Reflected cross-site scripting in e107 CMS before 2.3.13 lets unauthenticated attackers run arbitrary JavaScript in a victim's browser by sending a single crafted GET request to search.php with a manipulated author parameter. The parameter is split with parse_str() and its unescaped value is written into a rendered HTML input tag through the form handler's 'other' option key, so attackers can break out of the tag and inject an <img onerror=...> payload. Successful exploitation requires the victim to be lured into loading the attacker-supplied URL (CVSS 4.0 base 5.1, UI:A), and no public exploit code or confirmed active exploitation (CISA KEV) was identified at time of analysis.
In a2a-rs (a2a-server crate) through 0.4.3, the in-memory task store does not clamp client-supplied page-size and history-length parameters, so an authenticated low-privilege caller can request oversized result sets and pull task-history records beyond the pagination bounds the API is meant to enforce. Only deployments using InMemoryTaskStore that expose the task-list/history endpoints are affected, and the impact is limited to read-only disclosure of task metadata and message history - no write, no code execution and no denial of service. No public exploit code has been identified and the issue is not in CISA KEV, so this should be handled as a moderate-priority patch rather than an emergency (CVSS 4.0 base 5.1; assessed CVSS 3.1 AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:N/A:N).
Trigger.dev versions 4.5.0 through 4.5.x expose a delegated user-actor token through the Dashboard Agent input proxy, which persists the token in browser-readable session input records at payload.metadata.userActorToken. An attacker who already holds a chat's public session token can query the input records endpoint to recover that credential and replay it to exchange for an environment JWT, yielding low-privilege read access within the targeted environment; the CVSS 4.0 base score is 5.1 and the assessed CVSS 3.1 vector indicates network-reachable, low-privilege, user-interaction-dependent confidentiality impact only. No public exploit code or confirmed active exploitation (CISA KEV) was identified at time of analysis, and the issue is fully remediated in 4.6.0.
Stored cross-site scripting in Trudesk through 1.2.11 allows any authenticated user to upload script-bearing HTML or SVG files through the uploadlogo, uploadpagelogo, and uploadfavicon branding endpoints, which are then served unauthenticated from /assets/. Exploitation requires a valid low-privileged account (CVSS PR:L) and victim interaction when a page referencing the asset is loaded (UI:P); the impact is limited to integrity and cross-site scripting rather than remote code execution. No public exploit identified at time of analysis, and the CVSS 4.0 base score is 5.1 (medium).
Cross-site request forgery in WWBN AVideo before 24.0 allows an attacker to make a logged-in user's private playlists public, remove playlist videos, or forge video suggestions by luring the victim to an attacker-controlled page with an auto-submitting form. The flaw exists in legacy endpoints such as playlistStatus.php that are not covered by the application's .json.php-only autoCSRFGuard() token check. Exploitation requires an active authenticated AVideo session and user interaction, and no public exploit code was identified at time of analysis; impact is limited to the victim's playlist and suggestion state rather than server or account takeover.
An out-of-bounds read in device-mapper-multipath's multipathd daemon allows local users who can submit commands to the daemon's control interface to read past the end of a buffer by supplying a command whose format string ends in a trailing percent character, exposing unintended daemon memory (limited information disclosure) or crashing multipathd to cause a denial of service. Local attackers need no prior privileges on the host per the CVSS vector (PR:N), but exploitation requires a local account on a system where multipathd is running and reachable. Affected products include Red Hat Enterprise Linux 6, 7, 8, 9 and 10 plus Red Hat OpenShift Container Platform 4, which ships the component in its RHEL-based node stack. No public exploit code has been identified at time of analysis and there is no CISA KEV listing; impact is bounded to confidentiality and availability (5.1, C:L/A:L), with no code execution or integrity impact.
Incorrect authorization in hMailServer's Sender Rewriting Scheme (SRS) handling (versions 6.2.1 through 6.3.6) allows an authenticated, low-privileged party holding a valid server-signed SRS address to relay mail through the server to arbitrary external recipients without satisfying the SMTP authentication that the relay rules normally impose. The flaw only matters when SRS forwarding is explicitly enabled (it is off by default), and the signed token remains usable for roughly 21 days, so exposure is bounded to deployments that use the feature and to attackers who can obtain a token - either as the recipient of a forwarded message or as an account holder who forwards mail to an outside mailbox. Fixed in release 6.3.7; no public exploit identified at time of analysis.
