter passes the attacker-controlled function.type field into Python eval() calls in func_set_nnabla_version_decorate (utils.py lines 111 and 116). A duplicate network name lets a malicious second network bypass the version gate, which only checks the first same-name network. Exploitation requires the victim to run nnabla_cli convert on a crafted .nnp file (network attack vector, user interaction required, no privileges). The advisory states that the affected version is 1.39.0 and that the earliest affected version is unknown, and it gives no fixed version because the project is EOL. Recommended fixes are replacing eval() with getattr/hasattr, validating function.type against known types, and enforcing unique network names. The advisory does not state any detection guidance, and no vendor advisory is listed.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/3em0/cve_repo/blob/main/2026…n-in-NNP-Converter-('function.type'-eval()).md

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

None

User interaction

Passive

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:P/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

SonySony Neural Network LibrariesSony NNabla

Executive summary

Sony Neural Network Libraries (nnabla) up to and including v1.39.0 contains a code injection flaw in its command-line parameter encoder. An attacker who supplies a malicious parameter directory can run arbitrary operating system commands when a victim runs the documented encode_param command, and the advisory says this happens without any visible error. The advisory rates the issue High under CVSS v3.1 with a score of 8.8, and states that the project is end-of-life so no upstream patch is expected, with the workaround being to avoid processing parameter directories from untrusted sources.

Technical summary

The flaw is CWE-95 eval injection: the first line of each .txt parameter file is passed to Python's eval() without validation in load_param_in_txt() in encode_decode_param.py at line 37, reachable through nnabla_cli encode_param. Exploitation requires the victim to run the command on an attacker-supplied directory (CVSS AV:N, PR:N, UI:R, AC:L), and the advisory's proof of concept shows the canary command completing with exit code 0. Affected versions are stated as all versions up to and including v1.39.0, and the advisory says vulnerable code remains on the master branch; no fixed version is stated because the project is end-of-life and no upstream patch is expected. The suggested remediation is to replace eval() with ast.literal_eval() plus integer validation, and the stated workaround is to never run encode_param or decode_param on untrusted directories. The advisory gives no detection guidance beyond the proof of concept, and it does not state whether exploitation has been observed in the wild.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/3em0/cve_repo/blob/main/2026…)-Code-Injection-in-'nnabla_cli%20encode_param'-Leading-to-Arbitrary-Command-Execution.md

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

None

User interaction

Passive

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:P/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

SonySony Neural Network LibrariesSony Neural Network Libraries (nnabla)

Executive summary

A command injection flaw has been reported in the docker-mcp project by QuantGeekDev, affecting versions 0.1.0 and 0.2.0. The advisory says the flaw can be triggered remotely, and it does not state a severity score or a vendor recommendation.

Technical summary

The flaw is an OS command injection in the _build_windows_command function of docker_executor.py, triggered by manipulating the project_name argument. The advisory states the attack can be started remotely. Affected versions are listed as 0.1.0/0.2.0. Fixed versions are not stated. Workarounds are not stated. Exploitation in the wild is not stated. Detection guidance is not stated, and no CVSS score is given.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

None

User interaction

Passive

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

QuantGeekDevQuantGeekDev docker-mcp

Executive summary

A SQL injection flaw has been reported in the Online Cake Ordering System version 1.0 from itsourcecode. An attacker could potentially read, change, or disrupt the application's database, and the advisory describes the risk as serious. The advisory recommends using prepared statements, validating user input, limiting database account permissions, and running regular security audits; it does not state a vendor-issued fix.

Technical summary

The flaw is an SQL injection in the id parameter of /deletetransac.php, which the advisory says fails to sanitize or validate input before it is used in SQL queries. The proof of concept is a GET request with a crafted id value that uses GTID_SUBSET error-based injection against the host 154.219.114.125:9012. The advisory is inconsistent about preconditions: the root cause text says the attacker must log in with valid credentials, while the POC section says no authentication is required and exploitation does not require prior access. Affected version is listed as V1.0. A fixed version is not stated, no CVSS score is stated, and exploitation in the wild is not stated. The suggested repairs are prepared statements, input validation, minimal database permissions, and security audits; no workaround or detection guidance is given.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/RyanCarter9527/submit/issues/1

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

Low

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:P/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

itsourcecodeitsourcecode Online Cake Ordering System

Executive summary

A stored cross-site scripting flaw has been reported in GouguCMS up to version 4.08.18, affecting its user level management feature. An administrator account can save malicious script as a user level name, which then runs in the browser of any administrator who views the list. The advisory does not state a severity rating or a vendor fix, so the recommended response is not stated.

Technical summary

This is a stored cross-site scripting (XSS) vulnerability in the user level function of GouguCMS v4.08.18, in app/admin/controller/Level.php and related files. An authenticated administrator can submit HTML or JavaScript in the title parameter, which is stored in cms_user_level.title without output encoding and rendered as HTML in the admin user-level list. The advisory provides a proof-of-concept POST request to /admin/level/add that sets the title to a script tag, and notes the same value also appears in the user list via /admin/user/index. The advisory does not state a fixed version, a workaround, a CVSS score, or whether exploitation has been observed in the wild. Detection guidance beyond the reproduction steps is not stated.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/yuhuixiaowu/cve/blob/main/go…ucms_v4.08.18_XSS.md

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

High

User interaction

Passive

CVSS:4.0/AV:N/AC:L/AT:N/PR:H/UI:P/VC:N/VI:L/VA:N/SC:N/SI:N/SA:N/E:P/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

GouguCMSGouguCMS

Executive summary

A flaw has been reported in Fusion Creators Studio SCP Containment Breach Multiplayer, affecting versions up to 1.295. The advisory text describes a buffer overflow that could allow remote code execution, and it links to a public exploit repository. The advisory does not state a severity rating, and no vendor fix or recommendation is provided in the text.

Technical summary

The advisory describes a buffer overflow vulnerability that allows arbitrary remote code execution. The stated preconditions are subscribing to a malicious addon on the workshop or connecting to a server that runs the RCE addon, and then playing the game. The affected version is stated as up to 1.295 in the title, and no fixed version is stated. No workaround, CVSS score, or vendor patch guidance is stated. Detection guidance is not stated, and the referenced repository includes a proof-of-concept exploit file named Exploit.gs.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/drbloop2000/SCP-CBM-RCE/tree/main

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

None

User interaction

Passive

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:P/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

Fusion Creators StudioFusion Creators Studio SCP Containment Breach Multiplayer

Executive summary

The advisory reports a vulnerability in SteelRaven7 Ravenfield, but the text provided is a Steam Community web page and contains no description of the flaw itself. The severity and vendor recommendations are not stated.

Technical summary

The advisory text only says a vulnerability was detected in SteelRaven7 Ravenfield 25407073, followed by boilerplate Steam Community page content. The vulnerability class, attack vector, and preconditions are not stated. Affected and fixed versions are not stated beyond the version string 25407073 shown in the title. No workaround, detection guidance, CVSS score, or evidence of exploitation is stated. Defenders should consult the NVD entry and the vendor for technical details.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from steamcommunity.com/app/636480/events/590690402783479125

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

None

User interaction

Passive

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

SteelRaven7SteelRaven7 Ravenfield

Executive summary

A stored cross-site scripting flaw affects CloudClassroom-PHP-Project version 1.0 and prior, in the addnewstudent.php and studentdetails.php components. The advisory rates it Medium with a CVSS score of 6.1 and says the product is unpatched with no fix available yet. The advisory's recommended mitigations are output encoding, input validation, restricting access to the affected component, and reviewing sessions and credentials.

Technical summary

The flaw is stored (persistent) cross-site scripting (CWE-79) in addnewstudent.php, where the fname, lname, faname, addrs and course fields are saved without sanitization and echoed back without HTML encoding, and the name is rendered unencoded in studentdetails.php. The advisory states that no authentication is needed to inject the data, that a victim admin or faculty user must view the record (user interaction required), and that the attack is remote over the network via POST. Affected versions are listed as 1.0 (and prior), and the fixed version is listed as None. The advisory gives workarounds: htmlspecialchars encoding with ENT_QUOTES, input validation, Content-Security-Policy, prepared statements, network ACLs and WAF rules, and session invalidation. It provides log search indicators for suspicious values such as script, onerror, onload and svg tags. It does not state a CVE identifier, an official vendor response, or any evidence of exploitation in the wild.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/vinniboy021-blip/Advisory/bl…tudent-stored-xss.md

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

None

User interaction

Passive

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:N/VI:L/VA:N/SC:N/SI:N/SA:N/E:P/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

CloudClassroom-PHP-ProjectCloudClassroom-PHP-ProjectCloudClassroom-PHP-Project addnewstudent.php

Executive summary

The advisory describes a weakness in ggml-org llama.cpp, a C/C++ library for running large language models. The text provided is mostly repository content and does not describe the weakness, its severity, or a vendor fix, so these points are not stated.

