A vulnerability in JetBrains Datalore before version 2026.4 allows improper owner validation, which enables unauthorized users with limited privileges to read issues and runs from other workspaces. This flaw does not allow modification or denial of service but exposes some confidential information across workspace boundaries.
A vulnerability in JetBrains Datalore versions before 2026.4 allows notebook code to execute commands in the sidecar container via the mount broker. This could lead to high impact on confidentiality, integrity, and availability of the affected system.
A critical vulnerability in JetBrains Datalore versions before 2026.4 allows view-only collaborators to execute code within a notebook's computation environment. This flaw could lead to complete compromise of confidentiality, integrity, and availability of the affected system.
A vulnerability in actix-http's encoding::Decoder allows unbounded decompression of compressed input streams, which can lead to excessive memory and CPU consumption. This can cause denial of service when processing untrusted gzip, deflate, Brotli, or Zstandard compressed bodies. The issue affects versions prior to 3.20.1. Fixed in actix-http 3.20.1 and actix-web 4.18.2, where limits on decompression are enforced.
CVE-2026-123288 is a stored cross-site scripting (XSS) vulnerability in JetBrains Datalore versions before 2026.4. The issue arises from notebook charts via the Vega editor link, allowing malicious scripts to be stored and potentially executed. This vulnerability has a medium severity rating with a CVSS score of 5.4. It was fixed in version 2026.4.
Red Hat has released security updates for the Red Hat build of Keycloak 26.4.17 images used on OpenShift Container Platform. These updates address multiple vulnerabilities including account takeover, denial of service, information disclosure, request smuggling, cryptographic bypasses, and authorization bypasses. The fixes cover a wide range of issues affecting authentication, SAML processing, token validation, and underlying libraries such as Bouncy Castle and Netty. The update aligns the containerized image with the standalone product release and is intended for on-premise or private cloud deployments.
Saxy's XML encoder writes the target and the content of a processing instruction into the output document verbatim, without validation or escaping. The encoder builds the construct by concatenating both fields between <? and ?>, so neither field is checked against the constraints the XML specification places on it.
Because ?> terminates a processing instruction, a target or content value containing that sequence closes the construct at a position the attacker chooses. Everything the attacker writes after it is emitted as ordinary markup at document level, which allows arbitrary elements, attributes, and further processing instructions to be injected into the serialized document. Appending a matching <? keeps the encoder's own terminator well-formed, so the result still parses and the injection is silent rather than a visible corruption. The target is additionally unconstrained by the XML Name production, which permits injection before the remainder of the processing instruction.
The result is XML injection into a serialized document: the tree the application asks Saxy to encode and the tree a downstream parser reads back are no longer the same. Character data, attribute values, and comments are all escaped by the encoder, so processing instructions are the one construct that passes user bytes through untouched. Since Saxy.XML.processing_instruction/2 is a documented public builder, an application inherits the flaw without misusing the API. The consequences depend on what consumes the XML: forged records in a feed or export read by another system, injected stylesheet or rendering directives, or active content where the output is served to a browser as XML or SVG.
This issue affects saxy: from 0.7.0 before 1.7.0.
Node.js HTTP client undici versions 7.15.0 through 7.30.1 and 8.0.0 through 8.12.3 improperly verify Subresource Integrity (SRI) metadata by splitting the integrity string only on single spaces instead of all ASCII whitespace. This parsing flaw allows an attacker who controls the response body to bypass integrity checks if the SRI digest is preceded by whitespace such as a tab, resulting in acceptance of tampered content without error. The issue is fixed in versions 7.30.2 and 8.12.4, and users should upgrade accordingly or avoid whitespace around integrity values as a workaround.
A flaw was found in D-Bus. With a privileged Monitoring client attached (for example dbus-monitor or busctl monitor), an unprivileged local user can cause dbus-daemon to abort if a driver-originated broadcast such as NameOwnerChanged or ActivatableServicesChanged cannot be delivered to a recipient (for example due to a deny rule or an exceeded outgoing message queue). Driver broadcasts still use serial 0; when delivery fails, composing the monitor error reply asserts reply_serial != 0 and aborts the daemon.