Authorization flaw in Apache Pulsar's Functions Worker trigger REST endpoint allows an authenticated user who holds function-management permission on a namespace to cause the Worker to publish caller-controlled messages to a configured input topic that the user is not authorized to produce to, including topics in other tenants. Exploitation requires the Worker's own broker-client identity to be authorized to produce to the target topic and, in Pulsar 2.3.0 and later, to list that topic's subscriptions, which in practice mainly holds when the Worker identity is a superuser; no function instance needs to be running and no function code is executed. This is a moderate-severity integrity issue rather than a critical priority, and no public exploit was identified at time of analysis; vendor fixes are available in 4.0.14 and 4.2.5.
Server-side request forgery in the Metasploit Framework FTP client libraries lets a rogue or compromised FTP server redirect an operator's data connection to an arbitrary host and port by supplying a crafted 227 PASV (or 229 EPSV) response. The flaw affects Metasploit Framework builds containing the pre-fix lib/metasploit/framework/ftp/client.rb, lib/msf/core/exploit/remote/ftp.rb, and lib/rex/proto/ftp/client.rb code, and is rated CVSS 4.7 (Medium) with UI:R because the operator must actively connect to the malicious server. No public exploit identified at time of analysis; exploitation yields blind outbound TCP connectivity only, consistent with C:L/I:N/A:N.
Server-side request forgery in Metasploit Framework before 6.5.7 allows an attacker-controlled HTTP server to redirect requests made through send_request_cgi! to arbitrary hosts reachable from the operator's network, including internal HTTP services. The CVSS 3.1 vector (4.7, AV:N/AC:L/PR:N/UI:R/S:C/C:L/I:N/A:N) reflects that no privileges are required but the operator must interact with the malicious endpoint, and scope is changed because internal services can be reached. No public exploit was identified at time of analysis, and the issue is fixed in 6.5.7 via a new HttpRedirectPolicy that defaults to same_host.
hMailServer 6.0.0 through 6.3.6 writes SMTP SASL authentication exchanges to its log in cleartext for several common mechanisms, so a local user who can read the log - or anyone handed a copy of it - can recover mailbox credentials. Exposure requires SMTP logging to be enabled, which is off by default, plus local read access to the log files; on Windows before 6.3.6 every local user could read the log folder, lowering that barrier considerably. Captured material includes AUTH PLAIN passwords sent before EHLO or with non-standard spacing, XOAUTH2/OAUTHBEARER bearer tokens and SCRAM-SHA-256 client proofs (from 6.2.0, with logged salt and iteration count permitting offline guessing) and GSSAPI authenticators (from 6.3.2); CVSS 3.1 scores this 4.7 (AV:L/AC:H/PR:L/UI:N/S:U/C:H/I:N/A:N), and no public exploit was identified at time of analysis. Fix is released in v6.3.7.
hMailServer 6.2.25 through 6.3.6 exposes server-wide REST traffic counters to domain-confined authenticated users because GET /api/v1/metrics/history bypasses the authorization gate used by report routes. A domain administrator, an auditor role scoped to selected domains, or an API key limited to named domains can read 30-day aggregate traffic history for the entire server, though no message content, account data, or configuration beyond those counters is disclosed. No public exploit code has been identified at time of analysis, and the issue is assessed as a low-to-moderate priority information-disclosure flaw, not an urgent threat.
Missing authorization checks in Splunk Web's search-time sourcetype renaming feature let authenticated users who lack the admin and power roles create, move, edit, or delete their own sourcetype renaming configurations inside apps they are not permitted to access. Splunk Enterprise deployments below 10.4.4, 10.2.8, 10.0.11, and 9.4.16 are affected, but only when Splunk Web is enabled, since the missing permission check lives in that code path. Impact is limited to the integrity of renaming configurations in otherwise inaccessible apps - no confidentiality or availability effect - giving this a low CVSS of 4.3 (PR:L, I:L); no public exploit code was identified at time of analysis, so this is a genuinely low-priority issue despite affecting a widely deployed product.