Technical summary

The advisory does not state the vulnerability class, attack vector, or preconditions. It does not list affected or fixed versions, workarounds, CVSS score, or exploitation status. The only identifiable reference is to the llama.cpp master branch repository page, which is not a version statement. Defenders should consult the NVD entry at the provided URL or the upstream project's security policy for the specific details.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/ggml-org/llama.cpp

Impact if exploited

How it is exploited

Attack vector

Local

Attack complexity

Low

Attack requirements

None

Privileges required

Low

User interaction

None

CVSS:4.0/AV:L/AC:L/AT:N/PR:L/UI:N/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:P/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

ggml-orgggml-org llama.cpp

Executive summary

The advisory describes a flaw in smg-project smg versions 1.7.0 through 1.10.1 that appears to involve a crafted JPEG image causing excessive memory allocation. The advisory text is mostly a copied GitHub gist page and does not state a severity rating or a fixed version. No vendor recommendation is stated in the text provided.

Technical summary

The advisory text is a copied gist page containing a proof-of-concept Python script that writes JPEG files declaring dimensions of 65500x65500 and 20000x20000, which the script says would require a large byte allocation in a decoder, and PNG control files. The vulnerability class is not explicitly stated, though the script comments suggest resource exhaustion through image dimension handling. The attack vector and preconditions are implied to be supplying a crafted image to the affected software, but this is not explicitly stated. Affected versions are listed as 1.7.0/1.8.0/1.9.0/1.10.0/1.10.1, and fixed versions, workarounds, CVSS score, and detection guidance are not stated.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from gist.github.com/Yunzez/24a1bf4c226fbc324ea8da1d202c6d9b

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

None

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:N/VI:N/VA:L/SC:N/SI:N/SA:N/E:P/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

smg-projectsmg-project smg

Executive summary

A vulnerability was reported in EyouCMS up to version 1.8.3, a content management system. The advisory describes a server-side request forgery flaw that can be triggered from the public site, but it does not state a severity rating.

Technical summary

The advisory describes server-side request forgery caused by a missing SSRF check in the remote_file_to_local() function, which can be reached from the front end. It states the affected range as up to 1.8.3, but no fixed version is stated. Attack preconditions, CVSS score, and exploitation in the wild are not stated. No workaround is stated. No detection guidance is stated; the text mainly consists of repository file listings and references to a reproduction document and vulnerability report that are not included here.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/yy-t1/security-reports/tree/…9%8D%E5%8F%B0%20SSRF

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

None

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:P/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

EyouCMSEyouCMS

Executive summary

A flaw in AshPostgres, a database library used with the Ash framework, can let one tenant read another tenant's record in a multi-tenant application. The advisory rates it MEDIUM with a CVSS 4.0 base score of 6.0. The advisory points to a fix in version 2.15.2 and the fix commit 8aa8122, and it does not state any workaround.

Technical summary

This is an Incorrect Authorization flaw (CWE-863) in the ash_postgres data layer. When an upsert is skipped during Ash.bulk_create/4 or a single create with return_skipped_upsert?: true, the lookup of the existing row did not apply attribute multitenancy, so a user could receive another tenant's record that shares the same identity values. The advisory says the affected range is from 2.6.22 before 2.15.2 on Hex, and from commit af0903f up to commit 8aa8122 on GitHub. The CVSS vector is CVSS:4.0/AV:N/AC:L/AT:P/PR:L/UI:N/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N, which requires low privileges and a resource that uses attribute multitenancy with return_skipped_upsert?: true. The advisory does not state any workaround or in-the-wild exploitation, and it gives no specific detection guidance beyond the affected routines.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from cna.erlef.org/cves/CVE-2026-115635.html

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

Present

Privileges required

Low

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:P/PR:L/UI:N/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

ash-projectash-project ash_postgres

Executive summary

A flaw in the ash_postgres library, used with the Elixir Ash framework, lets a user in one tenant see records that belong to other tenants' data when certain relationships are loaded. The advisory rates it MEDIUM with a CVSS 4.0 score of 6.0. The advisory states the affected range is from 0.38.0 before 2.15.2 on Hex, and the fix is in commit 8c25c94 on GitHub; it does not explicitly state an upgrade recommendation beyond that fix.

Technical summary

This is an Incorrect Authorization flaw (CWE-863) where AshPostgres builds lateral join queries for many_to_many and through relationships without running the resource's read action, so attribute multitenancy is not applied to the join resource. The attack requires an authenticated user with low privileges (PR:L) and network access, and the preconditions are that the join resource uses :attribute multitenancy with rows from several tenants and that the relationship is loaded with a lateral join, such as with a limit, offset or sort. The advisory states the affected Hex versions as 0.38.0 up to but not including 2.15.2, and the GitHub affected range as commit 7115bec up to but not including 8c25c94. A workaround is not stated in the advisory. Detection guidance is not stated beyond the affected function names lateral_join_query/3, through_step_queries/4 and run_query_with_lateral_join/4 in lib/data_layer.ex.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from cna.erlef.org/cves/CVE-2026-115627.html

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

Present

Privileges required

Low

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:P/PR:L/UI:N/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

ash-projectash-project ash_postgres

Executive summary

A flaw in the Ash framework, an Elixir library from ash-project, lets a client read destination records that should only be reachable through a specific relationship. The advisory rates it MEDIUM with a CVSS 4.0 score of 6.0. The vendor's fix is in version 3.37.1, and the advisory says affected versions are 3.24.0 before 3.37.1.

Technical summary

This is an Incorrect Authorization flaw (CWE-863) in the Ash read relationships code, specifically lateral_join?/4 in lib/ash/actions/read/relationships.ex. When a has_many through relationship is loaded for a single record with a limit of 1, or an unsorted has_one through relationship is loaded with a limit of 1, the query omits the filter linking destination records to the source, so the load returns unrelated records; paging with an offset can enumerate more of them. The attack vector is network with low complexity, and the CVSS vector requires low privileges and no user interaction, with high confidentiality impact. The advisory states the affected range as 3.24.0 up to but not including 3.37.1, with fix commit e374633, and says the vulnerable Git range is 24f41e2 to e374633. The advisory does not state a workaround or any detection guidance, and does not state whether the flaw has been exploited in the wild.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from cna.erlef.org/cves/CVE-2026-114214.html

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

Present

Privileges required

Low

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:P/PR:L/UI:N/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

ash-projectash-project ash

Executive summary

A remote code execution weakness has been reported in DedeCMS up to version 5.7.118, a content management system. The advisory rates it 7.2 under CVSS v3.1 base and 8.8 for the other listed score, and it lists a code injection flaw in the update guide script.

Technical summary

The advisory describes a code injection flaw in dede/update_guide.php that can lead to remote code execution, mapped to CWE-94, CWE-96, and CWE-352. The table lists CVSS v3.1 scores of 7.2 / 8.8 without stating which score is the base score, and it does not state the attack vector, required privileges, or user interaction in plain terms. The advisory does not state a fixed version, a workaround, or any detection guidance. It also does not state whether exploitation has been observed in the wild.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/SDshadow/cve

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

High

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:H/UI:N/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:P/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

DedeCMSDedeCMS

Executive summary

A SQL injection flaw has been reported in the Performance Indicator System 1.0 from SourceCodester, a PHP and MySQLi web application. The advisory says an attacker can exploit it without logging in to read, change, or delete database data. The advisory does not state a severity rating, a fixed version, or a vendor patch.

Technical summary

The flaw is SQL injection in the 'contact' parameter of /opils/staff/save_profile.php, where user input is placed directly into SQL queries without validation. The advisory describes it as time-based blind and says no login or authorization is required. The affected version is listed as V1.0, and no fixed version is stated. No workaround is stated beyond the suggested repair of using prepared statements with parameter binding, strict input validation, least-privilege database accounts, and regular audits. The advisory gives no detection guidance, no CVSS score, and does not say whether exploitation in the wild has been observed.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/hubdk01/cve/issues/13

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

None

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:P/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

SourceCodesterSourceCodester Performance Indicator System

Executive summary

A SQL injection flaw has been reported in the Performance Indicator System 1.0 from SourceCodester, a PHP and MySQLi application. The advisory says an attacker needs no login to exploit it, which could expose or change database contents. The advisory does not state a vendor fix or official severity score, and it recommends using prepared statements, input validation, least-privilege database accounts, and regular security audits.

Technical summary

The flaw is SQL injection in the 'contact' parameter of /opils/staff/edit_dealer.php, which the advisory says is passed directly into SQL queries without validation. The advisory describes it as time-based blind injection, reproduced with sqlmap, and says no login or authorization is required. The affected version is listed as V1.0; a fixed version is not stated. No workaround beyond the suggested code fixes (prepared statements, input validation, minimal database permissions) is given, and no CVSS score or detection guidance is stated. Exploitation in the wild is not stated.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/hubdk01/cve/issues/12

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

None

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:P/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

SourceCodesterSourceCodester Performance Indicator System

Executive summary

A SQL injection flaw has been reported in the Performance Indicator System 1.0 from SourceCodester, a PHP and MySQLi application. The advisory says an attacker who needs no login could read, change, or delete database data and potentially disrupt the service. The advisory does not state a severity score or a vendor fix, and it recommends prepared statements, input validation, and least-privilege database accounts.

