A denial of service vulnerability exists in the joi library's email validation rule when configured to validate multiple addresses. The inefficient regular expression used for splitting input strings can be exploited by a remote attacker submitting inputs with long runs of whitespace, causing excessive CPU consumption and stalling the Node.js event loop. This affects Red Hat Hardened Images RPMs that embed joi for payload validation in various web dashboards and containerized services. No fix is currently available that meets Red Hat's criteria for ease of use and deployment.
A vulnerability (CVE-2026-116591) was found in the klauspost compress library's Zstandard (zstd) decompression implementation used in Red Hat Hardened Images RPMs. The flaw is an integer handling error in the decompression logic that leads to an out-of-bounds read and an unhandled panic, causing the application to crash and resulting in a denial of service (DoS). This affects applications decompressing untrusted zstd inputs. No direct workaround exists within the library itself, but filtering untrusted inputs or restricting decompression to trusted clients can mitigate risk until patched versions are applied.
The banks library versions prior to 2.4.3 contain a remote code execution vulnerability due to unsafe use of importlib.import_module and getattr on attacker-controlled import_path values in Tool JSON objects parsed from template bodies. This allows an attacker who can influence prompt content to execute arbitrary Python code in the hosting process. The vulnerability is distinct from a prior sandboxing fix and remains exploitable in version 2.4.2. A fix is available in version 2.4.3.
OpenAI research using an unreleased AI model discovered and exploited 948 distinct Linux kernel privilege escalation vulnerabilities in the kernel's master branch, generating working exploits and patches. The research demonstrated that unpatched Linux systems with low-privileged user access can be escalated to root by well-resourced attackers. OpenAI has submitted hundreds of patches and is working with the community to remediate these flaws. The company warns that outdated Linux systems should be considered effectively compromised if accessed by low-privileged users. Mitigations include applying patches promptly and using sandboxing and isolation technologies.
OpenHuman versions prior to 0.57.44 have a missing authentication vulnerability in the local /ws/dictation WebSocket endpoint. This endpoint does not validate clients or browser origins, allowing unauthenticated clients, including malicious web pages, to submit audio and trigger resource-intensive speech-to-text processing. Additionally, earlier versions accumulate audio in an unbounded buffer, enabling continuous streaming that can exhaust process memory and degrade or crash the application. The issue is fixed in version 0.57.44.
CVE-2026-107891 is a high severity vulnerability affecting multiple specific versions of the musl C library. The vulnerability has a CVSS 3.1 base score of 8.1, indicating high impact on confidentiality, integrity, and availability. No detailed description or technical specifics of the flaw are provided. There is no information about known exploitation in the wild or available patches.
This security update for python-Pillow addresses multiple vulnerabilities including decompression bomb bypasses, denial of service via infinite loops and memory exhaustion, heap buffer overflows, use-after-free, out-of-bounds reads, and potential code execution. The flaws affect various image format parsers and processing functions, such as MPO, DCX, BMP, PDF, TIFF, JPEG, GIF, and others. These issues can lead to denial of service, memory corruption, information disclosure, or remote code execution. The update fixes these issues to improve the security and stability of the Pillow library.
CVE-2026-76892 is a high-severity vulnerability in the tsparticles TypeScript library prior to version 4.3.1. The issue occurs in the HTTP mode of @tsparticles/mcp-server where JSON.parse is called outside a request-local error boundary in the POST /mcp handler. An unauthenticated attacker can send a malformed JSON body that triggers a SyntaxError, causing the Node.js process to terminate and the MCP service to become unavailable until restarted externally. This denial-of-service condition is fixed in version 4.3.1.
CVE-2026-63655 is a high-severity vulnerability in the tsparticles TypeScript library prior to version 4.3.1. The HTTP mode of @tsparticles/mcp-server accepts unauthenticated POST requests and accumulates request-body chunks in memory without size limits, allowing remote attackers to exhaust Node.js process memory and cause service termination. The vulnerability does not affect confidentiality or integrity. It is fixed in version 4.3.1.