Improper authorization in the Splunk Enterprise REST API allows an authenticated user without the 'admin' or 'power' roles to relocate field extractions and field transformations into apps they are not permitted to read, because the move operation never validates destination-app access. The issue affects deployments running versions below 10.4.4, 10.2.8, 10.0.11, and 9.4.16, and carries a CVSS score of 4.3 (AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:L/A:N), reflecting a limited integrity impact with no confidentiality or availability loss. No public exploit code has been identified at time of analysis and the issue is not listed in CISA KEV, so this is a low-to-moderate priority hardening item rather than an urgent exposure, though the required credential is a low-privilege one that many environments grant broadly.
Information disclosure in Splunk Enterprise (below 10.4.4, 10.2.8, 10.0.11, and 9.4.16) and Splunk Secure Gateway (below 3.10.12, 3.9.26, and 3.8.73) allows an authenticated user who does not hold the admin or power roles to view dashboard and report listings beyond the requested filter scope. The flaw is a low-severity (CVSS 4.3) NoSQL-style filter injection in Secure Gateway that is limited to objects the user can already access, so it widens listing visibility rather than granting access to forbidden objects. No public exploit code or CISA KEV entry was identified at time of analysis, and exposure requires that Secure Gateway be in use.
Missing authorization in Splunk SOAR (on-premises) below 8.7.0 allows any authenticated user who lacks the Automation Broker edit permission to retrieve the Automation Broker Docker Compose configuration file, which discloses the SOAR instance URL, the container image location, and the Broker version. The endpoint returns the file without enforcing the permission check that is supposed to gate it, so a legitimate low-privileged account can obtain deployment metadata it should not see. Impact is confined to confidentiality (CVSS 4.3, PR:L, C:L) with no secrets, credentials, or write access exposed; there is no public exploit identified at time of analysis and the issue is not in CISA KEV. Instances on 8.7.0 or later are not affected.
Remote unauthenticated denial of service in Apache bRPC through 1.17.0 allows a network attacker to abort the server process by sending a crafted HPACK header block over HTTP/2 or gRPC. The h2 parser is enabled by default on bRPC server ports and processes headers before authentication, so any exposed h2/gRPC listener is at risk. No public exploit identified at time of analysis, and the impact is limited to availability (process crash).
Paid AI features in GitLab Enterprise Edition can be accessed without a valid subscription by authenticated users on affected deployments because billing eligibility is not consistently enforced server-side. The issue affects all versions from 17.8 before 19.3.6, 19.4 before 19.4.3, and 19.5 before 19.5.1, with Community Edition and instances that do not expose the subscription-gated AI features out of scope. No public exploit was identified at time of analysis, and exploitation requires a valid GitLab EE account on an affected version.
Authorization bypass in the H5P WordPress plugin (versions through 1.17.9) lets an authenticated user manipulate a user-controlled object key to defeat the plugin's access-control checks and modify H5P content they are not authorized to change. The assessed vector (CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:L/A:N) limits impact to integrity only, with no confidentiality or availability effect, and low-privilege authentication is a hard prerequisite, so this is a moderate, not high-priority, issue. No public exploit code or CISA KEV entry is associated with this CVE in the available data (no public exploit identified at time of analysis).
DigitalME eRoom WordPress plugin versions through 1.7.1 exposes a network-reachable action without proper authorization checks, allowing an authenticated low-privilege user such as a subscriber to make limited integrity changes. The CVSS 4.3 vector requires existing credentials and confines impact to integrity only, so this is a low-to-moderate priority issue rather than a critical one; no public exploit identified at time of analysis and no active exploitation has been reported.