Technical summary

The flaw is a SQL injection in the 'contact' parameter of /opils/staff/adddealer.php, reached through a multipart/form-data POST request. The advisory states no login or authorization is required, and it describes a time-based blind variant that was confirmed with sqlmap. The affected version is listed as V1.0, and no fixed version is stated. No vendor patch, workaround beyond the suggested code fixes, CVSS score, or exploitation in the wild is stated. The suggested repair is to use prepared statements with parameter binding, validate and filter input, restrict database account privileges, and run regular security audits. The advisory gives no specific detection guidance such as log signatures.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/hubdk01/cve/issues/11

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

None

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:P/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

SourceCodesterSourceCodester Performance Indicator System

Executive summary

HTTP::Server::PSGI for Perl, in versions before 1.0056, can be crashed by a single request that declares a very large Content-Length. The advisory describes the impact as the server dying with an out-of-memory error on Linux. The advisory points to a code change that caps body reads at 64 KiB, and the fixed release is named as the version that includes that change.

Technical summary

The flaw is a resource exhaustion issue: the request body was read with sysread() using the full remaining Content-Length, so Perl tried to grow the buffer to that size before reading and failed with "Out of memory!" on Linux. The attack vector is a single crafted HTTP request sent to the server, with no authentication stated as required. The advisory says versions before 1.0056 are affected and the patch caps reads at 64 KiB (CVE-2026-115825 is referenced in the commit), but a specific fixed version number beyond the 1.0056 threshold is not stated. Workarounds are not stated, and no detection guidance is given. CVSS score and exploitation status are not stated.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/plack/Plack/commit/11545871c…9cfeb4ebc1069a.patch

Score

No CVSS score published.

Affects

HTTP::Server::PSGIHTTP::Server::PSGI

Executive summary

Starman, a Perl web server, has a flaw that lets an attacker exhaust server memory by sending chunked request bodies that the server keeps buffering without limit. The advisory does not state a severity rating or CVSS score. The fix is included in version 0.4019, and the advisory indicates the patch is available in the repository.

Technical summary

The advisory describes unbounded buffering of chunked request bodies in the chunk decoder, which can cause memory exhaustion, and the commit message also describes a quadratic-backtracking regex in chunk-size line parsing (CVE-2026-115822) that consumes CPU. The attack requires sending crafted chunked HTTP requests to a Starman server; the advisory does not state authentication or other preconditions. Affected versions are stated as those before 0.4019, and the fixed version is 0.4019. The fix replaces the regex with index() and an anchored parser, limits chunk-size lines to 4096 bytes, rejects sizes over 13 hex digits with 400 and closes the connection, and makes _prepare_env return an error status so bad bodies are not dispatched. No workarounds are stated, and no detection guidance or exploitation status is stated.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/miyagawa/Starman/commit/0f33…71988e22476acf.patch

Score

No CVSS score published.

Affects

StarmanStarman

Executive summary

Starman, a Perl web server, sends response bodies for HEAD requests and for 1xx, 204 and 304 responses, which can cause response desynchronization on keep-alive connections. The advisory does not state a severity rating or an exploitation status, and the fix is in Starman versions 0.4019 and later.

Technical summary

The flaw is a response desynchronization issue (HTTP response smuggling class) in Starman before 0.4019 for Perl, where bodies are written for HEAD, 1xx, 204 and 304 responses even though those responses must not carry a body. On keep-alive or pipelined connections, the extra bytes are read as the start of the next response, and a front end that pools upstream connections can be affected. The advisory does not state a CVSS score, attack preconditions beyond keep-alive or pipelined connections, or a workaround. The fix is in commit 91c47c529cd1e100f938486896ea9083ec3ec7bf, which skips body output for these responses and closes object bodies without reading them, and it adds a test file t/no_body_response.t. Detection guidance is not stated, and the advisory names no specific fixed release beyond the version threshold of 0.4019.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/miyagawa/Starman/commit/91c4…ea9083ec3ec7bf.patch

Score

No CVSS score published.

Affects

StarmanStarman

Executive summary

The advisory says Starman, a Perl web server, is affected by a flaw that lets a remote sender exhaust CPU on the server by sending chunked request bodies with long runs of hex digits. The advisory does not give a severity rating or CVSS score. The vendor-side fix is a code change that rejects malformed or overly long chunk-size lines with a 400 error, and the advisory says versions before 0.4019 are affected.

Technical summary

The flaw is a regular expression backtracking problem (quadratic complexity, a form of algorithmic CPU exhaustion) in parsing the chunk-size line of chunked request bodies, which the commit message says ran again over the whole buffer after every read. The attacker needs only to send an HTTP request with Transfer-Encoding: chunked; the advisory says 32 KiB of input cost about 1.7 seconds of worker CPU. The fix bounds chunk-size lines to 4096 bytes, parses them with an anchored linear-time pattern, rejects sizes with more than 13 significant hex digits, and returns 400 and closes the connection on malformed input. The advisory states affected versions as before 0.4019, but it does not state a fixed version number explicitly, and no workaround is stated. No detection guidance or evidence of exploitation in the wild is stated.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/miyagawa/Starman/commit/0f33…71988e22476acf.patch

Score

No CVSS score published.

Affects

StarmanStarman

Executive summary

Starman, a Perl web server, has a flaw where a client that sends a request body and then stalls can keep a server worker busy indefinitely, which can exhaust workers with only a few connections. The advisory describes the severity only in terms of workers being taken; a CVSS score is not stated. The fix adds a total time limit for reading request bodies, and the advisory's guidance is to update Starman to a release at or after 0.4019 (the fixed version is stated as 0.4019 in the advisory's wording of affected versions before 0.4019).

Technical summary

The weakness is a resource exhaustion issue (missing timeout on request body reads): read_timeout applied only to request headers, so a Content-Length or chunked body could stall or trickle indefinitely. The attack requires only network access to the Starman listener and a few open connections to tie up workers, and no authentication is mentioned. Affected versions are stated as before 0.4019 and the fix adds a --body-timeout option (body_timeout, default 300 seconds, 0 disables) that responds with 408 Request Timeout and closes the connection. The advisory does not state a CVSS score, a CWE, exploitation in the wild, or detection guidance beyond the timeout behavior, and it recommends a reverse proxy that buffers request bodies as a mitigation. Workaround details beyond setting body_timeout are not otherwise stated.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/miyagawa/Starman/commit/89df…15586ba2588558.patch

Score

No CVSS score published.

Affects

StarmanStarman

Executive summary

A vulnerability has been reported in Alfresco Activiti, a workflow product, affecting versions up to 7.21.0-rc.692. The advisory says it can be exploited remotely to inject code, and the vendor did not respond when contacted about it. The advisory does not state a severity score or a fix.

Technical summary

The flaw is described as code injection in the JuelScriptEngine.eval function of JuelScriptEngine.java, which is reached by performing a manipulation that the advisory does not specify. The attack can be initiated remotely. The advisory does not state any required privileges or user interaction, and it does not state a CVSS score. The advisory affects versions up to and including 7.21.0-rc.692, and no fixed version is stated. No workaround or detection guidance is stated. The vendor was contacted early but did not respond.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

Low

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

AlfrescoAlfresco Activiti

Executive summary

KodCloud KodExplorer, a web-based file manager, is affected by a vulnerability in versions up to 4.55. The advisory says the flaw can be exploited remotely and that exploit details are public, but it does not state a severity score. The advisory does not state any vendor-recommended fix, and the vendor did not respond to early disclosure.

Technical summary

The flaw is an unrestricted upload issue in the Online Editor Save Endpoint, located in app/controller/editor.class.php, where manipulation of the path/filestr argument is reported. It can be attacked remotely, and the advisory says an exploit has been publicly disclosed and may be used. Affected versions are stated only as up to 4.55, and no fixed version is stated. No workaround, detection guidance, CVSS score, or authentication requirement is stated in the advisory.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

Low

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:P/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

KodCloudKodCloud KodExplorer

Executive summary

A flaw has been found in KodCloud KodExplorer, a file management product, up to version 4.56. The advisory describes it as a server-side request forgery issue that can be triggered remotely. The vendor did not respond when contacted about the disclosure, and no fix or vendor recommendation is stated.

Technical summary

The flaw is server-side request forgery in the request_url_safe function of app/function/web.function.php, within the Remote Download component. An attacker can trigger it remotely by manipulating input to that function. The advisory states affected versions as up to 4.56 but does not state any fixed version. No workaround, exploitation status, CVSS score, or detection guidance is stated. The vendor was contacted early but did not respond.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

Low

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

KodCloudKodCloud KodExplorer

Executive summary

A SQL injection flaw was reported in the Performance Indicator System 1.0 from SourceCodester, a PHP and MySQLi web application. The advisory says an attacker needs no login to exploit it, which could expose or change database contents and disrupt the service. The advisory does not state a fixed version or a vendor patch, and it recommends code fixes such as prepared statements and input validation.

Technical summary

The flaw is SQL injection in the 'id' parameter of /opils/admin/deleteproduct.php, where user input is placed directly into SQL queries without validation. The advisory states that no login or authorization is required, and it provides a POST request example. Boolean-based blind and time-based blind techniques were demonstrated with sqlmap. The affected version is listed as V1.0, and no fixed version is stated. No vendor workaround is stated beyond the suggested remediation of parameterized queries, input filtering, least-privilege database accounts, and security audits. No detection guidance or CVSS score is stated.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/hubdk01/cve/issues/5

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

Low

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:P/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

SourceCodesterSourceCodester Performance Indicator System

Executive summary

A weakness has been reported in SH3, a product whose Gestão and Arrecadação management modules are affected. The advisory says the flaw is hard to exploit remotely, but a public exploit has been disclosed and may be used. The vendor's recommendation is to upgrade the affected component to version 2026.0.2.0.

