Tandoor Recipes instances running versions prior to 2.6.10 let any authenticated user infer sensitive database values - including password hashes, email addresses, and administrator status - by supplying crafted sort_order or order_field values that Django's ORM resolves as double-underscore foreign-key traversal paths during recipe search and RecipeBook ordering. The exposure is scored 6.5 (CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:N/A:N, CWE-943), with confidentiality impact rated High but reach constrained by the need for valid credentials and the indirect, oracle-style nature of the leak. No public exploit code has been identified at time of analysis, and there is no confirmed active exploitation.
SQL injection in the 'Salary Payments via ECS' report of Frappe HR (HRMS) lets an authenticated HR/payroll user read sensitive database records by supplying crafted report filter values. The vulnerability affects Frappe HR releases before 15.60.3 on the 15.x line and before 16.7.1 on the 16.x line, and is rated CVSS 3.1 6.5 (AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:N/A:N); because the CVSS vector specifies PR:L, exploitation requires a valid low-privileged account and is therefore an insider or post-compromise vector rather than an anonymous remote attack. No public exploit code and no confirmed active exploitation (CISA KEV) were identified at time of analysis, and the issue is resolved by vendor patches 15.60.3 and 16.7.1.
IBM i releases 7.3, 7.4, 7.5, and 7.6 contain an improper authentication weakness that lets a remote attacker who already holds valid IBM i credentials bypass intended security restrictions and read limited information they should not be able to access. The flaw is network-reachable with low attack complexity and no user interaction (CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:N/A:N, score 4.3), but it cannot be triggered by an unauthenticated outsider, so realistic risk is confined to malicious insiders or attackers who have already obtained a low-privileged account. No public exploit code and no confirmed active exploitation have been identified at time of analysis, and IBM has published vendor-supplied fixes.
Information disclosure in Mattermost (versions 11.7.x through 11.7.9, 11.8.x through 11.8.5, 11.9.x through 11.9.1, and 11.10.0) allows an authenticated member of a channel to recover the server-side webhook URLs and authentication secrets embedded in interactive-message posts. The server fails to strip action integration secrets before broadcasting the post_deleted WebSocket event and before returning thread API responses, so a channel member who deletes or observes deletion of an interactive-message post - or simply reads the threads REST API - can harvest credentials intended to stay server-side. Exploitation requires a valid authenticated account with membership in a channel that actually uses interactive message/action integrations, and there is no public exploit identified at time of analysis; the issue is a moderate-severity confidentiality exposure rather than an urgent emergency, with an assessed CVSS of 6.5 (AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:N/A:N).
Authenticated users of Frappe HRMS (Frappe HR) can read sensitive HR records belonging to other employees because the product's PWA employee-data API endpoints fail to enforce object-level authorization prior to versions 15.58.3 (15.x branch) and 16.4.3 (16.x branch). Rated CVSS 6.5 (medium) with vector AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:N/A:N, the flaw is a read-only confidentiality breach: no integrity or availability impact, and it requires a valid employee-level session, so anonymous external attackers cannot exploit it. Exploitation only works where the deployment exposes the affected PWA endpoints; no public exploit code has been identified at time of analysis, and there is no CISA KEV confirmation of active exploitation. Vendor-released patch: 15.58.3 and 16.4.3.
ERPNext instances prior to 15.118.0 (v15 branch) and 16.27.0 (v16 branch) expose financial data through the Sales Person-wise Transaction Summary report because the report's server-side query omits the requesting user's document-permission match conditions. An authenticated attacker holding only a limited role that can run this report can read Sales Order/Sales Invoice, Sales Team, and item-level financial rows for records the role should never see, resulting in a high-confidentiality, read-only disclosure (CVSS 6.5, AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:N/A:N; CWE-862 Missing Authorization). No public exploit code has been identified at time of analysis, and the flaw is not listed in CISA KEV, so risk is driven by the low exploitation complexity for any logged-in low-privilege user rather than by observed attacks.
Authenticated guest accounts on Mattermost 11.7.x through 11.7.9, 11.8.x through 11.8.5, 11.9.x through 11.9.1, and 11.10.0 can be escalated to team or channel administrator by abusing the member roles API endpoint, which fails to validate a crafted combination of guest and admin scheme roles. The flaw requires only a valid guest login and network reachability to the API, so any server that provisions guest accounts is exposed, though no public exploit code or confirmed active exploitation has been identified at time of analysis. Vendor advisory MMSA-2026-00752 tracks the issue, and Mattermost has published it on its security updates page.
