Threat briefs

The TranslatePress plugin for WordPress is vulnerable to unauthenticated stored XSS through improper sanitization of comment data, allowing attackers to inject persistent malicious scripts.

The One User Avatar | User Profile Picture WordPress plugin versions 2.5.4 and earlier contain a stored XSS and RCE vulnerability via improper file type validation in the avatar upload process.

An unauthenticated stored XSS vulnerability in LiteSpeed Cache versions 7.8.1 and below allows attackers to inject malicious scripts into WordPress comments by bypassing wp_kses input sanitization.

The Forminator Forms WordPress plugin (up to v1.57.0.1) is vulnerable to unauthenticated Stored Cross-Site Scripting (XSS) via the Rich-Text Textarea field, allowing malicious script execution in the context of victim browsers.

NUMail contains an unauthenticated OS command injection vulnerability allowing remote attackers to execute arbitrary system-level commands on affected servers.

A public proof-of-concept exploit has been released for CVE-2025-4611, a medium-severity vulnerability in the Slim Seo product that allows low-privilege remote exploitation.

CVE-2026-23489 is a blind remote code execution vulnerability in the GLPI Fields plugin (<= 1.23.2) that allows authenticated attackers to execute arbitrary PHP code via the dropdown-generation feature.

Publicly available proof-of-concept exploits targeting CVE-2023-2640 and CVE-2023-32629 allow unprivileged users to gain root access on Ubuntu systems by bypassing permission checks in the overlayfs subsystem.

Detection of CVE-2023-22515 exploitation attempts targeting Atlassian Confluence servers by sending crafted HTTP requests to specific vulnerable endpoints, potentially leading to unauthorized access and privilege escalation.

Changeweb Unifiedtransform version 2.0 contains an incorrect access control vulnerability allowing unprivileged users to modify student records via the /students/edit/{id} endpoint.

An improper authorization vulnerability (CVE-2026-71504) in Dolibarr prior to version 24.0.0 allows authenticated users to overwrite the credentials of any account via the Members REST API.

Unitree G1 EDU firmware versions 1.5.2 and earlier contain a chained vulnerability in the BLE GATT server and WiFi provisioning stack, allowing unauthenticated proximate attackers to achieve root-level remote code execution.

Unitree G1 EDU firmware through 1.5.2 is susceptible to unauthenticated remote code execution allowing root-level command injection via a chained exploit targeting the WebRTC-to-DDS bridge, hardcoded credentials, and path traversal in the knowledge upload API.

The openssl_encrypt library before 1.4.9 is vulnerable to key substitution attacks due to improper fingerprint validation when loading identities from local identity.json files.

An improper input validation vulnerability (CVE-2026-54721) in the Silverstripe userforms module allows authenticated attackers to achieve remote code execution by injecting malicious payloads into the email subject field.

A vulnerability in LiteLLM, tracked as CVE-2024-4786, allows remote authenticated attackers to circumvent established security controls.

JFrog Artifactory suffers from a path traversal vulnerability that allows an authenticated user to write files to unauthorized locations on the server by manipulating remote-repository configurations.

Roundup of Microsoft security advisories published in July 2026.

The NLTK library's TweetTokenizer is vulnerable to a ReDoS attack due to an unbounded regular expression, allowing unauthenticated attackers to trigger CPU exhaustion.

NLTK versions prior to 3.10.3 are vulnerable to remote code execution due to improper validation of dotted names during the unpickling of transition-parser models.

The openssl-encrypt Python package before version 1.4.9 is vulnerable to a symlink-following flaw in its verify-usb functionality, allowing attackers with physical access to removable drives to achieve arbitrary code execution via crafted __pycache__ files.

OpenRemote versions prior to 1.28.0 contain an information disclosure vulnerability in the Notification REST API that allows authenticated administrators to access sensitive data across all system realms.

A vulnerability in Baserow's Application Builder allows unauthenticated attackers to bypass permission checks and retrieve sensitive data by leveraging improperly handled access control logic.

The execute_ruby tool fails to sanitize the pseudo-terminal library, allowing attackers to escape the Ruby sandbox and execute arbitrary system commands via spawn entry points.

The get-html-skeleton tool contains an SSRF vulnerability via insufficient URL validation, allowing remote callers to exfiltrate cloud instance metadata or internal credentials.

The mcp-http-server package in UI-TARS-desktop exposes unauthenticated MCP transports globally, allowing remote attackers to execute arbitrary system commands and access the filesystem.

The Telnyx MCP server improperly bound its HTTP listener to all interfaces and failed to enforce authentication, allowing unauthenticated remote attackers to trigger tool dispatch and leak internal API credentials.

The mcp-router CLI versions prior to 0.6.3 default to binding the MCP aggregator service to all network interfaces without mandatory authentication, exposing fronted MCP servers to unauthorized remote access.

An out-of-bounds read vulnerability (CVE-2025-2399) in Mitsubishi Electric CNC Series controllers allows remote attackers to trigger a denial-of-service condition via crafted packets sent to TCP port 683.

Xiiaozet LK100W devices running firmware prior to v2.1.240 are vulnerable to multiple high-severity flaws, including OS command injection and authentication bypass, which could allow remote attackers to achieve full device compromise.