Technical summary

The flaw is classified as a password hash computed with insufficient computational effort, which weakens protection of stored credentials if an attacker obtains the hashes. The attack can be carried out remotely, but the advisory says it is highly complex and rated difficult to exploit. A public exploit has been disclosed. The advisory states that upgrading to version 2026.0.2.0 mitigates the issue; it does not state the vulnerable version range. No workaround, CVSS score, or detection guidance is stated in the advisory.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

High

Attack requirements

None

Privileges required

High

User interaction

None

CVSS:4.0/AV:N/AC:H/AT:N/PR:H/UI:N/VC:L/VI:N/VA:N/SC:N/SI:N/SA:N/E:P/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

SH3SH3 Módulo de Gestão/Módulo de Arrecadação

Executive summary

YesWiki versions from 4.5.0 before 4.7.0 contain an SQL injection flaw in the Bazar entry creation feature, where a date field from the request is inserted into a database query without escaping. The advisory rates it High with a CVSS score of 8.2 and says the fix is in version 4.7.0, so users should upgrade to 4.7.0 or later.

Technical summary

The flaw is CWE-89 SQL injection in PageManager::save(), where the date_maj_fiche POST field is concatenated raw into an INSERT SET statement via the forcedDate parameter instead of being escaped like the other columns. The advisory describes a network attack with low complexity and no privileges required on the authenticated path, and says an unauthenticated path via action=saisir_fiche works on a default install, with confidentiality impact rated High. Affected versions are stated as >= 4.5.0, < 4.7.0 in the package data and <= 4.6.7 in the description, and the patched version is 4.7.0. The advisory suggests escaping $forcedDate at the sink or validating date_maj_fiche as a strict datetime, but does not state any other workaround. Detection guidance is not stated, and whether the flaw is exploited in the wild is not stated.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/YesWiki/yeswiki/security/adv…/GHSA-wjf7-hrf3-9gcx

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

None

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:L/VA:N/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

YesWikiyeswiki

Executive summary

YesWiki, a wiki package used with ActivityPub federation, has a missing authorization flaw in its sync feature. A malicious federated server can cause a YesWiki administrator's routine sync to delete, overwrite, or create entries that belong to other federated actors. The advisory rates it High (8.1 CVSS) and states the fix status is not fixed, with the affected range given as versions 4.6.2 through 4.6.7; the advisory does not state a patched version for the sync path.

Technical summary

The flaw is CWE-862 (missing authorization) in ActivityPubService::syncActorPosts, which processes Delete, Update, and Create activities from a remote outbox without the assertOwns() and sameHost() checks used by the inbox handler, and runs a host-scoped garbage-collection delete. Exploitation requires an administrator to trigger the sync action (a GET request with no CSRF token), and the attacker needs no privileges on the wiki; the CVSS vector is CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:N/I:H/A:H. The advisory lists the affected range as >= 4.6.2 and <= 4.6.7 in the body, while the metadata says '< 4.7.0' and the patched version is listed as 4.7.0, but it also says the fix status is 'not fixed' for the sync path. Workarounds are stated as avoiding sync with untrusted actors and adding a CSRF token to the sync action, which the advisory suggests but does not confirm as implemented. No detection guidance is stated, and no exploitation in the wild is stated.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/YesWiki/yeswiki/security/adv…/GHSA-cr64-h6xp-9g2j

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

None

User interaction

Passive

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:N/VI:H/VA:H/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

YesWikiyeswiki

Executive summary

YesWiki, a wiki platform, has a critical flaw that lets anyone on the network read data from its database, including user credentials, without logging in. It can also be used to make the site unavailable by overloading the database. The advisory recommends updating to 4.7.0 or 5.0.0-alpha4, and restricting network access to the public Bazar entries API until then.

Technical summary

The flaw is an unauthenticated SQL injection (CWE-89) in the Bazar entries search API, reachable through GET /api/forms/{formId}/entries and /api/entries/bazarlist. A backtick in the query or searchfields field name breaks out of a backtick-quoted SQL alias, because the code uses mysqli_real_escape_string, which does not escape backticks. Injected SELECT subqueries can read any table the database user can access, and results are returned in the JSON response, while SLEEP-based subqueries can exhaust database connections. The advisory states the vector is network with low complexity and no privileges or user interaction, and gives CVSS 3.1 vector AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:H with a score of 9.1. Fixed versions are stated as 4.7.0 and 5.0.0-alpha4, while the affected range is inconsistently described as 4.6.0 before 4.7.0 and 5.0.0-alpha1 before 5.0.0-alpha4 in one place and as 4.6.7 through 5.0.0-alpha3 in the detailed text. Recommended interim mitigation is restricting network access to the public Bazar endpoints. Detection guidance is not stated, and no in-the-wild exploitation is mentioned.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/YesWiki/yeswiki/security/adv…/GHSA-ccp2-6q93-hggh

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

None

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:N/VA:H/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

YesWikiyeswiki

Executive summary

YesWiki, a wiki package distributed through Composer, has a stored cross-site scripting flaw affecting versions from 4.2.0 up to but not including 4.7.0. The advisory rates it Moderate (CVSS 6.4) and says any logged-in user who can comment on a page can run scripts in the browsers of everyone who views that page, including administrators. The advisory lists 4.7.0 as the patched version, but its body says no fix exists in any published release, so administrators should confirm the status with the vendor before relying on an upgrade.

Technical summary

The flaw is CWE-79 stored cross-site scripting in comment bodies. Comments are saved with only whitespace trimming and are rendered with the Twig raw filter, and the wiki raw-HTML syntax is enabled by default through allow_raw_html set to true, so any authenticated user with comment access on a page can inject script through POST /api/comments. The CVSS 3.1 vector is AV:N/AC:L/PR:L/UI:N/S:C/C:L/I:L/A:N, and the advisory says no user interaction beyond viewing the page is needed. The header lists affected versions as >= 4.2.0, < 4.7.0 and patched version as 4.7.0, but the body says the confirmed affected range is >= 4.2.0, <= 4.6.7 and that the fix status is not fixed. The advisory suggests sanitizing comment bodies at storage or setting allow_raw_html to false as an interim step, but it does not state a fixed release that the body confirms. No detection guidance or indicators of exploitation in the wild are stated.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/YesWiki/yeswiki/security/adv…/GHSA-wq96-vjjh-hfc6

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

Low

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:L/VI:L/VA:N/SC:L/SI:L/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

YesWikiyeswiki

Executive summary

YesWiki versions before 4.7.0 contain an unauthenticated server-side request forgery flaw. Any anonymous visitor can make the YesWiki server fetch arbitrary internal or external URLs and read the responses back. The advisory rates it Moderate with a CVSS 3.1 base score of 5.3, and the fix is to upgrade to version 4.7.0.

Technical summary

The flaw is CWE-918 server-side request forgery in the wiki raw-inclusion syntax [[|URL]], which fetches attacker-supplied URLs without scheme, host, or IP validation and echoes the response body. It is reachable without authentication through the page render handler using the content parameter, and also stored via a page body saved by a registered editor that fires when anonymous visitors view it. The advisory states the affected versions are less than 4.7.0 and the patched version is 4.7.0. The CVSS vector is CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:N/A:N. A workaround is not stated, and no detection guidance beyond the proof of concept's access log observation is provided. No CVE ID is listed in the advisory, and exploitation in the wild is not stated.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/YesWiki/yeswiki/security/adv…/GHSA-x9j5-r988-9jcc

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

None

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:L/VI:N/VA:N/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

YesWikiyeswiki

Executive summary

YesWiki, a wiki package distributed through Composer, has a cross-site request forgery flaw in its admin despam action. An unauthenticated attacker who gets a logged-in administrator to view a planted wiki page could permanently delete pages or restore old revisions, with the front page and configuration pages at risk. The advisory rates it High (CVSS 8.1) and recommends adding CSRF token checks, but it contradicts itself on which versions are fixed.

Technical summary

The flaw is CWE-352 in tools/security/actions/despam.php, where the destructive clean (suppr[]) and revision-restore (rev[]) branches are guarded only by a session-based UserIsAdmin() check, with no CSRF token, Origin, or Referer verification. The advisory describes the attack as requiring default configuration (default_write_acl='*', allow_raw_html=true), an attacker-created wiki page containing an auto-submitting form, and an administrator visiting that page; the session cookie's SameSite=Lax setting still allows the same-site POST. Affected versions conflict within the advisory: the header lists affected versions as before 4.7.0 with patched version 4.7.0, while the body states a confirmed affected range of <= 4.6.8 and says the fix status is not fixed. The advisory also says the v5 pre-release 5.0.0.0-alpha3 is affected, though it is outside the stable range it scopes. The stated CVSS 3.1 vector is AV:N/AC:L/PR:N/UI:R/S:U/C:N/I:H/A:H with a base score of 8.1. Recommended fixes are to require a CSRF token on the clean and revision-restore branches, with CSP script-src and form-action 'self' and a permissions.action entry for despam as defense-in-depth. Detection guidance and exploitation in the wild are not stated, and the advisory notes no CVE is assigned.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/YesWiki/yeswiki/security/adv…/GHSA-cv88-hx5q-vq72

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

None

User interaction

Passive

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:N/VI:H/VA:H/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

YesWikiyeswiki

Executive summary

YesWiki versions before 4.7.0 contain a stored cross-site scripting flaw in the filtertags action. An anonymous visitor who can create or edit a page could store script that runs for every later viewer of that page. The advisory rates it High and says the fix is to upgrade to version 4.7.0.