Mattermost Server versions 11.7.x through 11.10.x can disclose the stored client secret of an outgoing OAuth connection to an attacker who holds the administrative permission to manage those connections. The platform fails to require re-entry of the stored secret during connection validation, so an authenticated high-privileged user can repoint the connection's token URL at an attacker-controlled endpoint and trigger validation, causing the secret to be sent to that endpoint (CWE-201, information disclosure). Our assessment rates this CVSS 6.5 (CVSS:3.1/AV:N/AC:L/PR:H/UI:N/S:U/C:H/I:N/A:N), with the practical risk tempered by the privilege prerequisite - the feature must be enabled and at least one outgoing OAuth connection with a stored secret must already exist; no public exploit was identified at time of analysis and there is no KEV listing. Exploitation is remote, requires no victim interaction, and yields confidentiality impact only (no integrity or availability impact).
Incorrect authorization in Devolutions Server 2026.3.8.0 and earlier lets an authenticated user who holds only View permission on a connection retrieve stored add-on passwords, because the connection data returned to the client includes add-on properties unredacted rather than masked. Exploitation requires a pre-existing low-privilege account and only exposes secrets for connections that account can already view, so the impact is high-confidentiality but scoped to credentials stored in add-on properties; no public exploit code has been identified at time of analysis and there is no confirmed active exploitation (CISA KEV). The vendor has released a fix, making this a genuine but moderate-priority information-disclosure issue rather than a critical emergency.
Authorization bypass in SFM Software Technologies Inc.'s Costifier allows any authenticated user to reach or alter resources belonging to other users by manipulating a trusted, user-controlled object identifier. The flaw is rated CVSS 6.5 (CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:H/A:N), a moderate-severity CWE-639 (Authorization Bypass Through User-Controlled Key) issue that requires only low-privilege credentials - remote unauthenticated exploitation is not possible, and no user interaction is needed. All Costifier versions through the 2026-10-16 build are affected; the vendor was contacted but did not respond, so no patched version is confirmed, and no public exploit identified at time of analysis. Because integrity impact is rated High while confidentiality and availability are None, the primary risk is unauthorized modification of other users' records rather than bulk data exposure.
Memory exhaustion in Text::CSV_XS before 1.65 allows remote attackers to degrade or crash long-running Perl services by repeatedly submitting malformed CSV records. Each parse failure in getline() - and the trailing row in getline_all() - abandons the already-parsed field array without freeing it, so the process's resident memory grows with every bad record, with wide records amplifying the per-record leak; authentication requirements are not a barrier per the assessed vector (PR:N, UI:N). No public exploit identified at time of analysis, and EPSS is low at 0.19% (8th percentile), but any internet-facing daemon or worker parsing untrusted CSV is exposed.
Unauthenticated request-context confusion in Wheels 4.0.6 through 4.1.2 lets any remote client push a normal request into a separate test application context by supplying an X-Wheels-Test-Context header, a WHEELS_TEST_CONTEXT cookie with any value, or a test-runner path anywhere in the URL - including the query string. Because the alternate context reuses the same configuration, environment, datasource and debug settings, there is no confidentiality or privilege gain; the realistic impact is limited to bypassing maintenance mode when it was enabled only at runtime, obtaining a second independent set of in-memory rate-limit counters, and triggering extra application-startup side effects and memory consumption under repeated requests. No public exploit code has been identified at time of analysis, and the vulnerability is fixed in Wheels 4.2.0.
Out-of-bounds reads and writes in the X.Org X client libraries (libX11, libXi, libXpm, libXrender and libXtst) can be triggered when a connecting X11 client parses crafted replies or events returned by a malicious, compromised, or man-in-the-middled X server, causing client crashes and, in a subset of cases, limited memory disclosure. The lead issue in the advisory is an out-of-bounds write in libX11's XKB action parser when a XkbGetMap reply contains keycode values above max_keycode + 1; the family also includes multiple out-of-bounds reads in libXi's XI2 class parsing and XQueryDeviceState/XListInputDevices replies, all of which are fixed in libX11 1.8.14, libXi 1.8.4, libXpm 3.5.20, libXrender 0.9.13 and libXtst 1.2.6. The assessed vector (CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:N/I:N/A:H, 6.5) indicates exploitation is unauthenticated but requires user interaction, with impact limited to denial of service rather than confirmed code execution; no public exploit identified at time of analysis, and no KEV entry or EPSS score was supplied.
