| CVE |
Vendors |
Products |
Updated |
CVSS v3.1 |
| Apache Wicket 10.10.0, 10.11.0 and 9.24.0 do not set the Content-Security-Policy headers on a page whose render request handler is wrapped in more than one IRequestHandlerDelegate. The framework unwraps only a single delegate level when deciding whether a request renders a page, so with a deeper chain, for example an application-provided delegating request mapper combined with CryptoMapper, the page is served without any CSP header even though CSP is enabled, and no error or warning is raised. The nonce-based protection Wicket applies to inline scripts and styles, together with any other CSP directives configured by the application, is thereby lost for those pages, exposing them to content injection the policy was meant to prevent. Applications on stock Wicket that do not wrap request handlers themselves are not affected. Versions before 10.10.0 and 9.24.0 handled nested delegates correctly. Users are recommended to upgrade to version 10.12.0 or 9.25.0, which fix the issue. |
| Incorrect order of security checks in Apache Wicket.
org.apache.wicket.protocol.http.ResourceIsolationRequestCycleListener obtains the page a request targets before it consults its resource isolation policies, because those policies are passed that page. Obtaining the page instantiates it, and instantiating a page runs the configured org.apache.wicket.authorization.IAuthorizationStrategy. An application that refuses an unauthenticated visitor by throwing org.apache.wicket.RestartResponseAtInterceptPageException therefore throws it from inside the isolation listener, before any policy has run. That exception replaces the request handler, so the policies are never consulted and a request that came from another origin is not refused.
The exception's constructor stores the URL of the request it interrupted, so that Component#continueToOriginalDestination() can resume it once the visitor has signed in. The stored URL is the one the policies would have refused. On sign in Wicket redirects to it, and that redirect is same-origin, so the policies allow it and the listener runs in the newly authenticated session. The whole URL is stored, its query string included, so an ordinary org.apache.wicket.markup.html.form.Form is resumed with the field values the original request carried in its query string and Form#onSubmit() runs with values another origin chose. A form POST is not needed for this; a plain top-level navigation carries them.
An application is affected where it registers ResourceIsolationRequestCycleListener, which is opt-in and is not registered by default, and where it refuses unauthorized access with RestartResponseAtInterceptPageException followed by continueToOriginalDestination(). No application override is involved on that path: IAuthorizationStrategy#isInstantiationAuthorized() documents throwing that exception, org.apache.wicket.authroles.authentication.AuthenticatedWebApplication throws it and org.apache.wicket.authroles.authentication.panel.UsernamePasswordPanel continues to the original destination, and org.apache.wicket.authorization.strategies.page.SimplePageAuthorizationStrategy in wicket-core installs the same pattern. A visitor who is already signed in is not affected, because no interception occurs. Request-body parameters stored alongside the intercepted URL are not applied to the resumed form: the resumption is a redirect, so the resumed request is a GET, and form components read their values from the query string on a GET.
As a workaround, refuse unauthorized access with org.apache.wicket.RestartResponseException instead, which does not store the original destination; visitors are then not returned to the page they asked for after signing in.
This issue affects Apache Wicket: from 9.1.0 through 9.24.0, from 10.0.0 through 10.11.0.
Releases before 9.1.0, which is every 8.x release and 9.0.0, are not affected, because the resource isolation this issue concerns was introduced in 9.1.0. They are not protected either. Their only cross-origin defence is org.apache.wicket.protocol.http.CsrfPreventionRequestCycleListener, which works out where a request came from by reading its Origin header and falling back to its Referer header. Neither header need arrive: a browser does not send Origin on the plain GET a link uses, and the document on the other origin decides through its own referrer policy whether a Referer reaches the application at all. That mechanism therefore never made the guarantee the Fetch Metadata based policy makes, and on those releases a listener invocation from another origin is not reliably refused in the first place. Anyone on a release before 9.1.0 who wants this protection should upgrade to 9.x or 10.x and register ResourceIsolationRequestCycleListener.
Users are recommended to upgrade to version 9.25.0 or 10.12.0, which fix the issue. |
| Insufficient session expiration in Apache Wicket.
org.apache.wicket.protocol.ws.WebSocketSettings#newWebSocketRequest() wraps the WebSocket upgrade request in an org.apache.wicket.protocol.ws.api.ServletRequestCopy, which is kept for the life of the connection and used to build the request for every frame that arrives on it. Copying the request is deliberate and is documented as such, because the container recycles the upgrade request once the handshake is over. The container session is copied along with it, into an org.apache.wicket.protocol.ws.api.HttpSessionCopy holding the session attributes as they stood when the connection was opened. That second copy is not necessary, because a session, unlike a request, is not discarded at the end of a request. Every frame afterwards resolves its Wicket session through ISessionStore#lookup(), which reads it out of the copy, so the connection is detached from the live container session for as long as it stays open.
