CVE Vulnerability Catalog

A vulnerability exists in NGINX Plus and NGINX Open Source within the ngx_http_slice_module module. When the slice directive with unnamed regex captures is configured or during a background cache update, unauthenticated attackers can send requests causing uninitialized memory access in the NGINX worker process, leading to limited memory disclosure or a restart.

ERPNext prior to versions 15.111.0 and 16.22.0 is vulnerable to server-side template injection (SSTI) through a configuration field. An authenticated user with a standard operational role can exploit this flaw to disclose data outside their normal permission scope.

Metabase from version 1.57.0 to 1.57.19.1, 1.58.14.1, 1.59.10, and 1.60.4 contains a vulnerability that allows an attacker with access to database connection configuration to read arbitrary files from the Metabase server's filesystem. By adding unsafe JDBC parameters to a MySQL or MariaDB connection, the driver can read files from the Metabase host and expose their contents through queries or validation error messages.

A vulnerability in Vaultwarden prior to 1.36.0 allows an attacker to hijack a victim's account via SSO. The issue is due to missing verification of the IdP email_verified claim when linking an identity to an existing local account.

A vulnerability in Vaultwarden prior to version 1.36.0 allows an unauthenticated attacker to hijack a user session by manipulating the SSO authorization flow. The lack of binding the OAuth state parameter to the browser session and acceptance of attacker-controlled PKCE parameters enable triggering authentication at the identity provider and redeeming tokens.

A vulnerability in Tabby (formerly Terminus) prior to version 1.0.234 allows code execution via command injection when dropping files. The lack of neutralization of command substitution metacharacters ($(...) and `...`) enables an attacker to execute arbitrary code after the victim presses Enter.

A vulnerability in Penpot prior to version 2.15.0 allowed an authenticated file editor to reach internal-only endpoints via remote image import. The issue was due to missing destination filtering in the HTTP client used for image downloading.

Penpot before version 2.15.0 has a vulnerability in the ReplServer that bound to 0.0.0.0:4403 and exposed an unauthenticated /execute endpoint. This allowed anyone on the network to execute arbitrary JavaScript code on the server.

In Directus prior to version 12.0.0, when response caching is enabled, the cache-key derivation omits authorization context such as share tokens, roles, and policies. This allows different users or anonymous clients requesting the same URL to receive a cached response without permission re-evaluation.

In Directus prior to version 12.0.0, the SSRF protection on the file-import-from-URL feature can be bypassed using the address 0.0.0.0 because the system treats it as a keyword for local interfaces but never blocks the literal address. On Linux and macOS, connecting to 0.0.0.0 reaches localhost, allowing an authenticated user with file-upload rights to fetch internal services via the /files/import endpoint and retrieve the response as a downloadable file.

FastGPT version 4.15.0-beta4 has a vulnerability in reverse-call endpoints under /api/invoke/* that authenticate only via a JWT signed with a default constant secret 'token'. An unauthenticated attacker can self-sign an HS256 JWT and access cross-tenant user PII or write arbitrary content to chat files.

A vulnerability in FastGPT allows a low-privileged tenant user to access quotes or full-text content from another tenant's dataset by manipulating the initialId parameter in the POST /api/core/chat/record/getCollectionQuote endpoint. The issue affects versions from 4.14.17 to 4.15.0-beta5.

A vulnerability in Cursor editor before 03/31/2026 allowed attacker-controlled web content to connect to an unauthenticated local agent endpoint, enabling code execution in the cloud agent sandbox and access to files, repositories, environment variables, credentials, and GitHub App tokens. The issue was fixed on 03/31/2026 by requiring authentication for the relevant endpoint.

A vulnerability in the HTTP/2 profile on a BIG-IP virtual server allows an unauthenticated attacker to remotely cause increased memory utilization. This can lead to system performance degradation until the TMM process is forced or manually restarted.

An injection vulnerability exists in NGINX Ingress Controller when configured with CRDs or Ingress annotations. An authenticated attacker with write access can inject arbitrary NGINX configuration directives.

In Cloudreve prior to version 4.16.1, a WebDAV account scoped to a configured folder can bypass the restriction by sending paths containing encoded ".." characters (e.g., %2e%2e). The stripPrefix function in pkg/webdav/webdav.go joins the decoded request suffix to the account root without checking containment, allowing reading and listing files outside the folder, and for writable accounts also creating, overwriting, moving, or deleting them.

Cloudreve versions 4.12.0 through 4.16.1 issue OAuth access tokens without the client_id claim, causing the JWT verifier to not load token scopes into the request context. The RequiredScopes function treats such requests as non-scoped session authentication, allowing a low-scope OAuth token to call APIs requiring higher scopes, including file, share, workflow, user settings, WebDAV accounts, and potentially admin scopes.

A vulnerability in NGINX Plus and NGINX Open Source occurs when a map directive uses regex matching and a string expression references the map's regex capture variables before referencing the map output variable. An unauthenticated attacker can send crafted HTTP requests, causing a heap buffer overflow in the NGINX worker process, leading to a restart or potential code execution.

A permission control vulnerability in the file system allows unauthorized access to confidential data. Successful exploitation may compromise service confidentiality.

A vulnerability in ICU Scandinavia Boomerang allows an unauthenticated remote attacker to retrieve sensitive credential files (service accounts, SMTP) by requesting specific XML files from the webroot via static HTTP.