Flatpak is a Linux application sandboxing and distribution framework. Prior to versions 1.14.0 and 1.15.10, a malicious or compromised Flatpak app using persistent directories could access and write files outside of what it would otherwise have access to, which is an attack on integrity and confidentiality.

When persistent=subdir is used in the application permissions (represented as --persist=subdir in the command-line interface), that means that an application which otherwise doesn't have access to the real user home directory will see an empty home directory with a writeable subdirectory subdir. Behind the scenes, this directory is actually a bind mount and the data is stored in the per-application directory as ~/.var/app/$APPID/subdir. This allows existing apps that are not aware of the per-application directory to still work as intended without general home directory access.

However, the application does have write access to the application directory ~/.var/app/$APPID where this directory is stored. If the source directory for the persistent/--persist option is replaced by a symlink, then the next time the application is started, the bind mount will follow the symlink and mount whatever it points to into the sandbox.

Partial protection against this vulnerability can be provided by patching Flatpak using the patches in commits ceec2ffc and 98f79773. However, this leaves a race condition that could be exploited by two instances of a malicious app running in parallel. Closing the race condition requires updating or patching the version of bubblewrap that is used by Flatpak to add the new --bind-fd option using the patch and then patching Flatpak to use it. If Flatpak has been configured at build-time with -Dsystembubblewrap=bwrap (1.15.x) or --with-system-bubblewrap=bwrap (1.14.x or older), or a similar option, then the version of bubblewrap that needs to be patched is a system copy that is distributed separately, typically /usr/bin/bwrap. This configuration is the one that is typically used in Linux distributions. If Flatpak has been configured at build-time with -Dsystembubblewrap= (1.15.x) or with --without-system-bubblewrap (1.14.x or older), then it is the bundled version of bubblewrap that is included with Flatpak that must be patched. This is typically installed as /usr/libexec/flatpak-bwrap. This configuration is the default when building from source code.

For the 1.14.x stable branch, these changes are included in Flatpak 1.14.10. The bundled version of bubblewrap included in this release has been updated to 0.6.3. For the 1.15.x development branch, these changes are included in Flatpak 1.15.10. The bundled version of bubblewrap in this release is a Meson "wrap" subproject, which has been updated to 0.10.0. The 1.12.x and 1.10.x branches will not be updated for this vulnerability. Long-term support OS distributions should backport the individual changes into their versions of Flatpak and bubblewrap, or update to newer versions if their stability policy allows it. As a workaround, avoid using applications using the persistent (--persist) permission.

Apache Tomcat contains a path equivalence vulnerability that allows a remote attacker to execute code, disclose information, or inject malicious content via a partial PUT request. This vulnerability can be chained with CVE‑2026‑34486.

Accessibility. An authorization issue was addressed with improved state management.

Erlang Erlang/OTP SSH server contains a missing authentication for critical function vulnerability. This could allow an attacker to execute arbitrary commands without valid credentials, potentially leading to unauthenticated remote code execution (RCE). By exploiting a flaw in how SSH protocol messages are handled, a malicious actor could gain unauthorized access to affected systems. This vulnerability could affect various products that implement Erlang/OTP SSH server, including—but not limited to—Cisco, NetApp, and SUSE.

Roundcube Webmail before 1.5.10 and 1.6.x before 1.6.11 allows remote code execution by authenticated users because the from parameter in a URL is not validated in program/actions/settings/upload.php, leading to PHP Object Deserialization.

An earlier fix for an Inter-process Communication (IPC) vulnerability, CVE-2011-3079, added authentication to communication between IPC endpoints and server parents during IPC process creation. This authentication is insufficient for channels created after the IPC process is started, leading to the authentication not being correctly applied to later channels. This could allow for a sandbox escape through IPC channels due to lack of message validation in the listener process.

Last updated 25 August 2025

The iframe sandbox rules were not correctly applied to XSLT stylesheets, allowing an iframe to bypass restrictions such as executing scripts or navigating the top-level frame.

Last updated 24 July 2024

Last updated 24 July 2024

Buffer overflow in the nsXMLHttpRequest::AppendToResponseText function ...

Last updated 25 August 2025

Redis is prone to a (Debian-specific) Lua sandbox escape, which could result in remote code execution.

An elevation of privilege vulnerability exists when an attacker establishes a vulnerable Netlogon secure channel connection to a domain controller, using the Netlogon Remote Protocol (MS-NRPC), aka 'Netlogon Elevation of Privilege Vulnerability'.