Ending the session through the container therefore does not end the connection. After an idle timeout, an HttpSession#invalidate() call made from a filter or from a security framework, a container-managed logout, or a session replacement, the connection goes on resolving the org.apache.wicket.Session instance that existed when it was opened, together with that session's attributes and, for applications built on wicket-auth-roles, the signed-in state held on that same instance, so an authorization decision taken inside a WebSocket callback is taken against an identity that has already been revoked. The copy is also what the connection writes through: a Session#invalidateNow() issued from inside a WebSocket callback invalidates the copy and leaves the container session valid, and session attributes set over the connection are written to the copy rather than to the session.
An application is affected where it uses Wicket's native WebSocket support on pages or shared resources that require an authenticated session, and where sessions are ended by the container rather than by Wicket. Session#invalidateNow(), and AuthenticatedWebSession#invalidate() and #signOut() which call it, remain effective for a WebSocketBehavior on a page: they clear the page manager, and the connection can then no longer resolve the page it was opened for. A WebSocketResource is not limited in that way, because a frame naming a resource is served against a throwaway page and never consults the page manager, so such a connection outlives Wicket's own sign-out as well. Some containers close WebSocket connections themselves when the HTTP session they were opened under is destroyed, in which case the connection ends regardless; that is a property of the container rather than of Wicket. Applications that do not use the native WebSocket support are not affected.
This issue affects Apache Wicket: from 8.0.0 through 8.19.0, from 9.0.0 through 9.24.0, from 10.0.0 through 10.11.0. Older, unsupported releases from 6.0.0 onwards are also affected. Users are recommended to upgrade to version 8.20.0, 9.25.0 or 10.12.0, which fix the issue. |
| A flaw was found in Red Hat Enterprise Linux 9 kernel builds. An incomplete backport of the CVE-2026-46099 fix in RHEL 9.6 through 9.10 left a local_bh imbalance in seg6_input_core(), where the drop path returns with bottom halves still disabled. This can lead to softirq starvation and denial of service on systems with SRv6 tunnels configured. Upstream and RHEL 10 are not affected. |
| Prefab UI's Generative UI sandbox runs caller-supplied Python inside Pyodide, Python compiled to WebAssembly, in a Deno subprocess. The feature is documented as isolation for untrusted code, including LLM-written UI. From prefab-ui 0.14.0 through 0.20.2 the launcher starts Deno with unrestricted --allow-read, --allow-write, and --allow-net, and the runner leaves the Deno API on the JavaScript global object. Pyodide's js module proxies that global, so the Python can call Deno.readTextFileSync and Deno.writeTextFileSync. A caller who can supply the code can read and write files as the server OS user, and can persist code execution by writing a .pth file into the server virtualenv's site-packages. The usual remote path is FastMCP's GenerativeUI provider, which registers the MCP tool generate_prefab_ui and forwards that tool's code argument into this sandbox. On a FastMCP server with no authentication configured, any client who can reach the server can call the tool. With authentication, any client allowed to call tools can. Generative UI is opt-in, and the host must have Deno installed. Confirmed on prefab-ui 0.20.2. Fixed in 0.21.0: after the runtime and Prefab sources load, the runner revokes Deno read, write, and network permission before it accepts code, the process is started with --no-prompt so a permission request cannot grant them again, and bootstrap network access is limited to cdn.jsdelivr.net. The fix is commit 030f074 (pull request 538). |
| A Beta9 tenant can attach a volume when deploying a function, endpoint, or task queue. The gateway builds the host path by joining the shared volumes root, the workspace name, and the volume identifier from the stub configuration. Nothing checks that the identifier names a volume that workspace owns. Parent-directory segments in the identifier collapse the joined path to the worker's filesystem root, and the worker bind-mounts that path into the sandbox read-write unless the volume was explicitly marked read-only. A tenant with a workspace token can therefore read and write the worker's files, including other tenants' volumes stored on that machine. Confirmed against gateway 0.1.769 and worker 0.1.745. Fixed in gateway 0.1.776 and worker 0.1.756: volume identifiers and mount paths are validated before a mount is built, including for stubs saved before the check existed, and the gateway rejects a volume that does not exist or is not owned by the workspace. |
| Beta9 turns each shell command in a tenant image definition into a Dockerfile RUN line and executes it with Buildah on the build worker. The worker image sets BUILDAH_ISOLATION=chroot, and no buildah invocation passes --isolation, so the step only changes its filesystem root. It keeps the worker pod's PID, network, IPC, and user namespaces, runs as uid 0, and holds CAP_DAC_OVERRIDE. PID 1 in that procfs is the worker entrypoint, so /proc/1/root is the worker container's filesystem and is writable. That filesystem contains the runc, buildah, and runsc binaries the worker uses to start tenant containers, the worker's Kubernetes service account token, and the host directories mounted into the pod. Worker pods are privileged. Clip v2 layered builds are the default, so a tenant with a workspace token reaches this by submitting one image build. The read-only proof confirmed the shared namespaces, CAP_DAC_OVERRIDE, and that the worker's runc binary is visible at /proc/1/root. Fixed in worker 0.1.756: Buildah RUN steps, including the WORKDIR mkdir, use OCI isolation with private PID, IPC, and network namespaces, and tenant ARG values are passed only into the build container. |
| When `[oauth2_client].ACCOUNT_LINKING` is set to `auto` (not the default) and registration is enabled, the link-account sign-up form looked up the existing account to link by the user name or email submitted in the form instead of the identity asserted by the OAuth2 provider. An attacker with an account on a configured OAuth2 provider could submit another user's name, link their external identity to that Gitea account and be signed in as it without its password, including as an administrator. The link persisted for later sign-ins. Automatic linking now only happens when the existing account matches the provider's email or provider-derived user name; otherwise the account password is required. |
| This issue was addressed by enforcing library validation on the affected binaries. This issue is fixed in Xcode 27. An app may be able to break out of its sandbox. |
| Uncontrolled Recursion vulnerability in Apache Thrift Java bindings.