Technical summary

The flaw is stored cross-site scripting (CWE-79) in the filtertags action, where attacker-controlled token values are echoed into page HTML without output encoding, including the button inner text, the title parameter, and the tagbadges value. The payload is injected through the action attribute syntax, which the wakka formatter processes before its allow_raw_html check, so it bypasses that control. Preconditions stated are that default page ACLs on a fresh install are read=* and write=*, and the advisory reports anonymous exploitation using a hashcash token fetched without authentication. Affected versions are stated as before 4.7.0 and the patched version as 4.7.0. The advisory does not state a workaround. It recommends output-encoding the echo sites in tools/tags/actions/filtertags.php. Detection guidance is not stated.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/YesWiki/yeswiki/security/adv…/GHSA-f53x-3gqx-xq76

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

None

User interaction

Passive

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:N/VI:N/VA:N/SC:L/SI:L/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

YesWikiyeswiki

Executive summary

A stored cross-site scripting flaw in YesWiki lets a low-privilege authenticated user upload a malicious SVG file through an image field, and the script runs for people who open the entry or the stored file. The advisory says the earlier fix for a related form-upload issue did not cover this image field, so the flaw remains on the patched version. The advisory lists the affected version as before 4.7.0 and the patched version as 4.7.0, and it recommends applying the same upload sanitization to the image field.

Technical summary

This is a stored cross-site scripting issue (CWE-79) caused by unrestricted SVG upload through the ImageField class in tools/bazar/fields/ImageField.php, which stores files without the sanitizer that the sibling FileField class received in the earlier fix. The attacker needs a valid authenticated account (PR:L) and user interaction (UI:R), and posts a multipart request to /api/entries/{formId} with an SVG file that is then served as image/svg+xml on the wiki origin. The advisory states the affected versions as less than 4.7.0 and the patched version as 4.7.0, and it says the fix for this specific field is not yet present. No workaround is stated. No detection guidance is stated beyond the proof-of-concept description. The advisory does not state any exploitation in the wild or a CVSS score.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/YesWiki/yeswiki/security/adv…/GHSA-c5q6-7233-3v6g

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

Low

User interaction

Passive

CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:P/VC:N/VI:N/VA:N/SC:L/SI:L/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

YesWikiyeswiki

Executive summary

Snipe-IT versions before 8.5.0 contain a flaw that lets an attacker reuse a captured SAML login assertion to obtain multiple valid sessions for the same user. The advisory rates the severity as Moderate and says the fix is in version 8.5.0.

Technical summary

This is a race condition (TOCTOU, CWE-367) in the SAML assertion nonce check, enabling capture-replay (CWE-294). The nonce check and insert were not atomic and the saml_nonces.nonce column had a non-unique index, so two concurrent POST /saml/acs requests with the same captured assertion could both log in. Affected versions are stated as <= 8.4.1 and the patched version is 8.5.0, with fixes in commit 5b0a779c07 and a unique index migration that runs saml:clear_expired_nonces first. Preconditions are SAML SSO enabled with valid IdP metadata, an attacker holding one valid assertion, and the ability to submit it twice within a sub-second race window. No workaround is stated. Exploitation in the wild is not stated, and no CVE ID is listed (the advisory says No known CVE). Specific detection guidance is not stated.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/grokability/snipe-it/securit…/GHSA-w49g-f2jc-qxhg

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

Low

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

grokabilitySnipe-ITgrokability snipe-itSnipe-IT

Executive summary

Snipe-IT versions up to and including 8.6.2 can send LDAP login credentials, including the service account password and end-user passwords, in cleartext when an attacker on the network path interferes with the encrypted connection setup. The advisory rates the issue Moderate and recommends upgrading to version 8.6.3 or later, which fixes the flaw.

Technical summary

The flaw is a cleartext transmission of sensitive information issue (CWE-319) caused by an algorithm downgrade during negotiation (CWE-757). In Ldap::connectToLdap(), the return value of ldap_start_tls() was ignored, so when StartTLS failed, the code still performed an ldap_bind() in cleartext. Exploitation requires LDAP authentication enabled with ldap_tls = 1, and an attacker on the network path between Snipe-IT and the LDAP server, who answers the StartTLS request with a Protocol error or the server does not offer StartTLS. Affected versions are stated as <= 8.6.2 and the patched version is 8.6.3, fixed in commit 91ea06904f. The advisory does not state any workaround. It does not state whether exploitation has occurred in the wild, and no CVSS score is given. No detection guidance is provided beyond the stated exploitability conditions.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/grokability/snipe-it/securit…/GHSA-wx48-r3qh-jvh5

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

Present

Privileges required

None

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:P/PR:N/UI:N/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

grokabilitySnipe-ITgrokability snipe-itSnipe-IT

Executive summary

A flaw in Snipe-IT, an IT asset management application, lets users who have been deactivated sign in through Google single sign-on if their Google account is still active. The advisory rates it Moderate and says it affects versions 8.6.2 and earlier. The vendor recommends upgrading to version 8.6.3 or later.

Technical summary

The Google SSO callback in GoogleAuthController::handleGoogleCallback looks up the local user by username matching the Google account email and logs them in without checking the activated flag, which is a missing authorization check (CWE-285) and an improper authentication weakness (CWE-287). The advisory says exploitation requires Google SSO to be enabled, a deployment that uses email-as-username or an email-shaped username on the account, a deactivated account, and a still-live Google identity. Affected versions are stated as 8.6.2 and earlier, and the fix is in 8.6.3 (commit 5fea2fadc0), which checks the activated flag after lookup. The advisory does not state a CVE ID, a CVSS score, or any workaround other than upgrading. It does not state whether exploitation has been observed in the wild, and it gives no specific detection guidance.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/grokability/snipe-it/securit…/GHSA-xf36-wx48-5rqp

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

Present

Privileges required

None

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:P/PR:N/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

grokabilitySnipe-ITgrokability snipe-itSnipe-IT

Executive summary

Snipe-IT versions 8.6.3 and earlier have a flaw in the API asset-creation path that can store sensitive custom field values as plaintext when the caller lacks the permission needed to write encrypted fields. The advisory rates the issue Low and says it is fixed in version 8.7.0, which the vendor recommends installing.

Technical summary

The flaw is classified as missing encryption of sensitive data (CWE-311) and an always-incorrect control flow implementation (CWE-670) in Api\AssetsController::store. When the caller lacks assets.view.encrypted_custom_fields, the encryption call is skipped and the unencrypted value is written to the database column because the block lacks a continue statement, unlike the update path. The advisory states the affected versions are 8.6.3 and earlier and the fixed version is 8.7.0, with the fix in commit 34dafe9eed. The advisory does not state a CVE ID, a CVSS score, or any workaround. It also does not state whether exploitation has been observed or provide specific detection guidance beyond the description of the impact.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/grokability/snipe-it/securit…/GHSA-jh4g-98fm-jcqc

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

Low

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:N/VI:L/VA:N/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

grokabilitySnipe-ITgrokability snipe-itSnipe-IT

Executive summary

Edgeless Systems Contrast versions from 1.21.0 before 1.25.0 contain a flaw in the imagepuller component that can select the wrong registry configuration when both a concrete domain and its parent wildcard are configured. The advisory rates it Low, with a CVSS v3.1 base score of 2.6, and the vendor's fix is to upgrade to 1.25.0 or later.

Technical summary

The flaw is classified as CWE-1289, improper validation of unsafe equivalence in input, in the imagepuller registry selection logic. It is exploitable only when the registry pull configuration contains both a concrete domain (for example registry.example.com) and its parent domain as a wildcard (for example .example.com), which the advisory describes as very unlikely in practice. The advisory lists affected versions as 1.21.0 through 1.24.1 and the patched version as 1.25.0, with the fix made in pull request #2682. The CVSS v3.1 vector is AV:A/AC:H/PR:L/UI:N/S:U/C:L/I:N/A:N, giving a score of 2.6. The advisory states a workaround: replace the wildcard entry with a list of concrete entries. No detection guidance is stated, and no CVE identifier is stated (the advisory says No known CVE). Exploitation in the wild is not stated.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/edgelesssys/contrast/securit…/GHSA-w4p9-c7h8-jvgr

Impact if exploited

How it is exploited

Attack vector

Adjacent network

Attack complexity

High

Attack requirements

Present

Privileges required

Low

User interaction

None

CVSS:4.0/AV:A/AC:H/AT:P/PR:L/UI:N/VC:L/VI:N/VA:N/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

edgelesssysEdgeless Systemsedgelesssys contrastEdgeless Systems Contrast

Executive summary

Rclone, an open-source tool for copying and syncing files to cloud storage, has a credential exposure flaw in its Backblaze B2 backend. When a custom download host is configured, the live account authorization token and SSE-C encryption keys can be sent over cleartext or to another host during an HTTP redirect, and the advisory rates it Moderate with a CVSS 3.1 base score of 5.3. The advisory says the fix is to update to v1.75.2, with workarounds that include enforcing HTTPS on the download URL and using least-privilege application keys.