Denial of service in libtirpc allows an authenticated remote RPC client to exhaust server memory by sending oversized wrapped-data requests to an RPC service that is configured with RPCSEC_GSS integrity (rpc_gss_svc_integrity) or privacy (rpc_gss_svc_privacy) protection. Only services explicitly running in that non-default, Kerberos-backed mode are exposed - most commonly NFS deployments on Red Hat Enterprise Linux 6 through 10, Red Hat Hardened Images, and Red Hat OpenShift Container Platform 4 - and the impact is availability-only (CVSS 6.5, A:H) with no confidentiality or integrity effect. No public exploit code was identified at time of analysis and the issue is not confirmed actively exploited (no CISA KEV listing), so it warrants medium-priority patch tracking rather than emergency response, although affected RPC services can become unstable or terminate under the resource exhaustion.
Missing authorization filtering in Open Zaak's POST search endpoints (/zaken/api/v1/zaken/_zoek and /documenten/api/v1/enkelvoudiginformatieobjecten/_zoek) allows an authenticated API client holding a valid OAuth/JWT token with the zaken.lezen or documenten.lezen scope to retrieve case or document metadata outside the case types and confidentiality levels that token was authorized for. Exploitation requires a legitimate token (authenticated, PR:L) and does not require user interaction; the exposure is limited to metadata, with document searches additionally requiring the attacker to already know the target UUIDs, and document content remaining protected by a separate download authorization. No public exploit code has been identified at time of analysis, and the issue is fixed in Open Zaak 1.27.2, 1.28.2, and 1.29.0, so all earlier releases in the affected branches should be considered vulnerable.
Stored cross-site scripting in the WordPress plugin PowerPack Addons for Elementor (slug powerpack-lite-for-elementor) through version 3.1.0 lets an authenticated attacker with contributor- or editor-level access inject malicious script through the plugin's Elementor widgets, which then executes in the browser of any user who later views the affected page. CVSS 3.1 scores this 6.5 (AV:N/AC:L/PR:L/UI:R/S:C/C:L/I:L/A:L), reflecting a low-privileged but authenticated requirement plus user interaction from the victim viewer, so exploitation is not fully automatic. No public exploit code was identified at time of analysis and the issue is not in CISA KEV; the practical risk is a moderate, deliverable-but-conditional stored XSS rather than an emergency.
Stored cross-site scripting in the WordPress plugin The Events Calendar (versions through 6.18.1) lets an authenticated attacker with event-authoring privileges inject persistent JavaScript that executes when another user, such as an administrator, views the affected event page. Exploitation requires a low-privilege WordPress account (PR:L) and victim interaction (UI:R), and the flaw carries a CVSS 3.1 score of 6.5 (AV:N/AC:L/PR:L/UI:R/S:C/C:L/I:L/A:L). No public exploit has been identified at time of analysis, and the issue was reported by Patchstack.
Stored cross-site scripting in the MapSVG WordPress plugin (mapsvg-lite-interactive-vector-maps) through version 8.16.1 allows an authenticated low-privilege user to persist JavaScript that later executes in the browser of any higher-privileged user who views the affected map content. The issue is scored CVSS 6.5 (CVSS:3.1/AV:N/AC:L/PR:L/UI:R/S:C/C:L/I:L/A:L), reflecting that exploitation requires a pre-existing authenticated account with enough rights to submit or edit MapSVG input, and that the payload only fires when a victim - typically an administrator or editor - renders the page; no public exploit identified at time of analysis and there is no CISA KEV listing. Deployment risk is concentrated on sites that grant content-entry or map-editing roles to untrusted or semi-trusted users, since anonymous visitors cannot trigger the stored payload.
Stored cross-site scripting in the GeoDirectory - WP Business Directory Plugin and Classified Listings Directory for WordPress (all versions up to and including 2.8.191) allows authenticated attackers with subscriber-level access or above to inject arbitrary JavaScript into a listing's post_content that then executes in the browser of any visitor viewing that listing page. The attacker must own an already-published listing, because the GeoDir_Post_Data::can_edit() ownership check limits who can store the payload and ajax_save_post() re-applies the existing post_status on edit, so changes to a published listing bypass re-moderation and go live immediately; sites that disable open registration, disallow front-end submissions, or hold listings in moderation are substantially less exposed. This moderate-severity flaw (CVSS 6.4, CWE-79) was reported by Wordfence, and no public exploit identified at time of analysis (it is not listed in CISA KEV).