This issue affects Apache Thrift: before 0.24.0.
Users are recommended to upgrade to version 0.24.0, which fixes the issue. |
| Highflame ramparts 0.8.1 through 0.8.10 contains a security scan bypass vulnerability that allows attackers to hide malicious payloads because collect_bundle_siblings and walk_bundle_subdir skip nested directories and .pyc files. Attackers can distribute a skill bundle with an unchecked-hash .pyc under scripts/<subdir>/__pycache__/ that ramparts silently ignores, while CPython executes it when the skill runs. |
| Nexterm through 1.2.2-BETA contains a missing authorization vulnerability that allows authenticated users to hibernate, resume, or terminate other users' remote sessions via session lifecycle routes. Attackers who obtain a session ID can call the hibernate, resume, or DELETE connection endpoints to disrupt live SSH, RDP, VNC, or SFTP sessions. |
| LearnHouse before 1.1.1 contains a missing authorization vulnerability in the assignment API that allows authenticated users to read other users' grades and submissions. Attackers can supply another user_id to the submissions grade endpoint or query submission read functions to view other students' work, grades and instructor feedback. |
| WebDyne versions from 2.014 before 3.022 for Perl issue session cookies without the Secure, HttpOnly and SameSite attributes.
WebDyne::Session sets its session cookie with only a name, value and path.
Browsers then send the session ID over plain HTTP and with cross-site requests, and expose it to page scripts. A network attacker or an injected script can steal the session, and another site can make requests that carry it. |
| A vulnerability was found in SourceCodester Online Food Ordering System 1.0. The impacted element is an unknown function of the file order.php. The manipulation of the argument price/title results in client-side enforcement of server-side security. It is possible to launch the attack remotely. The exploit has been made public and could be used. |
| A vulnerability was identified in iFlytek astron-agent up to 1.1.2. This impacts the function FileInfoV2Service.listKnowledgeByPage of the file FileInfoV2Service.java of the component Console Knowledge-List Endpoint. Such manipulation of the argument space-id leads to authorization bypass. The attack may be launched remotely. The exploit is publicly available and might be used. The vendor was contacted early about this disclosure but did not respond in any way. |
| Crass is a Ruby CSS parser that's fully compliant with the CSS Syntax Level 3 specification. Prior to 1.0.7, when the `:preserve_comments` option is not enabled, which is the default, Crass discards CSS comments by recursively consuming the next token. An attacker-provided stylesheet containing a very large number of adjacent comments can cause excessive recursion and trigger a `SystemStackError`, denying service to the parsing application. Preserving comments avoids this recursive path. This issue is fixed in version 1.0.7. |
| Forge (also called node-forge) is a native implementation of Transport Layer Security in JavaScript. Prior to 1.4.1, the affected implementation used forge.tls.createConnection(), TLS record processing, process() to process the relevant input or security decision. node-forge processed received TLS records recursively, consuming one JavaScript stack frame per record in a single input chunk. A remote peer could send thousands of small records together and exhaust the call stack, causing an uncaught exception and denial of service. This issue is fixed in version 1.4.1. |
| Forge (also called node-forge) is a native implementation of Transport Layer Security in JavaScript. Prior to 1.4.1, the affected implementation used asn1.fromDer(), BIT STRING decoding, maxDepth to process the relevant input or security decision. node-forge decoded ASN.1 stored inside BIT STRING values without consistently applying the configured maxDepth limit and recursively retained copies at each nesting level. Crafted DER containing nested BIT STRINGs could cause excessive recursion, CPU use, and quadratic memory consumption during asn1.fromDer(). This issue is fixed in version 1.4.1. |
| IXP Manager is a full-stack web application for managing Internet Exchange Points. Prior to version 7.3.0, an authenticated user can view and edit a resource belonging to another user without authorization. The API-key edit/update path in `ApiKeyController` selects records by identifier without checking that `user_id` matches the caller, allowing access to another user's API-key record. This issue is fixed in version 7.3.0. |