The GutenKit WordPress plugin versions before 2.5.5 improperly sanitize custom CSS saved against posts, allowing users with Author role and above to inject JavaScript that executes in visitors' browsers. This vulnerability affects sites running WordPress versions below 7.0.
CVE-2026-92856 is a vulnerability in the SAML brokering component of the Red Hat build of Keycloak 26.6. The flaw allows Keycloak to ignore ProxyRestriction security instructions from an upstream identity provider, enabling users to access downstream services even when such access is explicitly forbidden. This bypasses intended identity delegation boundaries and allows unauthorized session establishment in downstream services. Exploitation requires a specific SAML brokering configuration and the presence of ProxyRestriction conditions from a trusted upstream provider. The vulnerability has a moderate severity rating with a CVSS score of 4.9.
CVE-2026-123269 is an out-of-bounds read vulnerability in the libvmdk library, specifically in the function libvmdk_descriptor_file_read_header within the Descriptor File Parser component. The vulnerability requires local access to exploit and has publicly available exploit code. A patch has been identified to address this issue.
Multiple vulnerabilities have been identified in JetBrains YouTrack, a web-based platform for project management, issue and bug tracking, helpdesk, and workflow management. No specific details about the nature, impact, or affected versions of these vulnerabilities are provided.
Python filelock versions from 3.30.0 up to 4.2.1 on POSIX platforms have a race condition vulnerability in StrictSoftFileLock and AsyncStrictSoftFileLock. This occurs because the coordination directory is validated but the claims directory is accessed by pathname, allowing a local user to swap in a symbolic link during lock acquisition. This can lead to claim files being created in unintended directories or cause lock timeouts. Additionally, inspecting or recovering locks can disclose file names or allow a blind force_break to let two processes enter a critical section simultaneously. The issue is fixed in version 4.2.1.
Pillow, a Python imaging library, has a resource allocation vulnerability in EPS rendering for versions 2.3.0 through 12.3.0. The issue arises because scaled image dimensions are not validated before invoking Ghostscript, leading to excessive CPU and disk usage when processing untrusted EPS or PostScript files with an application-selected scale greater than one. Although decompression-bomb protection exists, it only checks the output after Ghostscript completes rendering, allowing resource consumption beyond limits. This vulnerability is fixed in version 13.0.0.
LibreOffice includes a dBASE database driver, which reads dBASE tables and their .ndx index files when a folder of dBASE files is used as a data source. A heap buffer overflow existed when an index page was read. The number of entries recorded in a page was not checked against the number of entries the page could hold, which comes from a different field of the index header. A crafted index could also make a page refer back to itself, so that reading it recursed without end. In fixed versions the index header and the entry count of each page are limited to what an index page can hold, and a page that refers back to one of its parent pages is not followed.
FreeSWITCH versions prior to 1.11.2 contain an authentication bypass vulnerability in their inbound ICE handling. The software accepted ICE connectivity checks without verifying the STUN MESSAGE-INTEGRITY attribute, allowing a network peer to spoof connectivity checks, redirect media traffic, and potentially eavesdrop or inject media on unencrypted legs. This affects all WebRTC and SIP legs that negotiate ICE. The issue is fixed in version 1.11.2 by enabling ice_verify_message_integrity.
CVE-2026-116571 is a stack-based buffer overflow vulnerability in FreeSWITCH prior to version 1.11.2. It occurs in the rtmp_rtp2rtmpH264() function of the mod_rtmp module when processing oversized SPS and PPS parameter sets from an H.264 call bridged to an RTMP endpoint. The vulnerability can cause FreeSWITCH to abort and terminate all active calls. The mod_rtmp module is disabled by default in stock builds. This issue is fixed in FreeSWITCH version 1.11.2.