Cross-site request forgery bypass in Electron before 42.11.3, 43.7.0, 44.3.0, and 45.0.0-alpha.6 lets untrusted content loaded by a desktop app make credentialed requests to other sites by corrupting how the framework attributes the initiator of links opened into new windows. Because Electron re-issued such top-level navigations as application-initiated loads rather than preserving the originating page's identity, the request could be judged same-site, causing SameSite=Strict cookies to be attached and Sec-Fetch-Site to omit the true initiator. Rated CVSS 4.2 (AV:N/AC:H/PR:N/UI:R, C:L/I:L/A:N), exploitation requires an app that both loads untrusted content and allows it to open popups (popups are disabled by default), a victim who opens the crafted link into a new window, and a target site that depends on SameSite cookies rather than anti-CSRF tokens; no public exploit code identified at time of analysis and the issue is not in CISA KEV.
Keycloak can bypass mandatory OTP enrollment during login when a realm orders the 'Update User Locale' required action before 'Configure OTP,' allowing an attacker who already holds a valid password to access the account without setting up the required one-time password credential. The issue affects Red Hat build of Keycloak and Red Hat Single Sign-On 7 and carries a low CVSS score of 4.2 because exploitation requires a non-default administrative ordering and prior knowledge of valid credentials. No public exploit code has been identified at time of analysis.
A race condition in Zephyr RTOS's Infineon PSOC 6 BLESS Bluetooth HCI driver lets a connected BLE peer within radio range obtain small amounts of adjacent RAM (up to one HCI/ACL payload) or cause duplicate-dropped HCI packets, because the driver stores its outgoing TX descriptor in a file-scope static that is populated before the serialising semaphore is taken. Only builds using hci_infineon_psoc6_bless.c with CONFIG_BT_HCI_ACL_FLOW_CONTROL=y (the default on this driver's boards) and an application that returns -EINPROGRESS from its L2CAP recv callback and later calls bt_l2cap_chan_recv_complete() are realistically exposed, since that pattern moves the ACL buffer unref onto a preemptible thread that calls bt_send() directly, outside the cooperative TX work queue. Impact is bounded - the driver only reads through the descriptor, so there is no write primitive, no memory corruption, and the leaked bytes are not attacker-chosen; the assessed vector is CVSS:3.1/AV:A/AC:H/PR:N/UI:N/S:U/C:L/I:N/A:L (4.2). No public exploit identified at time of analysis; an upstream fix commit is published alongside a GitHub security advisory.
Cross-site request forgery in Splunk Enterprise allows an unauthenticated attacker to trick a user who holds the edit_federated_providers capability into disabling all configured federated search providers, temporarily preventing federated searches from retrieving remote data. The issue affects releases below 10.4.4, 10.2.8, 10.0.11, and 9.4.16, and exploitation requires the victim to be authenticated, privileged, and induced to click a crafted link; no public exploit code was identified at time of analysis.
Null-pointer dereference in Qt's VectorImage QML type lets a crafted SVG crash Qt-based applications that parse and render it, causing a local denial of service. The flaw affects Qt builds that expose the VectorImage QML type; an unauthenticated local user must deliver the SVG and a victim must actively open or load it, and impact is limited to availability with no code execution or data exposure. No public exploit identified at time of analysis, and the issue is low-to-moderate priority because exploitation requires a specifically exposed SVG rendering path and user action.
Out-of-bounds memory read in Electron's macOS offscreen rendering (OSR) path lets an attacker who already has code execution inside the GPU process feed inconsistent frame geometry - a content_rect or pixel_size not backed by the underlying shared-memory mapping or IOSurface - to the main process, which then reads past the end of the mapped buffer while building paint-event images. The result is disclosure of main-process memory contents or an application crash, and exploitation is confined to macOS applications that enable OSR; Windows, Linux, and default windowed Electron apps are unaffected. This is a sandbox/boundary-escalation step rather than a standalone entry point: the assessed vector is CVSS:3.1/AV:L/AC:H/PR:H/UI:N/S:C/C:L/I:N/A:L (3.9), and attack complexity is high with no public exploit identified at time of analysis. Fixed upstream in Electron 42.11.6, 43.7.3, 44.4.3, and 45.0.0-alpha.9.