Stored cross-site scripting in the HivePress - Business Directory, Listings & Classified Ads plugin for WordPress (all versions up to and including 1.7.32) allows an authenticated attacker with subscriber-level access or above to inject arbitrary web scripts through the 'message' parameter, which are then rendered to other users. Exploitation is conditional: the site administrator must have enabled user listing submission (a non-default setting), and a victim must subsequently log in through the login form embedded on the attacker's listing page for the payload to execute in their browser. This is a medium-severity issue (CVSS 6.4, CWE-79); no public exploit code was identified at time of analysis and there is no CISA KEV confirmation of active exploitation, so it should be treated as a moderate, conditional priority rather than an emergency.
Stored cross-site scripting in the Cooked - Recipe Management plugin for WordPress (all versions up to and including 1.16.1) lets an authenticated attacker with contributor-level access or above embed arbitrary JavaScript through the [cooked-timer] shortcode, which is stored and then executed in the browser of any user - including an administrator - who later views the injected post, page, or recipe. The root cause is insufficient input sanitization and output escaping in the plugin's shortcode handling (CWE-79), and the CVSS 3.1 vector AV:N/AC:L/PR:L/UI:N/S:C/C:L/I:L/A:N places it at 6.4, a moderate-severity issue contingent on the attacker already holding a contributor+ account on a site running the vulnerable plugin. There is confirmed actively exploited status? No - no public exploit identified at time of analysis, and no CISA KEV listing; treat this as a routine patch-management priority rather than an emergency.
Stored cross-site scripting in the FiboSearch - Ajax Search for WooCommerce WordPress plugin, in all versions up to and including 1.34.2, lets an authenticated attacker with Subscriber-level access or above inject arbitrary JavaScript through the 'purchased_plan' parameter that later executes in the browser of any user viewing the affected admin page. Exploitation is gated by a non-default precondition: the site owner must have already completed the standard Freemius opt-in flow, after which a Subscriber can trigger the vulnerable handler simply by loading /wp-admin/profile.php with crafted query parameters. Scored CVSS 3.1 6.4 (AV:N/AC:L/PR:L/UI:N/S:C/C:L/I:L/A:N), the issue is confirmed fixed in 1.34.3; no public exploit code has been identified at time of analysis, and it is not listed as actively exploited.
Stored cross-site scripting in the GeoDirectory WordPress plugin (all versions up to and including 2.8.192) allows an authenticated user with Subscriber-level access to inject arbitrary JavaScript into the 'twitterusername' custom field, which is then stored and executed whenever another user - ideally a higher-privileged admin or moderator - views the affected listing page. The only input restriction on that field is a client-side HTML 'pattern' attribute, which is trivially bypassed, while server-side sanitization and output escaping are insufficient. Exploitation is confirmed to require an account with listing-creation privileges and a victim who views the injected listing; no public exploit code has been identified at time of analysis, and the overall risk is assessed as moderate priority.
Stored cross-site scripting in the Premium Packages - Sell Digital Products Securely WordPress plugin, affecting all versions up to and including 7.2.8, allows an authenticated attacker with subscriber-level access or above to inject arbitrary script into an order 'note' that executes when a privileged user opens the affected order-notes page in wp-admin. The payload defeats WordPress' sanitize_textarea_field() filtering because it is saved as a hex-escaped sequence such as \x3cimg onerror=...\x3e that is only reconstituted into a live HTML tag by stripcslashes() during admin-side rendering, after sanitization has already run. No public exploit code has been identified at time of analysis, and exploitation is conditional on the victim (typically an administrator) viewing the poisoned notes view, so the practical risk is moderate rather than critical; the vendor has committed an upstream code fix in the plugin's order-notes view.
Stored cross-site scripting in the Mang Board WordPress plugin (versions up to and including 2.4.4) allows authenticated attackers with subscriber-level access or higher to inject arbitrary scripts into pages via the 'content' parameter by omitting the data_type parameter in a modify request, which bypasses the HTMLPurifier sanitization branch and leaves only reversible mbw_htmlspecialchars2 encoding as a barrier. Exploitation requires only that a victim view the injected Mang Board page, resulting in in-browser script execution with limited confidentiality and integrity impact (CVSS 6.4); this is a moderate-severity issue representing a real but not emergency-level priority, and no public exploit identified at time of analysis. Sites that disable open registration meaningfully limit the attacker pool, and exploitation is specific to versions up to and including 2.4.4.