Technical summary

This is a cleartext transmission of sensitive information issue (CWE-319) in the rclone b2 backend, where the default Go HTTP client follows redirects from the custom --b2-download-url host. A same-host HTTPS-to-HTTP redirect causes the Authorization header to be resent in cleartext, and the SSE-C customer key headers are forwarded to any redirect target host. The advisory lists rclone versions before v1.75.2 and states affected versions as <= v1.75.0 with patched version v1.75.2. The attack vector is Adjacent with high complexity, no privileges or user interaction required, and CVSS 3.1 vector AV:A/AC:H/PR:N/UI:N/S:U/C:H/I:N/A:N. The Remediation section says to update to v1.75.1, which conflicts with the patched version v1.75.2 listed elsewhere in the advisory. Workarounds are to enforce an explicit https:// scheme on --b2-download-url, block plaintext redirects at the CDN, unset the download URL, and use scoped application keys. No CVE identifier is stated in the advisory, and no detection guidance or evidence of exploitation in the wild is stated.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/rclone/rclone/security/advis…/GHSA-gpw7-77cq-9f72

Impact if exploited

How it is exploited

Attack vector

Adjacent network

Attack complexity

High

Attack requirements

Present

Privileges required

None

User interaction

None

CVSS:4.0/AV:A/AC:H/AT:P/PR:N/UI:N/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

rcloneBackblazercloneBackblaze B2

Executive summary

A flaw in the rclone file sync tool lets an attacker who can create a specially named directory in the source tree change ownership and timestamps of a directory outside the destination the operator chose. The advisory rates it Moderate, with a CVSS score of 7.4 for the root-run case and a conservative estimate of about 6.8. The advisory recommends fixing the trimming logic so the name is checked after trimming, and it notes the affected versions are 1.75.1 and earlier, with 1.75.2 as the patched version.

Technical summary

This is a path traversal (CWE-22, CWE-59) in rclone's local backend, where translateLink trims the .rclonelink suffix from the OS path after the containment check, so validated paths can be shifted one directory up. It requires a local destination with -l and --metadata, and an attacker who can create a directory named .rclonelink in the copied tree. The advisory reports the escape affects metadata only (ownership, timestamps, xattrs) because content writes are blocked by os.Root hardening, and that a root-run case allows ownership change. It lists affected versions as <= 1.75.1 and patched version 1.75.2, but also states that v1.75.0 and master HEAD 4479a3b are affected by its own testing. The suggested fix is to trim the remote name before confinement and re-validate the final path. The advisory does not state a workaround, and it does not state any detection guidance.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/rclone/rclone/security/advis…/GHSA-55qp-jrwr-x827

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

High

Attack requirements

Present

Privileges required

None

User interaction

Passive

CVSS:4.0/AV:N/AC:H/AT:P/PR:N/UI:P/VC:N/VI:H/VA:N/SC:N/SI:H/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

rclonegithub.com/rclone/rclonerclone

Executive summary

Rclone versions from 1.52.0 before 1.75.2 contain a cross-site request forgery flaw in the pCloud backend's localhost OAuth setup callback. A malicious web page open in a user's browser during the pCloud authorization flow can silently point the user's pCloud remote at an attacker-controlled host, causing later file transfers to go to that host. The advisory rates it Moderate with a CVSS score of 6.8, and the fix is to upgrade to v1.75.2.

Technical summary

The flaw is a missing state check (CWE-352 cross-site request forgery) on rclone's localhost OAuth callback at 127.0.0.1:53682 for the pCloud backend, combined with trust in an unauthenticated hostname parameter (CWE-346). An attacker needs a victim to load a malicious page while rclone config or rclone authorize pcloud is running, and the advisory notes Chrome Private Network Access can block subresource requests while top-level navigation is not blocked. Affected versions are stated as >= 1.52.0 and <= 1.75.1, with patched version v1.75.2. The advisory does not state a workaround. Detection guidance beyond the proof of concept is not stated, and no CVE identifier is listed.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/rclone/rclone/security/advis…/GHSA-hvpr-p4pv-4f46

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

High

Attack requirements

Present

Privileges required

None

User interaction

Passive

CVSS:4.0/AV:N/AC:H/AT:P/PR:N/UI:P/VC:H/VI:H/VA:N/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

rclonepCloudrclonepCloud backend

Executive summary

The advisory says rclone, an open-source file synchronization tool, has an authentication bypass in its serve webdav feature affecting versions from 1.64.0 before 1.75.2. An unauthenticated remote attacker can learn whether arbitrary paths exist and whether they are files or folders, though file contents and writes are not reachable. The advisory rates it Moderate with a CVSS v3 base score of 5.3 and recommends upgrading to 1.75.2.

Technical summary

The flaw is missing authentication (CWE-306) caused by the basicAuth and MiddlewareAuthCustom HTTP middlewares in lib/http skipping all OPTIONS requests rather than only genuine CORS preflights. An unauthenticated network attacker can send OPTIONS requests to rclone serve webdav and read the Allow header, which differs for existing files, existing collections, and missing paths, enabling namespace enumeration. The advisory states the affected range as >= 1.64.0 and <= 1.75.1, with patched version 1.75.2. No workaround is stated. The advisory does not state whether the issue is exploited in the wild, and it gives no detection guidance beyond the proof of concept. The advisory says rclone serve http and other listed servers are not affected.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/rclone/rclone/security/advis…/GHSA-gj73-fh6v-92fj

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

None

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:L/VI:N/VA:N/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

rclonerclone

Executive summary

Rclone versions from 1.37 before 1.74.2 contain a server-side request forgery flaw that lets unauthenticated attackers, when the optional --rc-serve feature is enabled without RC authentication, redirect HTTP and NetStorage backend requests to hosts they choose. The advisory rates it Moderate with a recommended CVSS 3.1 score of 5.9. The vendor-side fix is to upgrade to version 1.74.2, which the advisory lists as the patched version.

Technical summary

This is a server-side request forgery (CWE-918) caused by a scheme-relative named-remote root, such as //host/path, that replaces the configured URL authority for the http and netstorage backends. Exploitation requires --rc-serve enabled, no RC HTTP authentication, a reachable RC listener, and knowledge of an existing http or netstorage remote name; the advisory notes the default bind is loopback and that --rc-no-auth is not set in the tested configuration. Affected versions are stated as 1.37 or later up to and including 1.74.1, with patched version 1.74.2. The advisory says the attacker can cause same-scheme HEAD and GET requests to attacker-selected hosts with response bodies returned, send configured HTTP headers to those hosts, and obtain time-bound NetStorage signatures, but no write primitive is demonstrated. A proposed origin-check patch is described, but the advisory does not state a workaround, and no detection guidance is given. Whether the issue is exploited in the wild is not stated, and no CVE identifier is listed on the advisory.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/rclone/rclone/security/advis…/GHSA-xqmq-v4wj-rw5m

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

High

Attack requirements

Present

Privileges required

None

User interaction

None

CVSS:4.0/AV:N/AC:H/AT:P/PR:N/UI:N/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

rclonencwrclone

Executive summary

Rclone versions before 1.75.2 contain a path traversal flaw in its WebDAV server (rclone serve webdav) that lets an unauthenticated remote client create, overwrite, and move files and directories above the folder that rclone is configured to serve. The advisory rates it High with a CVSS 3.1 base score of 7.5, and it describes the impact as integrity and availability rather than data disclosure. The vendor's fix is upgrading to version 1.75.2.

Technical summary

The flaw is a CWE-22 path traversal caused by unvalidated '..' path elements in WebDAV requests, such as MKCOL /. which plants a directory node whose backend path is '..'. The attack needs no authentication, which the advisory says is the default for rclone serve webdav, and requires two HTTP requests (an MKCOL followed by a PUT). The advisory says it applies when the served remote has a non-empty root and the backend joins root and remote before encoding, such as webdav, ftp, sftp, http, or memory, and that local and s3 backends are not affected. Affected versions are stated as <= 1.75.1 and the patched version as 1.75.2; the advisory says the lower bound is not established. Workarounds stated: running with --read-only blocks the attack (the advisory reports MKCOL returning 405), and the advisory proposes code fixes that add path checks. Detection guidance beyond the proof-of-concept request sequence is not stated, and no CVE identifier is stated in the advisory.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/rclone/rclone/security/advis…/GHSA-44pm-2q5p-fm63

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

None

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:N/VI:H/VA:N/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

rclonencwrclonencw rclone

Executive summary

The rclone tool's WebDAV backend can send a custom credential header to a different host when the configured WebDAV server issues a cross-host redirect, which could expose a reusable login credential. The advisory rates it Low, with a CVSS 3.7 score and a note that no CVE is assigned. The advisory states that affected versions are v1.56.0 through v1.75.1 and that v1.75.2 is the patched version.