Out-of-bounds memory read in Electron's frame-capture path (webContents.beginFrameSubscription) affects all versions prior to 42.11.6, 43.7.3, 44.4.3, and 45.0.0-alpha.9. The main process trusted frame geometry and pixel-format metadata relayed from the GPU process - content_rect, coded_size, pixel_size, and pixel_format - without validating it against the actual shared-memory or IOSurface mapping, so a malicious set of values causes the browser to read past the end of the buffer while assembling captured frames, leaking process memory or crashing the app. Exploitation requires that the target application actually use the frame-subscription API and that the attacker already hold code execution inside the GPU process through some other bug; no public exploit code and no active exploitation were identified at time of analysis, and the CVSS 3.9 score reflects that steep prerequisite rather than the severity of the memory-safety class.
Path traversal in the Undertow web server's PathResourceManager, which serves static files from the filesystem, lets remote attackers read files outside the intended static-content root when a symbolic link resolves into a sibling directory whose name begins with the same characters as an authorized safe directory (for example 'reports' versus 'reports-internal'). The root cause is a string-prefix check that never verifies a full path-segment or directory boundary, so it fails to confirm the link's canonical target actually resides inside the approved directory. Affected deployments are those bundling Undertow's static resource handling, as reflected by Red Hat's CPEs for JBoss Enterprise Application Platform 7 and 8, Red Hat Fuse 7, Red Hat Build of Apache Camel for Spring Boot 4, Red Hat Single Sign-On 7, and RHEL 8/9 packages; the assessed severity is low (CVSS 3.1 base 3.7, AV:N/AC:H/PR:N/UI:N/C:L/I:N/A:N) because exploitation demands a pre-existing symlink plus a directory-name collision, and the outcome is limited to low-confidentiality file disclosure with no write or availability impact. No public exploit code and no confirmed active exploitation were identified at time of analysis, and the signals align on low real-world risk.
Denial of TLS certificate issuance and renewal in Progressive Robot hMailServer 6.0.0 through 6.3.6 stems from the ACME HTTP-01 challenge responder handling only one connection at a time. A remote unauthenticated attacker can open idle connections during a certificate authority validation so the single worker waits up to five seconds per connection and the CA request times out, causing validation failure; sustained interference across renewals can let the certificate expire and make clients reject the server's TLS, although mail service continues on the current certificate until then. The issue is low-severity and availability-only, with no public exploit identified at time of analysis, and is fixed in 6.3.7, which serves up to 64 concurrent connections.
Unbound 1.0.1 through 1.26.1 can be induced to downgrade a configured island-of-security trust anchor to insecure when a remote unauthenticated attacker injects a forged NSEC response with the SOA bit cleared. The flaw affects only resolvers configured with both a root trust anchor and an additional trust-anchor for an island whose parent is provably insecure under the root chain, and it causes the DNSSEC validator to accept and cache the NSEC without the required RRSIG validation. Impact is integrity-only (CVSS 3.7, AC:H), and no public exploit code was identified at time of analysis.
Node.js 26.12.0 can cache invalid ZIP local headers within its zlib handling, so an application that passes attacker-influenced ZIP or zlib data through those APIs may end up with corrupted header state affecting integrity of subsequently processed data. The vendor rates this Low, scope is unchanged, and only integrity impact is assessed (C:N/I:L/A:N) with high attack complexity, meaning exploitation depends on specific application behavior rather than merely running Node.js. The fix ships in Node.js 26.12.1; no public exploit code has been confirmed at time of analysis and EPSS/KEV signals are absent from the available data.
Node.js 26.12.0 mishandles filesystem permission tree nodes for wildcard path grants when the experimental Permission Model is enabled, potentially permitting file access beyond the intended grant scope. The issue is local, requires high attack complexity, and only affects deployments that explicitly enable the Permission Model with wildcard --allow-fs-read/--allow-fs-write specifications and allow an attacker to execute code inside that restricted context. It is rated low severity (CVSS 3.6); no public exploit identified at time of analysis, and Node.js 26.12.1 contains the vendor fix.