Stored cross-site scripting in the SureDash - Community, Courses & Member Dashboard WordPress plugin affects all versions up to and including 1.12.0 and lets an authenticated attacker with subscriber-level access or above inject arbitrary JavaScript through the 'display_name' parameter, which is stored and later rendered as live HTML. The payload fires when any user hovers over a comment's like counter on a community page, because the plugin's tooltip handler assigns the attacker-controlled data-tooltip-description value directly to innerHTML and the browser re-parses the decoded entity payload as markup. This is a moderate-risk authenticated stored XSS (CVSS 6.4, CWE-79) rather than an urgent emergency; exploitation requires a registered account, an active SureDash community/comment interface, and a victim performing the hover interaction, and no public exploit code or confirmed active exploitation (CISA KEV) was identified at time of analysis. An EPSS score was not included in the available intelligence, and the vendor has released version 1.12.1 which removes the condition.
Stored cross-site scripting (XSS) in the WebberZone Top 10 - Popular Posts WordPress plugin, a moderate-severity issue (CVSS 6.4), affects all versions up to and including 4.5.2 and allows an authenticated attacker with contributor-level privileges or higher to inject arbitrary web scripts via crafted block JSON. The attacker sets an undeclared is_widget flag to '1' and supplies a malicious instance_id value that is concatenated into an unescaped HTML class attribute, so the payload executes when a victim-potentially an administrator-views the injected page. No public exploit code has been identified at the time of analysis, and exploitation requires a WordPress site that permits contributor-level or higher accounts and a non-default block configuration.
Stored cross-site scripting in the Tutor LMS - eLearning and Online Course Solution WordPress plugin (all versions up to and including 4.1.0) allows an authenticated attacker with subscriber-level access or above to inject arbitrary JavaScript via the display_name parameter, which is insufficiently sanitized on input and not escaped on output. The payload is only stored and rendered when a Legal Consent record exists for the attacker's account, causing a 'View Logs' button to appear in the Consent Logs column of wp-admin/users.php; an administrator or other privileged user must then open that page for the script to execute, so exploitation is gated on consent logging being enabled and on a privileged user visiting the affected admin screen. Rated CVSS 6.4 (CWE-79), this is a moderate-severity issue with no confirmed active exploitation (not in CISA KEV) and no public exploit identified at time of analysis, though the Subscriber-level prerequisite is trivially met on sites with open user registration.
Stored cross-site scripting in the MaxFoundry Social Share Buttons plugin for WordPress, affecting all versions up to and including 1.30, allows an authenticated attacker with subscriber-level access or higher to inject arbitrary JavaScript through the 'form' parameter containing 'share_url', which then executes in the browser of any user who views the injected page. Exploitation is gated by the plugin's review-nag logic: the nonce required to reach the vulnerable save-network handler is only rendered in the DOM after a 7-day timer has elapsed since the attacker's first wp-admin visit, so a low-privilege account and a multi-day wait are both required, which is why the assessed vector raises attack complexity to High and this is treated as a moderate-priority issue rather than an emergency. No public exploit code was identified at time of analysis, and there is no confirmed active exploitation (CISA KEV).
Stored cross-site scripting in the WordPress Profile Builder plugin (CozmoLabs) affects all versions up to and including 4.0.3 and allows an authenticated subscriber-level user to inject persistent JavaScript that executes in the browsers of anyone viewing an affected page. The payload is written through the 'display_name' parameter and, critically, is stored in wp_comments.comment_author, so it survives deleting or renaming the attacker's account and continues to fire as long as the poisoned comment is rendered. Exploitation requires the target site to allow self-service registration or the attacker to already hold a low-privilege account; no public exploit identified at time of analysis, and the plugin's own vendor-reported severity is moderate (CVSS 6.4, PR:L, scope-changed).