Technical summary

This is an information disclosure flaw, classified as CWE-200 and CWE-522, in which operator-configured custom headers set via the WebDAV headers option are re-sent on cross-host HTTP 3xx redirects. Exploitation requires that an attacker can influence the server's redirect response, such as through an open redirect, a malicious or compromised endpoint, or an on-path position on a cleartext or same-host hop. The advisory states the fixed version is v1.75.2 and that the default configuration is affected, independent of the auth_redirect option. Standard Authorization headers are stripped by Go across registrable domains, but custom-named headers are not. The advisory does not state a workaround beyond avoiding reusable secret headers on remotes that may redirect to other hosts, and it does not state any detection guidance or known exploitation in the wild.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/rclone/rclone/security/advis…/GHSA-3rqf-7h69-gg5v

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

High

Attack requirements

Present

Privileges required

None

User interaction

None

CVSS:4.0/AV:N/AC:H/AT:P/PR:N/UI:N/VC:L/VI:N/VA:N/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

rclonerclone

Executive summary

A flaw in the stable self-update feature of rclone, a command-line file sync tool, lets an attacker who already controls the trusted release download path install a malicious replacement program, affecting versions 1.55.0 through 1.75.1. The advisory rates the issue High, with a CVSS 3.1 score of 7.5. The advisory recommends applying the proposed fix, which resolves the checksum only from the signed portion of the response, but it states that no upstream fixed release was available at the time of testing and the listed patched version is 1.75.2.

Technical summary

The vulnerability is an improper verification of a cryptographic signature (CWE-347). The verifier checks a PGP clearsigned SHA256SUMS block but then looks up the archive hash in the original response buffer, which can include unsigned bytes before the signed block, so an attacker-supplied unsigned checksum line is accepted. The attack requires control of authenticated responses from the stable release distribution path, such as origin, CDN, or TLS termination compromise, or interception with a trusted certificate, and user interaction in running stable selfupdate without --check. Affected versions are stated as >= 1.55.0 and <= 1.75.1, with patched version 1.75.2. The advisory states a proposed fix of resolving the hash from block.Bytes, which was tested but is not described as shipped upstream. Workarounds are not stated. Detection guidance beyond the proof-of-concept test is not stated.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/rclone/rclone/security/advis…/GHSA-w33h-3qgq-8mv8

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

High

Attack requirements

Present

Privileges required

None

User interaction

Passive

CVSS:4.0/AV:N/AC:H/AT:P/PR:N/UI:P/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

rclonencwrclone

Executive summary

The advisory says rclone serve s3 versions 1.75.0 before 1.75.2 contain an integer overflow that lets an unauthenticated network client bypass the configured memory limit for multipart uploads, which can exhaust memory on the host running the service. The severity is rated Moderate, and the advisory lists 1.75.2 as the patched version. Users should upgrade to 1.75.2 or later.

Technical summary

This is an integer overflow (CWE-190) combined with unbounded resource allocation (CWE-770) in the multipart reorder-memory accounting of rclone serve s3. An attacker who can reach the S3 listener opens two concurrent multipart PUT requests for the current part, each declaring 5,000,000,000,000,000,000 bytes, and withholds the bodies, so the signed int64 accounting field wraps negative and out-of-order parts later pass the capacity check. The advisory says affected versions are 1.75.0 up to but not including 1.75.2, with the patched version listed as 1.75.2. Exploitation needs no credentials when serve s3 runs without --auth-key or --auth-proxy, and authenticated deployments require an accepted S3 credential. The advisory states that no workaround is available beyond binding to a trusted interface, which it says does not fix the arithmetic, and that disabling streaming is not a safe mitigation. Detection guidance is not stated, and the CVSS v3.1 base vector given is AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:L, with an overall score shown as 5.3.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/rclone/rclone/security/advis…/GHSA-qjw8-8hmp-c3m6

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

None

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:N/VI:N/VA:L/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

rclonerclone

Executive summary

Rclone versions from v1.68.0 before v1.75.2 contain an authorization bypass in its S3 server when run with --auth-proxy, letting one authenticated tenant interfere with another tenant's in-flight multipart uploads. The advisory rates it Moderate with a CVSS score of 6.4, and the recommended fix is to upgrade to v1.75.2.

Technical summary

The flaw is CWE-639, an authorization bypass through a user-controlled key. The multipart operations UploadPart, CompleteMultipartUpload and AbortMultipartUpload look up uploads by upload ID alone and do not check the caller's tenant, so a tenant who learns another tenant's bucket, object and uploadID can inject data into, complete or abort that upload. The advisory says the attacker must be an authenticated tenant on a server using --auth-proxy with several keys, and must obtain the victim's tuple, which it demonstrates through ListMultipartUploads while owning a same-named bucket. Affected versions are stated as >=v1.68.0 <= 1.75.1, with patched version v1.75.2. The advisory does not state a workaround, and detection guidance is not stated. It notes that at-rest content is not read and the destination key is set by the victim at initiate.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/rclone/rclone/security/advis…/GHSA-35g6-9fjv-rmx5

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

High

Attack requirements

Present

Privileges required

Low

User interaction

None

CVSS:4.0/AV:N/AC:H/AT:P/PR:L/UI:N/VC:H/VI:L/VA:L/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

rclonencwrclone

Executive summary

The advisory says OpenRemote core (openremote-manager) before 1.31.0 lets a registered edge gateway write and delete assets in any realm, breaking tenant isolation. The advisory gives two different severity ratings: it reports the finding as High, with a base score of 9.6 (Critical) under CVSS v3.1, while the NVD record and the GitHub advisory list it as Low. The fix is to upgrade to version 1.31.0.

Technical summary

The flaw is an authorization bypass through a user-controlled key: inbound gateway events skip the standard client event authorization, and the only mapping of gateway asset IDs (mapAssetId) shifts the first two characters by a per-gateway offset, which is reversible by anyone who knows the gateway ID. The attacker needs a registered, enabled gateway's service-user credentials, which a realm administrator can obtain, and then can send crafted AttributeEvent or DELETE AssetEvent messages over the gateway session to change attributes or delete leaf assets in other realms. The advisory states the affected versions as less than 1.31.0 and the patched version as 1.31.0, and it says the issue is present in all releases up to 1.26.1. It does not state a workaround. The advisory includes a proof-of-concept that is a standalone re-implementation of the ID transform and does not run against a live system, and it states that a full dynamic test was not performed. No detection guidance is stated.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/openremote/core/security/adv…/GHSA-8pjc-wq2p-3xr4

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

High

Attack requirements

None

Privileges required

Low

User interaction

None

CVSS:4.0/AV:N/AC:H/AT:N/PR:L/UI:N/VC:N/VI:L/VA:N/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

openremoteopenremote coreopenremote-manager

Executive summary

OpenRemote core versions before 1.31.0 contain an authorization bypass that lets an authenticated user in one realm read active gateway tunnel information belonging to another realm. The advisory rates this Moderate (CVSS 5.0) and states that the fix is to upgrade to version 1.31.0. The attacker must already know or guess a gateway asset ID in the victim realm.

Technical summary

The flaw is an authorization bypass (CWE-639, CWE-284) in the REST endpoint GET /gateway/tunnel/{realm}/{id}, where getGatewayActiveTunnelInfos does not filter returned tunnels by realm, unlike its sibling methods. Any authenticated non-restricted user can exploit it over the network by passing their own realm in the path and a gateway ID from another realm, leaking target host, target port, and gateway routing data. Affected versions are stated as less than 1.31.0 (package io.openremote:openremote-manager), with the patched version given as 1.31.0. The advisory states no workaround. It also states that exploitation in the wild is not stated, and that the finding was confirmed by static source comparison only, not dynamically executed. Detection guidance beyond the proof-of-concept request pattern is not stated.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/openremote/core/security/adv…/GHSA-p5wh-9wx4-w5f3

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

Low

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:L/VI:N/VA:N/SC:L/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

openremoteopenremote coreopenremote-manager

Executive summary

OpenRemote versions before 1.31.0 contain an authorization bypass that lets a user with admin rights in one realm impersonate a gateway that belongs to a different realm. The advisory rates it Low, with a CVSS v4 base score of 2.1, and the patched version is 1.31.0.

Technical summary

The flaw is an authorization bypass (CWE-639, with CWE-862 also listed) in GatewayService, where gateway websocket events are routed using only the gateway ID and are not checked against the realm of the GatewayConnector. An attacker needs a realm admin account with WRITE:ADMIN in one realm and knowledge of a gateway ID in another realm; they then create a service user named gateway-GATEWAY_ID in their own realm and connect over websocket. The advisory says the attacker could create SSH tunnels during initial sync and overwrite descendant assets of the Gateway asset, though it cannot read existing assets. Affected versions are stated as less than 1.31.0 and the patched version is 1.31.0. No workaround is stated, and no detection guidance is stated. The advisory states no known CVE ID exists, and it does not state whether the flaw is exploited in the wild.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/openremote/core/security/adv…/GHSA-2j4x-qh6q-q8mm

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

High

Attack requirements

None

Privileges required

High

User interaction

None

CVSS:4.0/AV:N/AC:H/AT:N/PR:H/UI:N/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

openremoteopenremote coreopenremote-manager

Executive summary

An unauthenticated visitor can modify or delete any visitor-submitted question in phpMyFAQ versions before 4.1.10 (the advisory lists affected versions as up to and including 4.2.0-beta). The advisory rates it High at CVSS 7.5 under the default configuration, and Critical at 9.1 when the option to delete open questions after answering is enabled. The advisory recommends upgrading to 4.1.10, which is listed as the patched version.

