🚨 CVE-2026-48797
Backpropagate is a Python library for fine-tuning large language models on a single GPU. In versions 1.1.0 and 1.1.1, the optional Reflex web UI exposes a training control plane without authentication: dataset upload, model load, training start/stop, multi-run orchestration, GGUF export, and HuggingFace Hub push. The CLI accepts two operator-facing flags intended as security controls: --auth user:pass — documented as "require HTTP Basic authentication on every request to the UI." and--share — documented as "expose the UI on a public address; requires --auth." When --auth user:pass is passed, the CLI prints Auth: enabled (user: <username>) to confirm to the operator that authentication is active, then exports BACKPROPAGATE_UI_AUTH=user:pass to the subprocess that launches the Reflex backend. The Reflex backend (backpropagate/ui_app/**) never reads BACKPROPAGATE_UI_AUTH. No authentication middleware is registered. No request-level guard runs. No WebSocket upgrade guard runs. Any client that reaches the bound port — local or remote, depending on whether --share is used — has full UI access. An inline comment at backpropagate/cli.py:1217-1218 in the v1.1.0 source documents the gap: "For Phase 1 the variable is exported but Reflex doesn't read it yet." This comment was internal-facing; the user-facing documentation (README, CHANGELOG, SHIP_GATE) advertised the contract as enforced. An attacker who reaches the bound port can read uploaded datasets, trigger arbitrary training runs against any local base models as well as read their paths, trigger HuggingFace Hub pushes and cause disk-fill DoS. This issue has been fixed in version 1.2.0. If developers cannot immediately upgrade to 1.2.0 run backprop ui with no flags so it binds to localhost, use SSH port-forwarding (ssh -L 7860:localhost:7860 <training-host>) instead of --share for remote access, and audit any host previously launched with --share, re-issuing any HF tokens used during those sessions.

🎖@cveNotify

🚨 CVE-2026-49268
A remote attacker can inject LDAP special characters into the Distinguished Name (DN) construction in DefaultLdapRealm class. User-supplied username input is directly concatenated into the LDAP DN template without any escaping of RFC 2253 special characters. This allows an attacker to manipulate the DN structure used for LDAP bind authentication, potentially bypassing authentication or impersonating other users.

This issue affects all Apache Shiro versions through 2.2.0, and 3.0.0-alpha-1 when using DefaultLdapRealm
Upgrade to Apache Shiro 2.2.1 or 3.0.0-alpha-2 or later, which fixes the issue.

🎖@cveNotify

🚨 CVE-2026-9591
Cross-site request forgery (CSRF) in NewsItemApiController in SimplCommerce prior to commit 6233d73e allows an unauthenticated remote attacker to create or modify news items as an administrator via a crafted form submitted to `/api/news-items`, due to missing anti-CSRF protection.

🎖@cveNotify

🚨 CVE-2026-35066
Dell PowerFlex Manager, version(s) [Versions], contain(s) an Improper Access Control vulnerability. A low privileged attacker with remote access could potentially exploit this vulnerability, leading to denial of service.

🎖@cveNotify

🚨 CVE-2026-47103
Python StateMachine versions 3.0.0 before 3.2.0 contains a remote code execution vulnerability that allows attackers to execute arbitrary code by supplying malicious SCXML documents containing crafted `<data expr="...">` attributes evaluated unsafely. The SCXMLProcessor passes attacker-controlled expression strings through a call chain ending in Python's built-in eval() without sandboxing, enabling arbitrary code execution in the context of the hosting process.

🎖@cveNotify

🚨 CVE-2026-49502
Dell PowerFlex Manager, version(s) [Versions], contain(s) an Improper Authentication vulnerability. An unauthenticated attacker with adjacent network access could potentially exploit this vulnerability, leading to Information disclosure, Information tampering, and Unauthorized access.

🎖@cveNotify

🚨 CVE-2026-53871
Hermes WebUI before 0.51.368 contains an authorization bypass vulnerability in the get_profile_cookie() function that accepts unauthenticated profile names from the hermes_profile cookie. An authenticated attacker can forge the hermes_profile cookie value to bypass profile-scoped authorization checks and access sessions, files, and resources across different profiles.