It was found that a malicious HVM guest administrator can cause DoS, specifically prevent use of physical CPU for significant, perhaps indefinite period. When a benign exception occurs while delivering another benign exception, it is architecturally specified that these would be delivered sequentially. There are, however, cases where this results in an infinite loop inside the CPU, which (in the virtualized case) can be broken only by intercepting delivery of the respective exception.

When a guest sets up a hardware breakpoint covering a data structure involved in delivering #DB (Debug Exception), upon completion of the delivery of the first exception another #DB will need to be delivered. The effects slightly differ depending on further guest characteristics:

Guests running in 32-bit mode would be expected to sooner or later encounter another fault due to the stack pointer decreasing during each iteration of the loop. The most likely case would be #PF (Page Fault) due to running into unmapped virtual space. However, an infinite loop cannot be excluded (e.g. when the guest is running with paging disabled).

Guests running in long mode, but not using the IST (Interrupt Stack Table) feature for the IDT entry corresponding to #DB would behave similarly to guests running in 32-bit mode, just that the larger virtual address space allows for a much longer loop. The loop can't, however, be infinite, as eventually the stack pointer would move into non-canonical address space, causing #SS (Stack Fault) instead.

Guests running in long mode and using the IST for the IDT entry corresponding to #DB would enter an infinite loop, as the stack pointer wouldn't change between #DB instances.

If a host watchdog (Xen or dom0) is in use, this can lead to a watchdog timeout and consequently a reboot of the host. If another, innocent, guest, is configured with a watchdog, this issue can lead to a reboot of such a guest.

A privileged user inside guest could use this flaw to crash the host kernel resulting in DoS.

For KVM virtualisation, it only affects the AMD processor support, as for Intel it already intercepts the #DB exception.

Upstream KVM patch: ------------------- -> http://permalink.gmane.org/gmane.linux.kernel/2082332

References: ----------- -> http://www.openwall.com/lists/oss-security/2015/11/10/1

Netatalk before 3.1.12 is vulnerable to an out of bounds write in dsiopensess.c. This is due to lack of bounds checking on attacker controlled data. A remote unauthenticated attacker can leverage this vulnerability to achieve arbitrary code execution.

A flaw was found in ruby gem-rack. This flaw allows a malicious actor to craft requests that can cause shell escape sequences to be written to the terminal via rack's Lint middleware and CommonLogger middleware. This issue can leverage these escape sequences to execute commands in the victim's terminal.

It was discovered that the System.arraycopy() method has a race condition between verifying source elements and storing them. An untrusted Java application or applet could possibly use this flaw to trigger a Java Virtual Machine memory corruption.

It was discovered that the JPEG decoder did not properly handle certain input streams. An untrusted Java application or applet could possibly use this flaw to trigger a Java Virtual Machine memory corruption.

It was discovered that certain medialib operations do not properly validate that mlib and raster images correspond to each other. A remote attacker could possibly use this flaw to trigger a Java Virtual Machine memory corruption.

It was discovered that the ServiceLoader did not perform exception handling in a secure manner. An untrusted Java application or applet could possibly use this flaw to bypass security mechanisms and perform operations with full permissions.

Last updated 29 November 2024

FasterXML jackson-databind 2.x before 2.9.7 might allow remote attackers to conduct server-side request forgery (SSRF) attacks by leveraging failure to block the axis2-jaxws class from polymorphic deserialization.

Buffer overflow in micq client 0.4.6 and earlier allows remote attackers to cause a denial of service, and possibly execute arbitrary commands, via a long Description field.

Buffer overflow in BSD-based telnetd telnet daemon on various operating systems allows remote attackers to execute arbitrary commands via a set of options including AYT (Are You There), which is not properly handled by the telrcv function.

The line printer daemon (lpd) in the lpr package in multiple Linux operating systems authenticates by comparing the reverse-resolved hostname of the local machine to the hostname of the print server as returned by gethostname, which allows remote attackers to bypass intended access controls by modifying the DNS for the attacking IP.

Some functions that implement the locale subsystem on Unix do not properly cleanse user-injected format strings, which allows local attackers to execute arbitrary commands via functions such as gettext and catopen.

rpc.statd in the nfs-utils package in various Linux distributions does not properly cleanse untrusted format strings, which allows remote attackers to gain root privileges.

Buffer overflow in Canna input system allows remote attackers to execute arbitrary commands via an SRINIT command with a long user name or group name.

Xpdf, as used in products such as gpdf, kpdf, pdftohtml, poppler, teTeX, CUPS, libextractor, and others, allows attackers to cause a denial of service (infinite loop) via streams that end prematurely, as demonstrated using the (1) CCITTFaxDecode and (2) DCTDecode streams, aka "Infinite CPU spins."