CVE-2026-115473 is an out-of-bounds read vulnerability in the cesanta mongoose embedded web server and network library. Versions prior to 7.22 improperly read bytes from short path buffers in the mg_path_is_sane function, leading to one-byte heap out-of-bounds reads. This can be triggered by unauthenticated remote requests to mg_http_upload or mg_http_serve_dir with crafted short paths. On systems without memory protection, this may expose adjacent memory bytes or affect path acceptance, and instrumented builds may crash. The issue is fixed in version 7.22.
CVE-2026-113780 is a critical out-of-bounds write vulnerability in the Espressif IoT Development Framework (esp-idf) affecting versions prior to 5.5.6, and certain 6.0.x and 6.1.x versions before 6.0.4 and 6.1.1 respectively. The flaw occurs in the esp_tee trap entry code, where a caller-controlled register frame is saved via a REE-supplied stack pointer without proper validation, allowing a local unprivileged REE application to write into TEE-exclusive memory. This breaks the isolation boundary between the Rich Execution Environment (REE) and Trusted Execution Environment (TEE) on specific Espressif devices, potentially enabling code execution, denial of service, or system crashes. The issue affects both secure-service ecall and interrupt entry paths and has been fixed in esp-idf versions 5.5.6, 6.0.4, and 6.1.1.
A stack-based buffer overflow vulnerability exists in the lwIP DNS resolver component of Espressif's ESP-IDF framework. The flaw occurs when the DNS resolver copies multiple cached IP addresses into a single-element buffer without bounds checking, leading to potential stack memory corruption. This can cause device reboots or, without stack protection, allow an attacker to overwrite stack variables and potentially redirect control flow. The vulnerability affects multiple ESP-IDF versions from 5.2.3 up to but not including fixed versions 5.3.7, 5.4.5, 5.5.6, and 6.0.3. The issue is rated high severity with a CVSS score of 7.4. Fixed versions are available for most affected branches except the end-of-life 5.2 line.
FreeRDP versions before 3.32.2 have a heap buffer overflow vulnerability in the transport_default_read_pdu function. This occurs when handling partial Early User Authorization reads during PROTOCOL_HYBRID_EX CredSSP. A malicious RDP server can exploit this by splitting the four-byte EUA result into short reads, causing an out-of-bounds write beyond the allocated 16 KiB buffer and resulting in client denial of service.
Traefik versions 2.x before 2.11.60 and 3.x before 3.7.16 have a vulnerability where ForwardAuth subrequests and Ingress NGINX auth-url requests are sent through a shared auth-server connection pool. This can expose connection-bound NTLM or Negotiate authentication tokens. An unauthenticated attacker can reuse another user's authenticated connection and impersonate that user to access backend services.
Red Hat has issued an important security update for webkit2gtk3 addressing multiple vulnerabilities in the Skia graphics library used by chromium-browser. These include use-after-free, cross-origin data disclosure, information disclosure, uninitialized resource usage, integer overflow, out-of-bounds read, memory corruption, and improper input validation issues. The update affects Red Hat Enterprise Linux 9 across various architectures and product variants.
The Red Hat Cost Management Metrics Operator for OpenShift has a security advisory addressing multiple vulnerabilities, including CVE-2026-6368. This operator collects OpenShift usage data by querying Prometheus hourly and uploads metrics to Red Hat's Cost Management service. The advisory references several CVEs with critical severity, including a notable out-of-bounds write vulnerability in the PCRE2 library (CVE-2026-103111) that could lead to arbitrary code execution or denial of service in applications using vulnerable JIT compilation of regular expressions. The advisory does not explicitly list affected versions or confirm patch availability for CVE-2026-6368 but provides guidance to apply all relevant errata and operator updates. No known exploits in the wild are reported.
A vulnerability in the jackson-dataformat-cbor library's CBOR parser allows an attacker to cause a denial of service via an algorithmic complexity issue. The parser accumulates semantic tags in a list that grows by a fixed increment, causing quadratic CPU and garbage collection overhead when processing a long run of tags. This leads to thread exhaustion and potential service disruption. The issue affects multiple versions and has a high severity rating.