🎖@cveNotify

🚨 CVE-2026-11407
Pimcore CMS/DXP version 12.3.8 contains a sandbox bypass vulnerability that allows authenticated administrative attackers to execute arbitrary methods on PHP objects by exploiting empty checkMethodAllowed() and checkPropertyAllowed() implementations in the custom Twig SecurityPolicy. Attackers can supply malicious Twig templates through the DataObject ClassDefinition Layout\Text component to perform arbitrary file reads, execute arbitrary database queries, and potentially achieve remote code execution via PHP object gadget chains, with the pimcore_* function wildcard further broadening the bypass to all Pimcore Twig functions.

🎖@cveNotify

🚨 CVE-2026-12529
A security vulnerability has been detected in SourceCodester CET Automated Grading System with AI Predictive Analytics 1.0. Affected is an unknown function of the file /index.php of the component Student Self-Registration Endpoint. The manipulation leads to improper access controls. Remote exploitation of the attack is possible.

🎖@cveNotify

🚨 CVE-2026-32682
When NGINX Gateway Fabric is configured using GRPCRoutes, an authenticated, remote attacker with permission to create or modify GRPCRoute resources can cause the NGINX Gateway Fabric control plane to terminate by sending undisclosed GRPCRoute configurations containing backendRef filters.


Note: Software versions which have reached End of Technical Support (EoTS) are not evaluated.

🎖@cveNotify

🚨 CVE-2026-48817
Starlette is a lightweight ASGI framework/toolkit. In versions 1.0.1 and below, when dispatching a request, HTTPEndpoint selects the handler by lowercasing the HTTP method and looking it up as an attribute with getattr, without restricting the lookup to a known set of HTTP verbs. When an HTTPEndpoint subclass is registered through Route(...) without an explicit methods= argument, the route does not constrain the method and every method reaches the endpoint. If a non-standard HTTP method whose lowercased name matches an attribute on the endpoint subclass reaches the endpoint, that attribute is invoked as if it were a request handler. An attacker can use this to reach methods that were never meant to be HTTP handlers, such as internal helpers, without the authorization checks applied by the intended public handler. An application (including Starlette-based frameworks like FastAPI) is affected if it registers an HTTPEndpoint subclass via Route(...) without explicitly setting methods=, and that subclass includes extra methods named like non-standard HTTP verbs that take one request argument and return a response. This issue has been fixed in version 1.1.0.

🎖@cveNotify

🚨 CVE-2026-48823
Shaarli is a personal bookmarking service. Versions 0.16.1 and prior contain a stored Cross-Site Scripting (XSS) vulnerability in the tag filtering functionality of Shaarli. An authenticated user can inject arbitrary JavaScript into the tags field when creating a bookmark (Shaare). The malicious payload is stored and later executed when users interact with the "Filter by tag" search feature on the homepage. User-supplied input in the tags field is not properly sanitized or output-escaped before being rendered in the tag filtering interface. When a bookmark is created with a malicious payload inside the tag field, the payload is stored in the database. Later, when a user searches using the "Filter by tag" functionality on the homepage, the application renders matching tags dynamically. If the tag value contains HTML with JavaScript event handlers, it is injected into the DOM. This impacts anyone interacting with the "Filter by tag" search functionality, administrators and privileged users. This issue has been fixed in version 0.16.2.

🎖@cveNotify

🚨 CVE-2026-50107
When NGINX Plus or NGINX Open Source is configured as the data plane for NGINX Gateway Fabric, an injection vulnerability exists in the NGINX configuration generator component of NGINX Gateway Fabric. User-supplied string values from the NginxProxy Custom Resource Definition (CRD) access log format setting are rendered directly into NGINX configuration templates without sanitization or escaping. An authenticated attacker with permission to create or modify these CRDs may craft values that inject arbitrary NGINX configuration directives. This is a control plane issue; there is no data plane exposure from the vulnerability trigger itself.


Note: Software versions which have reached End of Technical Support (EoTS) are not evaluated.

🎖@cveNotify

🚨 CVE-2026-54387
Tinyproxy through 1.11.3, fixed in commit ff45d3b, fails to reconcile conflicting Content-Length and Transfer-Encoding: chunked headers, forwarding both verbatim to the backend while using Content-Length to determine how many request body bytes to consume. Remote attackers can desynchronize the proxy and backend parser state, allowing injection of arbitrary HTTP requests to the backend to enable cache poisoning, access control bypass, and request hijacking.