An out-of-bounds read in the Zephyr RTOS Bluetooth Classic A2DP host stack allows a paired BR/EDR peer to retrieve roughly 245 bytes of adjacent RAM from the victim device over the air. The bug is reachable only in firmware built with CONFIG_BT_CLASSIC and CONFIG_BT_A2DP that registers at least one non-SBC codec endpoint (MPEG1, MPEG2, ATRAC or vendor), because the SBC path pins the codec IE length to 4 bytes; the attacker must also have completed pairing since the AVDTP L2CAP channel enforces BT_SECURITY_L2 (CVSS:3.1/AV:A/AC:L/PR:L/UI:N/S:U/C:L/I:N/A:N). Impact is limited to read-only disclosure of a bounded region of RAM with no write primitive or code execution, and no public exploit code was identified at time of analysis.
Acronis Cyber Infrastructure before builds 6.0.1-112, 6.1.1-64, and 6.2.1-71 contains an improper authorization check (CWE-863) that allows a high-privileged authenticated user to read and modify sensitive data beyond the intended scope. The issued CVSS 3.3 vector (AV:N/AC:H/PR:H/UI:N/C:L/I:L/A:N) reflects network reachability but requires prior high-privilege access and difficult-to-reproduce conditions, with only partial confidentiality and integrity impact and no availability impact. No public exploit was identified at time of analysis and the issue is not listed in CISA KEV, so exploitation appears limited to specific authorized-but-overprivileged scenarios.
A file descriptor leak in the containers/storage Go library's archive.ApplyLayer and ApplyUncompressedLayer functions causes descriptors to be left open whenever corrupt or malicious image layers or tar streams are unpacked, and repeated processing of such inputs can exhaust the hosting process's file descriptor limit. The weakness is unauthenticated in CVSS terms (PR:N) but local in scope (AV:L) and requires the victim to actually process attacker-influenced layers (UI:R), so an attacker must first get malformed image content in front of a build, pull, or sync loop. Impact is availability-only and confined to the affected process (C:N/I:N/A:L, S:U): no code execution and no data exposure; a single bad input does not cause denial of service, only sustained repetition does. No public exploit code has been identified at time of analysis and the issue is not in CISA KEV, and EPSS scoring was not supplied with the source data.
Electron's chrome.tabs.query() implementation returned the url and title of every tab in a session to extensions that held neither the 'tabs' permission nor a matching host permission, letting a loaded extension enumerate the URLs and titles of all webContents in that session. Only applications that explicitly pull in extensions they do not fully trust via session.extensions.loadExtension are exposed, and only on builds prior to 42.11.1, 43.5.1, 44.1.1, and 45.0.0-alpha.4; unauthenticated remote attackers cannot reach the flaw without such an extension already running inside the app. The disclosed data is limited to tab URLs and titles (no page content, DOM, cookies, or code execution), no public exploit code was identified at time of analysis, and the issue does not appear on CISA KEV, consistent with its low CVSS 3.1 score.
Permanent deletion of existing vulnerability records in GitLab EE is possible for authenticated users holding only read-only permissions when a specific error-handling path is triggered, because cleanup logic fails to distinguish pre-existing records from newly created ones. The issue affects GitLab EE from 17.8 before 19.3.6, 19.4 before 19.4.3, and 19.5 before 19.5.1, and is rated CVSS 3.1 (AV:N/AC:H/PR:L/UI:N/S:U/C:N/I:L/A:N) with high attack complexity and only low integrity impact. There is no public exploit identified at time of analysis, and the independent assessment classifies this as a low-severity, low-priority issue.
Authorization bypass in Apache Pulsar (2.8.0 through 4.0.13 and 4.1.0 through 4.2.4) allows an authenticated user holding only namespace- or topic-level consume permission to clear backlogs, expire messages for all subscriptions, or delete subscriptions that the namespace's Prefix subscription authentication mode or per-subscription role grants should deny them. The flaw only affects brokers with authentication/authorization enabled, the default authorization provider in use, and namespaces that actually configure Prefix subscription auth mode or explicit subscription permissions; default deployments are unaffected because consume permission already authorizes these operations there. Impact is limited to subscription-state integrity - cursor advancement, subscription removal, and skipping of not-yet-replicated geo-replication messages - with no message disclosure, no policy modification, and no broker availability impact; no public exploit identified at time of analysis and CVSS scores remain low (2.3 per the vendor's CVSS 4.0 vector).
