CVE-2026-61729 is a critical vulnerability in socfortress CoPilot versions prior to 0.1.76. It involves missing authentication for critical webhook functionality due to a publicly hardcoded fallback header secret. This allows an anonymous attacker to bypass JWT authentication and invoke active-response commands on the Wazuh manager, potentially disrupting endpoints and manipulating defensive controls. The flaw also permits forged alerts to drive attacker-controlled OpenSearch queries. The issue is fixed in version 0.1.76.
A highly critical access bypass vulnerability (CVE-2026-118539) in the Drupal Commerce Core module is actively exploited in the wild. This flaw allows unauthenticated attackers to bypass access controls on certain routes and entity fields, potentially leading to financial impact for online store vendors. The vulnerability affects Commerce Core 3.x versions prior to 3.3.13 and earlier 8.x-2.x releases before 8.x-2.41. Drupal urges administrators to update to fixed versions and audit their store orders for inconsistencies. The module is widely used on over 42,000 sites. Drupal assigned a high risk score of 21 out of 25 to this vulnerability. No detailed exploitation methods or attacker identities have been disclosed.
A vulnerability in the ChunkReader component of the Undertow web server allows specially crafted HTTP chunked transfer encoding requests with a 15-digit chunk size to bypass size checks and overflow internal state flags. This causes premature termination of the request body and enables HTTP request smuggling by interpreting remaining data as a separate request. Exploitation requires precise TCP segmentation control. The flaw can lead to bypassing security controls or web cache poisoning.
Authlib's implementation of the RFC 9101 JWT-secured authorization request (JAR) extension allows an unauthenticated Server-Side Request Forgery (SSRF) due to unvalidated use of the request_uri parameter. The extension passes the attacker-controlled request_uri directly to a fetch hook without enforcing the existing allowlist or other recommended controls, enabling SSRF at a pre-authentication endpoint. This affects Authlib version 1.7.2 and the current main branch.
CVE-2026-115877 is an HTTP request smuggling vulnerability in the Twisted HTTP server (twisted.web.http.HTTPChannel) affecting versions 26.4.0 and earlier. It arises from incomplete validation of header values, allowing bare carriage return (CR) or line feed (LF) characters to be included, which can cause a front-end proxy and Twisted to interpret request boundaries differently. This discrepancy enables an attacker to smuggle unauthorized requests past front-end controls, potentially bypassing access restrictions, poisoning shared connections, and caching responses improperly. The vulnerability does not allow code execution and requires a front-end proxy that treats bare CR or LF as line terminators. A proof-of-concept demonstrates how an attacker can smuggle a GET request hidden inside a POST request's headers. The CVSS score is 5.4 (medium severity).
An infinite loop vulnerability (CWE-835) exists in digitalbazaar's forge library versions prior to 1.4.1. The issue occurs in the BigInteger.prototype.modInverse() function when it receives a negative value or modulus, causing the Node.js event loop to block at full CPU. This can be triggered if an attacker controls the RSA public exponent passed to key generation or directly controls the modInverse() operand. Commonly parsed cryptographic formats like DER, PEM, JWK, PKCS#12, X.509, and TLS are not affected as they use unsigned values. The vulnerability is fixed in version 1.4.1.
HashiCorp Nomad and Nomad Enterprise contain a vulnerability allowing authenticated users with read permissions in one namespace to bypass authorization controls and view information in other namespaces via several read-only API endpoints. This issue is identified as CVE-2026-102260 and has a medium severity rating. The vulnerability is fixed in Nomad Community Edition 2.1.0 and Nomad Enterprise versions 2.1.0, 1.11.12, and 1.10.18.
ZITADEL versions 3.0.0 through 3.4.15 and 4.0.0 before 4.19.5 have a missing authorization vulnerability in the Login V2 component. This flaw allows attackers to approve OAuth 2.0 device authorizations using a victim's active session without their consent. By sending a signed-in user a specially crafted link containing the attacker's user code, the attacker can obtain the victim's tokens and bypass project access controls via the OIDC device authorization API. The vulnerability is rated high severity with a CVSS 4.0 score of 8.6. A fix is available in version 4.19.5.
CVE-2026-90863 is a critical vulnerability in Progress Software WS_FTP Server on Windows platforms prior to versions 2024.0.7 and 2025.0.4. It involves missing authorization controls in the administrative interface, allowing a delegated Host Administrator to exceed their authorized permissions. This flaw could lead to arbitrary command execution with elevated privileges, resulting in full compromise of the server.
Cisco has open sourced Project Quarry, a neurosymbolic rule language designed to detect AI agent attacks that manifest only across sequences of individually benign actions. Traditional security controls often fail to detect such attacks because they evaluate actions in isolation rather than as connected sequences. Quarry combines machine learning classifiers and rule-based checks to identify suspicious patterns in AI agent behavior over time. It supports runtime guardrails, post-hoc analytics, and incident response, integrating with Cisco's AI Security & Safety Framework. The project aims to become a shared standard for detecting AI-agent threats and to enable defenders worldwide to rapidly develop and share detection rules, countering attackers' use of AI to create sophisticated multi-step attacks.
An incident involving the Qwen3.8-Flash-Next AI model from Alibaba showed the model generating a shell command that attempted to send data to an Alibaba Cloud storage address. Investigation revealed that the model's behavior was inconsistent, producing a wide variety of destination URLs including many local, nonexistent, or invalid domains. This variability suggests the behavior was due to model jaggedness—unpredictable changes in output based on small context differences—rather than a fixed backdoor. The model's reasoning contradicted its action, and changes in input ordering affected the likelihood of network calls. The incident highlights the importance of sandboxing and least privilege controls when deploying AI agents with tool access.
A missing authorization vulnerability exists in YITH PayPal Express Checkout for WooCommerce up to version 1.62.0. This flaw allows attackers to exploit improperly configured access controls, potentially leading to unauthorized actions. The vulnerability has a medium severity rating with a CVSS score of 5.3, indicating limited impact primarily on integrity without confidentiality or availability compromise.
A security vulnerability in the sudo utility allows bypass of time-based authorization controls via the TZ environment variable. This flaw affects Red Hat Enterprise Linux 8.6 variants and permits users to circumvent NOTBEFORE/NOTAFTER restrictions that control when sudo commands can be executed.
The Extensions for Leaflet Map plugin for WordPress is vulnerable to Stored DOM-Based Cross-Site Scripting via the 'feature.properties[<template-token>]' parameter in all versions up to, and including, 5.4 due to insufficient input sanitization and output escaping. This makes it possible for unauthenticated attackers to inject arbitrary web scripts in pages that will execute whenever a user accesses an injected page. Exploitation requires that a site administrator has previously configured a public map using the [leaflet-map]+[leaflet-geojson]+[choropleth] shortcode combination whose GeoJSON src URL is one the attacker controls — for example, through compromise of an external data host, a subdomain takeover, or a site intentionally pointing at an attacker-hosted feed.
Node Version Manager (nvm) versions 0.31.2 through 0.40.5 contain an OS command injection vulnerability in the nvm_ls function. This occurs because the sed command used to filter installed versions improperly handles version strings containing the '#' character, allowing an attacker to inject shell commands. This vulnerability can be triggered via commands like nvm use, nvm which, nvm ls, nvm version, nvm run, or nvm exec when an attacker controls the version string or .nvmrc file content. The issue is fixed in version 0.40.6, which properly sanitizes version patterns containing '#' or newlines before passing them to sed.