🎖@cveNotify

🚨 CVE-2026-55201
Evil-WinRM through 3.9, fixed in commit 6ecd570, contains a path traversal vulnerability in the download_dir() function that allows a rogue or compromised remote Windows server to write files outside the intended download directory by returning filenames with traversal sequences from Get-ChildItem command output that are passed unsanitized to File.join(). Attackers controlling the remote server can exploit this to overwrite sensitive client-side files such as SSH authorized_keys or shell configuration files, achieving persistent access or privilege escalation on the client machine.

🎖@cveNotify

🚨 CVE-2026-48821
Shaarli is a personal bookmarking service. Versions 0.16.1 and prior contain a DOM-based Cross-Site Scripting (XSS) vulnerability in the Thumbnail Synchronizer feature. When an administrator runs the thumbnail update process, malicious bookmark titles are returned via an AJAX response and inserted into the DOM using innerHTML without proper sanitization. The issue originates from the interaction between the backend thumbnail update endpoint and the frontend JavaScript responsible for rendering update progress. On the backend, the ThumbnailsController::ajaxUpdate method returns bookmark data formatted using the 'raw' formatter. This includes the unescaped bookmark title in the JSON response. On the client side, the script thumbnails-update.js processes this AJAX response and dynamically updates the progress interface. Administrators using the thumbnail synchronization feature are affected and exploitation could lead to session hijacking, privilege escalation, backdoor injection and full compromise. This issue has been fixed in version 0.16.2.

🎖@cveNotify

🚨 CVE-2026-48979
PHP Standard Library (PSL) is set of APIs covering async, collections, networking, I/O, cryptography, terminal UI, etc. In versions 6.1.0, 6.1.1 and 6.2.0, the Psl\H2\ServerConnection does not validate that the total bytes received in DATA frames match the content-length header declared in the HEADERS frame, allowing request smuggling. This is in violation of RFC 9113 §8.1.1. A malicious client is able to send more DATA bytes than declared, smuggling additional content past application-level size limits and send fewer DATA bytes than declared and close the stream early, causing applications that trust the declared length to behave incorrectly.
The vulnerability is only reachable for consumers using Psl\H2\ServerConnection directly to accept untrusted client traffic. Consumers of documented high-level PSL APIs are not affected. This issue has been fixed in versions 6.1.2 and 6.2.1.

🎖@cveNotify

🚨 CVE-2026-49133
Typemill before 2.24.0 contains a path traversal vulnerability that allows authenticated attackers with Author-level privileges to read arbitrary files outside the content directory by supplying traversal sequences in the path query parameter passed to Storage::getFile() with an empty folder argument. Attackers can bypass traversal-prevention controls in Storage::getFolderPath() to access sensitive files.

🎖@cveNotify

🚨 CVE-2026-12530
Improper neutralization of argument delimiters in the install_packages() method in AWS Bedrock AgentCore Python SDK versions >= 1.1.3 and < 1.6.1 might allow a remote authenticated user to execute arbitrary commands within the Code Interpreter sandbox via crafted package name arguments.



To mitigate this issue, users should upgrade to version 1.6.1.

🎖@cveNotify

🚨 CVE-2026-48820
CakePHP is a rapid development framework for PHP. In versions 4.5.11 and earlier, 4.6.0 through 4.6.3, 5.0.0 through 5.1.6, 5.2.0 through 5.2.12, and 5.3.0 through 5.3.5, View::_getElementFileName() does not check that the resolved element path is within the application/plugin view template paths. When element names are created with specifically crafted user-supplied data this weakness can be leveraged to include other PHP files on the server. Patched releases are available in 5.3.6, 5.2.13, 5.1.7, 4.6.4, and 4.5.11.

🎖@cveNotify

🚨 CVE-2026-48990
joserfc is a Python library that provides an implementation of several JSON Object Signing and Encryption (JOSE) standards. In versions 1.3.4 through 1.6.5, joserfc accepts oversized RFC7797 b64=false JWS payloads without applying JWSRegistry.max_payload_length, which can lead to resource exhaustion. The normal JWS compact and flattened JSON paths reject payloads above the configured payload-size limit with ExceededSizeError. The RFC7797 unencoded payload paths do not make the same check. A valid b64=false compact or flattened JSON JWS can therefore deserialize successfully with a payload larger than JWSRegistry.max_payload_length. Applications that accept lower-trust JWS values and rely on joserfc to reject oversized token content during verification have a moderate availability risk. This issue has been fixed in version 1.6.7.

🎖@cveNotify