Technical summary

The flaw is an object-level authorization bypass (CWE-639) in POST /api/faq/create, where the client-supplied openQuestionID is cast to an integer and used directly as the key for UPDATE or DELETE queries on the faqquestions table, with no ownership, visibility, or permission check. Exploitation requires no account or CSRF token, only solving the graphical CAPTCHA, and the advisory states no user interaction is needed. With records.enableDeleteQuestion false (the shipped default), the attacker can rewrite a question's category and answer link (CVSS 7.5, integrity impact only); with it true, the question is permanently deleted (CVSS 9.1). The advisory lists the affected versions as <= 4.2.0-beta and the patched version as 4.1.10. Workarounds are not stated beyond the suggested code fix, and no detection guidance is given. No CVE ID is assigned in the advisory (it says 'No known CVE').

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/thorsten/phpMyFAQ/security/a…/GHSA-9m55-g84r-x6v7

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

None

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:N/VI:H/VA:N/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

thorstenphpMyFAQthorsten phpMyFAQphpMyFAQ

Executive summary

phpMyFAQ versions before 4.1.10 contain a server-side request forgery flaw in the PDF export feature, which can make the server fetch external images referenced in FAQ answers. The advisory rates it Moderate with a CVSS overall score of 5.0, and the fix is to update to version 4.1.10.

Technical summary

The flaw is server-side request forgery in the PDF export feature, which fetches image URLs in FAQ answers during PDF generation; in the tested setup a loopback-only HTTP service was requested and its response embedded in the PDF. Exploitation requires that localhost be listed in records.allowedMediaHosts and that an attacker can create or modify FAQ content, and the advisory states the full chain was not proven to be achievable by an entirely unauthenticated attacker. Affected versions are stated as <= 4.1.9 and the patched version as 4.1.10. The advisory does not state a workaround, and the CVSS vector given is CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:L/I:N/A:N. No CVE identifier is listed (the advisory reports 'No known CVE'), and no detection guidance beyond the proof-of-concept steps is stated.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/thorsten/phpMyFAQ/security/a…/GHSA-8wmv-9c4g-p5mg

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

Low

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:L/VI:N/VA:N/SC:L/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

thorstenphpMyFAQthorsten phpMyFAQphpMyFAQ

Executive summary

The advisory says vLLM versions from 0.20.1 before 0.29.0 contain a flaw in which a single crafted request with an out-of-vocabulary stop token ID can crash the inference engine and interrupt service for other users. The advisory rates the issue Moderate with a CVSS v3.1 base score of 6.5, and the fix is to upgrade to 0.29.0 or later.

Technical summary

The weakness is improper input validation (CWE-129, with related CWE-248, CWE-400 and CWE-787) in stop_token_ids, which were checked only to be integers and not against the model vocabulary. A remote attacker who can send OpenAI chat or completion requests, or the legacy generate endpoint, with an out-of-vocabulary ID can trigger an assertion and terminate the engine on the default V1 runner, or write outside the logits buffer on the V2 runner. The CVSS vector is CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H, so low privileges are required. Affected versions are stated as greater than or equal to 0.20.1 and less than 0.29.0, and the patched version is stated as 0.29.0, with the fix in PR #54196 that rejects invalid IDs as request errors. The advisory does not state any workaround, and it gives no detection guidance or information about exploitation in the wild. No CVE identifier is listed (it states 'No known CVE').

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/vllm-project/vllm/security/a…/GHSA-w9r6-29q4-pvpr

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

Low

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

vllm-projectVllmvllm-project vllmvLLM

Executive summary

The advisory says vLLM versions from 0.8.5 before 0.31.0 contain a denial of service flaw. A single crafted API request using prompt_logprobs against a connector-cached prefix can crash the engine process, dropping all in-flight requests from other users on that engine. The advisory rates it High (7.7 CVSS as stated), and the patch replaces uninitialized memory with zero and negative infinity defaults; the advisory states the fix has not landed upstream yet.

Technical summary

The flaw is classed as use of uninitialized resource (CWE-908) leading to uncontrolled resource consumption (CWE-400) in LogprobsTensors.empty_cpu() in vllm/v1/outputs.py, which uses torch.empty(). Exploitation requires the V1 engine with a KVConnectorBase_V1 subclass that reports externally computed tokens, plus a request with prompt_logprobs (or logprobs with echo=true); the advisory lists network attack vector, low complexity, and low privileges. The proposed patch uses torch.zeros() and torch.full(-inf), and the advisory says a Fixes trailer should reference commit 3aee6573dc58. Workarounds stated are disabling KV connectors, stripping prompt_logprobs, logprobs and echo at a gateway, or applying the patch. The advisory lists affected versions as v0.8.5 onward through HEAD, and the patched version is written as >= 0.31.0. Detection guidance is not stated, and whether the fix has been released in a tagged version beyond the advisory's note that main is still affected is not stated beyond that.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Dropped extracted product not named in the source: PyPI vllm

Summarized from github.com/vllm-project/vllm/security/a…/GHSA-rhcx-5729-88vg

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

Low

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:H/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

vllm-projectVllmvllm-project vllmvLLM

Executive summary

A flaw in vLLM, an open-source system for serving large language models, lets API clients bypass a frame-count limit on inline video uploads. The advisory says it can cause excessive memory and CPU use and can crash or OOM-kill the serving process, and it is rated High. The fix rejects per-request media decoder overrides by default, and the advisory says deployments should update to version 0.31.0 or later.

Technical summary

The weakness is CWE-770, allocation of resources without limits or throttling, in VideoMediaIO.load_base64 for video/jpeg input. A client that can call the chat endpoint can set media_io_kwargs.video num_frames to -1 or a very large value, which bypasses the frame cap and makes the API frontend decode and stack every frame in memory. Affected versions are stated as 0.18.0 up to but not including 0.31.0, and the patched version is stated as 0.31.0 or later; the advisory notes the issue was confirmed on 0.19.1. The fix is PR #58830, which rejects non-empty per-request media_io_kwargs by default, and deployments that trust their clients can opt in with --trust-request-mm-kwargs. The advisory does not state a CVSS score, a CVE assignment beyond 'No known CVE', any workaround beyond the flag described, exploitation in the wild, or specific detection guidance.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/vllm-project/vllm/security/a…/GHSA-4m6g-h9mv-wqjc

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

None

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

vllm-projectVllmvllm-project vllmvLLM

Executive summary

A denial of service flaw in vLLM, an open-source engine for running large language models, lets a remote attacker crash the engine by sending crafted multimodal request data. The advisory rates it Moderate with a CVSS v3.1 base score of 5.9, and says the fix is to upgrade to vLLM version 0.30.0 or later.

Technical summary

The flaw is improper input validation (CWE-20) in the disaggregated multimodal /inference/v1/generate endpoint, which accepted client-supplied kwargs_data, mm_hashes, and mm_placeholders without checking that they describe a consistent prompt layout. Overlapping or repeated placeholder ranges can reach model code and crash the engine, as reproduced with Qwen2.5-VL-3B-Instruct. Affected versions are stated as greater than or equal to 0.20.0 and less than 0.30.0, and the patched version is stated as greater than or equal to 0.30.0, with the fix in PR #51898. The CVSS vector is CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:N/I:N/A:H, requiring no privileges or user interaction. No workaround is stated, no CVE ID is assigned, and no detection guidance or evidence of exploitation in the wild is stated.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/vllm-project/vllm/security/a…/GHSA-34qw-q2h2-jp3m

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

High

Attack requirements

Present

Privileges required

None

User interaction

None

CVSS:4.0/AV:N/AC:H/AT:P/PR:N/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

vllm-projectVllmvllm-project vllmvLLM

Executive summary

vLLM, an open-source system for serving large language models, has a flaw in its Nemotron-VL image processors that lets an unauthenticated remote user exhaust server memory and deny service to other users on the same instance. The advisory rates the issue High, and the fix is to upgrade to vLLM 0.25.0 or later.

Technical summary

The weakness is CWE-400, uncontrolled resource consumption. Two Nemotron-VL processor modules set Pillow's Image.MAX_IMAGE_PIXELS to None at import time, which disables Pillow's decompression-bomb guard for the whole process. An unauthenticated caller who can reach /v1/chat/completions on an instance that has imported either processor can submit a crafted image that decompresses to a very large size, potentially causing the worker to be killed by the out-of-memory handler. Affected versions are listed as 0.10.2 before 0.25.0 in the NVD description and as <= 0.8.5 in the GitHub advisory, which the advisory does not reconcile; the patched version is stated as 0.25.0 or later. The fix in PR #47010 removes the module-level assignments and adds a decoded pixel-count limit configurable through VLLM_MAX_IMAGE_PIXELS. No workaround beyond upgrading is stated, and no detection guidance is stated. No CVSS score is stated, and the advisory gives severity as High. No known CVE ID is listed in the GitHub text, and exploitation in the wild is not stated.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/vllm-project/vllm/security/a…/GHSA-mm38-5g96-7j6x

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

None

User interaction

None

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

vllm-projectVllmPillowvllm-project vllmvLLMPillow