Stored cross-site scripting in the Quick PayPal Payments: Payment Button & Form WordPress plugin (all versions up to and including 6.0.4) allows authenticated attackers holding Contributor-level access or higher to inject arbitrary JavaScript through the plugin's 'form' parameter, which then executes for every viewer of the affected post. The sanitization bypass is notable because the payload survives wp_kses_post by using a quote-breaking, angle-bracket-free construction, so injected script fires even for Editors and Administrators who routinely preview a Contributor's pending content. No public exploit code has been identified at time of analysis, and this is a moderate-severity (CVSS 6.4) issue whose practical exposure depends on whether a site permits untrusted users to register or be assigned the Contributor role.
Stored cross-site scripting in the SQL Chart Builder plugin for WordPress (elvinhaci), affecting all versions from 0 through 3.0.6, allows an authenticated subscriber-level user to persist arbitrary JavaScript that executes for anyone viewing a rendered chart. The attacker seeds a payload into a user-controlled database column - for example display_name set via /wp-admin/profile.php - and the plugin later prints that value unescaped as a chart X-axis label, so execution depends on an administrator having published a chart whose SQL query projects such a column and on a victim subsequently loading that page. No public exploit code has been identified at time of analysis and the CVE is not listed in CISA KEV, so this is a moderate-risk, not a top-priority, issue whose real-world reach is limited by that required admin-created chart configuration.
Stored cross-site scripting in the WordPress Name Directory plugin (all versions up to and including 1.34.2) allows authenticated attackers holding contributor-level privileges or higher to inject arbitrary JavaScript via the 'dir' attribute of the plugin's shortcode, which is then persisted and executed in the browser of any user who later views the affected page. The issue is a plugin-specific input sanitization and output escaping failure (CWE-79) rather than a WordPress core privilege escalation, and it is remediated in version 1.34.3. No public exploit identified at time of analysis and no CISA KEV listing; the CVSS 3.1 base score is 6.4 (AV:N/AC:L/PR:L/UI:N/S:C/C:L/I:L/A:N), reflecting moderate severity with low confidentiality and integrity impact and scope change from the plugin context into the victim's browsing session.
Stored cross-site scripting in the Tourfic - AI Powered Travel Booking, Hotel Booking & Car Rental WordPress plugin affects all versions up to and including 3.0.1, allowing an authenticated attacker with contributor-level access or above to inject arbitrary web scripts via the 'type' parameter handled by the Wishlist shortcode. The injected payload is crafted without angle-bracket characters so it survives the wp_kses_post filter applied on save, and it then executes as a live event handler in the unauthenticated (guest) rendering branch whenever any visitor loads the affected page. Rated medium severity (CVSS 6.4, AV:N/AC:L/PR:L/UI:N/S:C/C:L/I:L/A:N) and reported by Wordfence; no public exploit code or CISA KEV entry was identified at time of analysis, and the practical risk is bounded by the requirement that a site grant the attacker contributor rights (for example through open registration or lax role assignment) and publish the injected content to guest-visible pages.
Stored cross-site scripting in the WPPool 'Dark Mode - Improve Accessibility with AI Powered Dark Theme' WordPress plugin (all versions up to and including 5.3.17) allows attackers holding a Contributor-level account or higher to inject arbitrary JavaScript via the 'icon_light' shortcode attribute, which then executes in the browser of any visitor who later loads the injected page. Authentication is required and the privilege gate is Contributor-and-above (CVSS vector PR:L, scope changed), so sites that tightly control or disable self-registration and editorial roles are largely insulated. No public exploit code and no confirmed active exploitation have been identified at time of analysis; EPSS and KEV data were not provided.
Authenticated users on Mattermost Server 11.7.x through 11.10.0 can quarantine posts inside direct and group messages they are not participants of by issuing a direct API call to the post-quarantine endpoint, because the server performs no DM/GM membership or authorization check before processing the request (CWE-862, Missing Authorization). The flaw is remotely reachable over the network but requires a valid low-privileged account (CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:L/A:L per the assessed vector), and the realistic impact is improper manipulation of the moderation/quarantine state of DM and GM posts rather than disclosure of message contents or full data compromise. There is no public exploit identified at time of analysis and no confirmed active exploitation (no CISA KEV entry); this is a moderate-severity authorization weakness that Mattermost tracks under advisory MMSA-2026-00756.