Missing authorization checks in the Apache Pulsar broker admin REST endpoints for non-persistent topics allow an authenticated user holding only produce or consume permission on a topic or its namespace to read, replace, or delete that topic's entry filter policy. Only non-persistent topics on Pulsar 3.3.0 through 4.0.13 and 4.1.0 through 4.2.4 are affected; persistent topics, namespace-level policies, and versions before 3.3.0 are not. Real impact is tightly gated: the operator must have deployed entry filter implementations (the filters directory is empty by default) and must have set allowOverrideEntryFilters=true (default false), and even then only subscription dispatchers created after the change are affected, so conversation of the vulnerable policy only reshapes which already-installed filters withhold messages. No public exploit code was identified at time of analysis, and despite the accompanying 'RCE' tag the description confirms the flaw does not permit loading filter code or running arbitrary code in the broker.
Cross-tenant information disclosure in Apache Pulsar's transaction administration REST API allows any authenticated user - including an account granted no permissions - to read transaction coordinator and transaction metadata belonging to other tenants. The flaw is reachable only when broker authentication and authorization are enabled and transaction support is turned on via transactionCoordinatorEnabled, which is disabled by default; affected read endpoints under /admin/v3/transactions validate authentication and broker/coordinator locality but skip tenant, namespace, and topic authorization checks. Exposure is limited to metadata and statistics such as coordinator service URLs, transaction topic and subscription names, the opening principal, and pending-acknowledgement state; message contents cannot be read and transactions cannot be aborted, scaled, or otherwise modified. No public exploit code or confirmed active exploitation was identified at time of analysis, and the vendor-assessed severity is low (CVSS 4.0 base 2.3; independently assessed CVSS 3.1 vector AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:N/A:N).
Apache Pulsar's WebSocket service performs an incomplete authorization check when a WebSocket consumer supplies a dead letter policy via consumer endpoint query parameters, letting an authenticated user who holds only CONSUME permission on a subscribed topic cause messages to be written into a topic of their choosing - potentially in another tenant's namespace - because the deadLetterTopic's PRODUCE permission is never verified and the write is performed with the WebSocket service's own broker client identity. Affected releases are 2.3.0 through 4.0.13 and 4.1.0 through 4.2.4; the practical exposure is limited because webSocketServiceEnabled is off by default, the attacker must be an authenticated user, the WebSocket service's broker role must already hold PRODUCE on the target, and the target must accept a schema-less producer (schema validation is not enforced by default). Impact is limited to unauthorised message injection with low integrity and availability effect (dead-letter messages are distinguishable by their properties, but those properties are unauthenticated and cannot be trusted to establish origin); no public exploit code or CISA KEV-listed active exploitation was identified at time of analysis.
pypdf versions before 6.21.1 can be induced to hand /DCTDecode-labeled image streams to the wrong Pillow image parser when the optional Pillow dependency is installed, allowing a crafted PDF to trigger unintended decoding of non-JPEG data. The attack requires the victim to load or extract the affected image with pypdf, and the assessed impact is limited to low availability effects with no confidentiality or integrity loss. No public exploit was identified at time of analysis, and the issue is fixed in 6.21.1 by restricting Pillow to the JPEG format.
Client-side script execution can be triggered in Nuxt applications running with pages disabled (all versions prior to 4.6.0) when attacker-controlled input bound to NuxtLink's `to` prop carries a leading ASCII C0 control character ahead of a `javascript:` URL. The no-pages router classifies the value as an internal route before the script-protocol check runs, so the unsafe href is emitted into server-rendered HTML; the victim must click the link for script to run in the application's origin. The issue is an incomplete fix for CVE-2026-53722, is unauthenticated (PR:N) but gated by significant preconditions - pages-disabled mode, reflected untrusted input, and a user click - and there is no public exploit identified at time of analysis; Nuxt 4.6.0 resolves it.
