vLLM is an inference and serving engine for large language models (LLMs). From 0.5.5 until 0.23.1rc0, integer truncation of tensor dimensions in vLLM's GGUF dequantize kernels (csrc/quantization/gguf/gguf_kernel.cu) causes partial tensor processing. The output tensor is allocated at full size via torch::empty (uninitialized memory), but the dequantize CUDA kernel processes only a truncated number of elements. The unfilled portion of the output tensor retains whatever was previously in GPU memory. In multi-tenant inference deployments, this residual GPU memory may contain tensor data from other users' inference requests, constituting information disclosure. This vulnerability is fixed in 0.23.1rc0.
### Summary
``DigestAuthMiddleware`` can send an authentication response after following a cross-origin redirect.
### Impact
If the client follows a redirect (the default option) to an attacker controlled domain, the attacker may be able to extract the auth digest.
This likely requires an open redirect vulnerability or similar on the target domain for an attacker to be able to execute. Further, the attacker is only receiving the digest, so should only be able to extract the user's credentials if the cryptography is weak or there is some kind of password reuse.
### Workaround
Disable ``follow_redirects`` if this is a concern.
-----
Patch: https://github.com/aio-libs/aiohttp/commit/38d16060037e1bfcd6d677abababa3c2a4bb58fa
### Summary
The contents of files that are specified by [`server.fs.deny`](https://vite.dev/config/server-options#server-fs-deny) can be returned to the browser on Windows.
### Impact
Only apps that match the following conditions are affected:
- explicitly exposes the Vite dev server to the network (using `--host` or [`server.host` config option](https://vitejs.dev/config/server-options.html#server-host))
- the sensitive file exists in the allowed directories specified by [`server.fs.allow`](https://vite.dev/config/server-options#server-fs-allow)
- either of:
- the sensitive file exists in an NTFS volume
- the dev server is running on Windows and the sensitive file exists in a volume that 8.3 short name generation is enabled (it is enabled by default on system volumes)
### Details
Vite’s dev server denies direct access to sensitive files through `server.fs.deny`, including entries such as `.env`, `.env.*`, and `*.{crt,pem}`. However, on Windows, the deny logic does not correctly normalize NTFS ADS path forms before access checks are applied.
Because of this, requests such as `/.env::$DATA?raw` are treated as allowed paths, while Windows resolves them to the original file's default data stream.
Similar to that, Windows allows accessing a file using a different name with the 8.3 short name compatibility feature. Vite did not reject accessing files via them.
### PoC
```bash
$ npm create vite@latest
$ cd vite-project/
$ npm install
$ npm run dev
```
Access via browser at `http://localhost:5173/.env::$DATA?raw`
<img width="388" height="129" alt="deecc1315123883cfd0f9c26a002845a" src="https://github.com/user-attachments/assets/895c6012-4e2e-4a35-babb-69bbf3ee7170" />
Example expected result:
- `/.env::$DATA?raw` returns the contents of `.env`
- `/tls.pem::$DATA?raw` returns the contents of `tls.pem`
The vulnerability arises when the system fails to properly validate the 'email' field during the authentication process, allowing unverified or fake email addresses to be accepted. This lack of validation enables the creation of user accounts with fake email addresses, facilitating the mass creation of fraudulent accounts. Successful exploitation of this vulnerability could allow an authenticated attacker to carry out various attacks, such as mass spam distribution, system abuse, or bypassing user controls, thereby compromising the security and integrity of the system.
Vulnerability involving the exposure of sensitive data provided without adequate protection. The API exposes email and phone number data from the ‘email’ and ‘telefon’ fields. This vulnerability is also present in the local database, as it contains accessible sensitive information such as data on minors and municipal users. Successful exploitation of this vulnerability could allow an unauthenticated remote attacker to gain access to sensitive information and data.
@microsoft/kiota-http-fetchlibrary provides TypeScript libraries for Kiota-generated API clients. In versions 1.0.0-preview.97 through 1.0.0-preview.101, `@microsoft/kiota-http-fetchlibrary`'s `RedirectHandler` is documented as stripping `Authorization` and `Cookie` from cross-origin redirect targets, but the default `scrubSensitiveHeaders` callback in `RedirectHandlerOptions` uses case-sensitive property deletion (`delete headers.Authorization`, `delete headers.Cookie`) on a headers object that `FetchRequestAdapter.getRequestFromRequestInformation` has already lower-cased. The delete therefore targets keys that do not exist, the scrub is a no-op, and any Bearer token or Cookie attached by a kiota-generated SDK is forwarded to an attacker-controlled host across a 30x redirect. This is reachable in the default middleware chain (`MiddlewareFactory.getDefaultMiddlewares`) with no custom configuration, and applies to every kiota-generated TypeScript SDK that uses `BaseBearerTokenAuthenticationProvider` or any other authentication provider that sets the `Authorization` request header. Version 1.0.0-preview.102 patches the issue.
Statamic is a Laravel and Git powered content management system (CMS). Prior to 5.73.23 and 6.20.0, an authenticated Control Panel user could view metadata and content for resources they don't have permission to view, including entries, assets, users, roles, groups, and other configured resources. Depending on the resource, this could expose titles, custom field values, entry content, asset metadata, and the existence of users, roles, and groups. No data could be modified. This has been fixed in 5.73.23 and 6.20.0.