Stored cross-site scripting in the Tourfic WordPress plugin (versions up to and including 3.0.1) allows attackers with subscriber-level or WooCommerce customer credentials to inject malicious JavaScript into comment content, which executes when victims view affected pages, REST API responses, or RSS feeds. Because WooCommerce customer registration is open without admin approval, any visitor can effectively become an attacker; the payload bypasses the plugin's own wp_kses_post() sanitization when served via the unauthenticated GET /wp-json/wp/v2/comments endpoint (content.rendered) and comment RSS feeds. No public exploit code or active exploitation has been identified at time of analysis.
Authentication bypass in Xiaozhi-esp32-server-java through 6.1.0 lets unauthenticated remote attackers request /api/device/ota with a valid MAC address and receive a device WebSocket token, enabling access to that device's channel to read private role system prompts, write conversation history, consume LLM/STT/TTS credentials, and overwrite device metadata. No public exploit was identified at time of analysis, and the assessed risk is MEDIUM rather than top-tier because the attacker must know or enumerate a provisioned device MAC and the OTA endpoint must be network-reachable; impact is bounded to a single device's WebSocket channel rather than full server compromise.
Niefy wx-api through 0.8.2 allows unauthenticated remote attackers to impersonate WeChat followers by presenting a raw openid cookie to fan-facing controllers that Apache Shiro excludes from its authentication filter chain - specifically /wxUser/getUserInfo and /wxUserTags. Anyone who already knows a victim's openid can read that fan's nickname, phone number, unionid and tag assignments, and add or remove tags on the account, giving cross-tenant fan PII exposure and unauthorized state changes. Impact is bounded to per-fan confidentiality and integrity loss with no code execution or availability effect, and exploitation is non-trivial because openids are opaque, non-sequential identifiers, so blind mass exploitation is impractical without a prior source of valid openids; this is a moderate-priority issue with no public exploit identified at time of analysis, although a public technical writeup describing the attack path and an associated cross-tenant IDOR is referenced.
Unauthenticated clients that can reach an Umbraco CMS backoffice can complete the SignalR negotiate handshake against the PreviewHub and BackofficeHub endpoints and receive the refreshed notifications those hubs broadcast, disclosing content keys for unpublished content and revealing when editors are actively working. The flaw affects Umbraco CMS releases before 17.7.1 and before 18.2.1 and is rated medium severity (vendor CVSS 4.0 base 6.3; independently assessed CVSS 3.1 AV:N/AC:H/PR:N/UI:N/C:L), with the practical leak limited to read-only metadata that only flows while editorial activity is in progress. No public exploit identified at time of analysis, and no evidence of confirmed active exploitation; the vendor has published patched releases and upstream fix commits.
An out-of-bounds read (heap over-read) in the FreeRDP Windows RAIL client (wfreerdp) versions 3.0.0 through 3.31.x can be triggered by a malicious or compromised RDP server that sends a full-length WINDOW_ORDER_FIELD_TITLE window-title update lacking NUL termination, a value that wf_rail_window_common passes directly to SetWindowTextW. This reverse (server-to-client) attack requires the victim to initiate a connection to the attacker-controlled server with RemoteApp/RAIL in use; impact is limited to window-title corruption, disclosure of adjacent heap memory, and client crashes, with no code execution or integrity impact assessed (CVSS:3.1/AV:N/AC:H/PR:N/UI:R/S:U/C:L/I:N/A:L). No public exploit code or confirmed active exploitation was identified at time of analysis.
Host header injection in Nezha 2.1.0 through 2.3.14 lets an attacker who can control or spoof the Host header on a victim's MCP request redirect file uploads and capture one-time download tokens that can be replayed to exfiltrate files. The flaw is in mintTransferTool, which builds the fs.download_url and fs.upload_url transfer links from the unvalidated Host header, so the observable impact is limited to low confidentiality and integrity loss rather than code execution. Exploitation is unauthenticated per the CVSS vector (PR:N) but conditional on the attacker influencing the reflected Host value - for example through a misconfigured or trusting reverse proxy, DNS rebinding, or a man-in-the-middle position on the victim's path - and on the victim actually invoking the MCP transfer tools. No public exploit code or confirmed active exploitation was identified at time of analysis, and the vendor has shipped a fix in Nezha 2.3.15.