CSS injection in Astro's server-side View Transition generator (versions 7.1.0 through 7.3.2) lets attacker-controlled strings placed into animation properties - duration, delay, easing, direction, fillMode, or animation name - break out of their generated declaration because the ';', '{' and '}' delimiters are serialized unescaped. An application is only affected if it feeds untrusted input into those transition properties; stock Astro sites using hardcoded values are not exploitable. The framework still escapes literal less-than characters, so the injected content cannot close the style element, inject HTML, or execute JavaScript - impact is confined to CSS-level UI manipulation (hiding, repositioning or spoofing content, adding generated text, and conditional resource loads that may leak data already present in selectable DOM attributes). Severity is low (CVSS 4.0 base 2.3; independently assessed as CVSS 3.1 AV:N/AC:H/PR:N/UI:R/S:U/C:L/I:L/A:N), with no public exploit identified at time of analysis and a vendor fix released in Astro 7.3.3.
Fail-open topic initialization in Apache Pulsar (1.22.0-incubating through 4.0.13, and 4.1.0 through 4.2.4) causes the broker to silently disable a namespace's encryption_required policy for a topic whenever any step of persistent-topic loading fails, such as a metadata-store read error or a topic-level policy load failure. An authenticated client that already holds produce permission on the affected topic (CVSS PR:L) can then publish unencrypted messages that the broker accepts and stores, so the namespace's end-to-end encryption guarantee is lost for that topic; the impact is integrity-only, since already-encrypted messages and the encryption keys are never exposed. Exploitation depends on an operational initialization failure the attacker neither triggers nor controls, and the condition recurs on every topic reload (unload, bundle reassignment, broker restart) while the root cause persists - no public exploit code identified at time of analysis, and CISA KEV does not list this issue.
WarehousePG 6.27.1 through 6.27.6 and 7.2.1 through 7.6.0 allow an authenticated low-privileged user who can open a utility-mode connection to invoke gp_add_coordinator_standby(), named gp_add_master_standby() in WHPG 6 and kept as an alias in WHPG 7, without the superuser check enforced by sibling segment add/remove functions. On a cluster with no standby coordinator configured, this creates a fake gp_segment_configuration entry that blocks administrators from adding a real standby until a superuser removes it. The issue is low severity and limited to integrity and availability impact with no user-data exposure; no public exploit was identified at time of analysis and it is not confirmed actively exploited (CISA KEV).
Trigger.dev before 4.5.12 permits an authenticated project member to prematurely complete another member's DEVELOPMENT-environment waitpoint by knowing the victim's friendly waitpoint ID and submitting a crafted completion request, bypassing the per-user ownership check. The impact is low and limited to integrity of a pending development waitpoint; production environments are not affected. No public exploit was identified at the time of analysis, and a vendor patch is available.
SvelteKit (@sveltejs/kit) versions before 3.0.0 can emit attacker-influenceable HTML or JavaScript into prerendered pages when an application passes untrusted data into a cache-control header via setHeaders; the value is inserted unescaped into a <meta http-equiv="cache-control"> tag. Exploitation requires a prerendered page plus an untrusted header value at prerender/build time, and the injected payload executes only when a user later visits that page (UI:P/UI:R), so impact is confined to the victim's browser context. This is a low-severity XSS issue (CVSS 2.3; assessed CVSS:3.1 vector with C:L/I:L), with no public exploit identified at time of analysis and no CISA KEV confirmation.
Reflected cross-site scripting in itsourcecode Online Cake Ordering System 1.0 allows unauthenticated remote attackers to execute script in a victim's browser via the useruid parameter on /404.php, provided the victim opens a crafted link. Publicly available exploit code exists, and the CVSS 4.0 score of 2.1 reflects the requirement for user interaction and limited impact; no CISA KEV active exploitation is reported.
SQL injection in the 'del' parameter of /transactionsave.php in itsourcecode Online Cake Ordering System 1.0 lets an authenticated low-privilege remote attacker manipulate backend SQL queries, and publicly available exploit code exists. Impact is limited to low confidentiality, integrity, and availability effects per the assessed vector CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:L, and the niche deployment footprint keeps overall risk low.