The GridTime 3000 GNSS Time Server leaks the access token in the URL parameters of some endpoints.
This issue affects GridTime 3000: from 1.0r0.03 through 1.1r0.0.
Exposure of sensitive information to an unauthorized actor in Cost Management Interactive Experiences allows an unauthorized attacker to disclose information over a network.
The Appointment Booking Calendar plugin for WordPress is vulnerable to Sensitive Information Exposure in versions up to, and including, 1.4.01. This is due to insufficient authorization and missing per-calendar ownership checks in the cpabc_appointments_calendar_load2() function, which is reachable via the cpabc_calendar_load2=1 query parameter in wp-admin and only checks is_admin() && current_user_can('edit_posts'), a capability available to Contributor-level users and above. This makes it possible for authenticated attackers with Contributor-level access and above to supply an arbitrary calendar ID via the id parameter and extract customer booking information, including email addresses, names, phone numbers, booking times, and comments, from any calendar managed by the plugin.
The Kadence Blocks — Page Builder Toolkit for Gutenberg Editor plugin for WordPress is vulnerable to Sensitive Information Exposure in all versions up to, and including, 3.7.5 via the editor_assets_variables. This makes it possible for authenticated attackers, with contributor-level access and above, to extract the site's connected Kadence account license key, license owner email, api_key, api_email, and license domain from the browser console by inspecting window.kadence_blocks_params.proData. Exploitation requires only that an administrator has previously connected a valid Kadence license; the full credential bundle is then readable by any Contributor-level user from the block editor client context without any server-side request manipulation.
The FireBox Popups – Increase Sales and Grow Your Email List plugin for WordPress is vulnerable to Sensitive Information Exposure in all versions up to, and including, 3.1.7 via the 'form_id' parameter. This makes it possible for unauthenticated attackers to extract download a full CSV export of all form submissions — including any personally identifiable information submitted by users — for any arbitrary form_id.
Steeltoe.Management.Endpoint versions prior to 4.2.0 and Steeltoe.Management.EndpointCore versions prior to 3.4.0 contain a vulnerability where sensitive configuration values, such as connection strings with embedded passwords, are not properly redacted in the /actuator/env endpoint responses. This occurs because the default sanitization list does not cover common connection string patterns, exposing sensitive information to unauthorized actors. Versions 4.2.0 and 3.4.0 include patches to address this issue. Mitigations include upgrading to fixed versions, removing the env endpoint from actuator exposure, adding connection string patterns to the sanitization keys, and requiring authorization on actuator endpoints.
CVE-2026-47340 is a vulnerability in Apache DolphinScheduler before version 3.4.2 that allows authenticated users to access alert instances linked to alert groups for which they lack permission. This exposure of sensitive information does not affect system integrity or availability but compromises confidentiality. The issue is fixed in version 3.4.2, and users are advised to upgrade to this version to remediate the vulnerability.
CVE-2026-12117 is an improper access control vulnerability in the social login connection endpoint of Devolutions Server version 2026.2.0. It allows an authenticated vault member to enumerate social login entry metadata that they should not be authorized to access via a crafted API request. The vulnerability has a medium severity with a CVSS score of 4.3. No official patch or remediation guidance has been provided yet by the vendor.
CVE-2026-8386 is an information exposure vulnerability in the WP Go Maps WordPress plugin before version 10.0.10. The plugin's public single-marker REST endpoint does not filter markers by approval state, allowing unauthenticated users to access marker data that administrators have not approved for public display. This data may include personally identifiable information (PII) in address and description fields, as well as geographic coordinates. The vulnerability has a medium severity with a CVSS score of 5.3.
The WP Go Maps WordPress plugin before 10.0.10 does not properly enforce the marker approval filter on the admin-ajax fallback for its datatables route, allowing unauthenticated visitors to retrieve marker records that the site owner has not approved for public display, including their title, category, address and description fields.
Nezha Monitoring is a self-hostable, lightweight, servers and websites monitoring and O&M tool. From version 2.0.0 to before version 2.0.14, private services (`EnableShowInService: false`) are enumerable via per-server endpoints, leaking name and timing data. This issue has been patched in version 2.0.14.
Nezha Monitoring is a self-hostable, lightweight, servers and websites monitoring and O&M tool. From version 1.4.0 to before version 2.0.9, any authenticated non-admin member can connect to the server-status WebSocket and receive telemetry for all servers, including servers owned by other users. The normal server list API filters objects by HasPermission, but the WebSocket stream treats the presence of any authenticated user as authorization for the full unfiltered server list. This issue has been patched in version 2.0.9.
Parse Server is an open source backend that can be deployed to any infrastructure that can run Node.js. From version 9.8.0 to before version 9.9.1-alpha.5, apps that enable MFA and deny get on the _User class via Class-Level Permissions could expose sensitive user data through the /login and /verifyPassword endpoints. These endpoints re-fetch the user through the access-controlled query pipeline (CLP, protectedFields, auth-adapter sanitizers) before responding. When that re-fetch was denied by the _User get permission, the server fell back to the raw database row, exposing raw authData (including MFA TOTP secrets and recovery codes) and fields hidden by protectedFields (when protectedFieldsOwnerExempt is false). /verifyPassword is the most severe: with only a username and password (no session or MFA token), an attacker who knows a victim's password could retrieve their MFA secret and recovery codes, defeating the second factor. This issue has been patched in version 9.9.1-alpha.5.