Multimodal prefix-cache poisoning in SGLang through 0.5.21 lets remote unauthenticated attackers alias one image's cached KV entry onto a different image by supplying a caller-controlled mm_hashes value on the /generate endpoint, enabling cross-request image content disclosure or response corruption in shared multi-tenant deployments. The behavior was the default in affected releases because the tokenizer manager trusted GenerateReqInput.mm_hashes as the multimodal prefix-cache identity before the fix; exploitation requires the attacker to know or predict the victim's content hash, which is the main limiting factor (AC:H/AT:P), and only deployments running multimodal models with the KV/prefix cache active and serving multiple parties are meaningfully exposed. No public exploit code and no confirmed active exploitation were identified at time of analysis, and independent assessment rates this a moderate-priority multi-tenant data-integrity/disclosure flaw rather than an urgent one.
Reflected cross-site scripting in the Gwolle Guestbook WordPress plugin (all versions up to and including 5.1.1) allows unauthenticated attackers to inject arbitrary script into pages generated by the plugin's permalink handling, executing in the browser of any victim lured into clicking a crafted link. Tracked as CWE-79 and rated CVSS 6.1 (AV:N/AC:L/PR:N/UI:R/S:C/C:L/I:L/A:N), the flaw requires no authentication but is gated on victim interaction, which keeps it a medium-severity, per-victim attack requiring individual luring. No public exploit code was identified at the time of analysis, and the impact is confined to the victim's browser session (script execution in the site's origin context) rather than granting server-side access.
DOM-based cross-site scripting in the Tutor LMS eLearning and online course plugin for WordPress (versions up to and including 4.0.8) lets unauthenticated attackers inject arbitrary JavaScript that runs in a victim's browser when the crafted 'redirect_url' parameter is processed by the Topbar component's back-button handler. Because the CVSS vector carries UI:R, the attacker must convince a victim to open a specially crafted URL and trigger the back-button action - the flaw cannot be exploited passively against a server, and no public exploit code has been identified at time of analysis. Scored 6.1 (AV:N/AC:L/PR:N/UI:R/S:C/C:L/I:L/A:N) and assessed as a moderate-severity client-side issue rather than a top-tier priority, it still allows session-scoped script execution and potential redirect or content-injection abuse on any WordPress site running the affected plugin version.
Denial of service and limited memory disclosure in GIMP's Bohemia Interactive (PAA) texture importer, triggered when a user opens or previews a specially crafted .paa file through the affected plug-in. An integer overflow in the image-dimension calculation during decompression causes an undersized heap buffer to be allocated, and the subsequent out-of-bounds read crashes the plug-in process and may expose fragments of heap memory (CVSS 3.1 base 6.1, AV:L/PR:N/UI:R, C:L/A:H). The attack requires the victim to deliberately load the malicious file locally, so there is no public exploit identified at time of analysis and no confirmed active exploitation (CISA KEV) - this is assessed as a moderate-severity, low-priority issue in practice.
Authenticated cross-work attachment disclosure in O2OA through 10.0.2-ce lets any low-privileged user, using the endpoint jaxrs/attachment/download/{id}/work/{workId}, read files attached to other users' process works by supplying an attachment UUID they should not be able to reach. The root cause is an IDOR (CWE-639) in x_processplatform_assemble_surface, where the download actions never confirm the requested attachment belongs to the referenced work or draft, so the confidentiality of stored documents is exposed (vector reports C:H with no integrity or availability impact). Exploitation requires a valid low-privileged session plus knowledge of a specific, non-sequentially-enumerable attachment UUID, which keeps mass exploitation impractical; no public exploit code was identified at time of analysis and there is no CISA KEV entry confirming active exploitation.
Authenticated HQL injection in Erupt through version 2.3.0 lets a user who already holds query permission on an entity with a String-keyed linkTree target escape that entity's tree-scoped data filter and read records belonging to other mapped entities. The flaw is reached by posting unescaped quote characters into the linkTreeVal parameter of POST /erupt-api/data/table/{erupt}, which allows attackers to append always-true OR clauses or boolean subqueries to the generated Hibernate query. Exploitation is not trivial: it depends on valid low-privileged credentials plus a specific data-model configuration (a String-keyed linkTree target with query permission), and the impact is limited to confidentiality - no write or availability effect - so this is best treated as a moderate-priority authorization/data-exposure issue, and no public exploit code has been identified at time of analysis.
Directory traversal in the WordPress plugin 'Debloat - Remove Unused CSS, Optimize JS' (all versions through 1.3.0) allows remote attackers to read arbitrary files on the server via crafted input in the 's' search parameter. According to our independent assessment, the CVSS vector is AV:N/AC:H/PR:N/UI:N (unauthenticated, no user interaction), so the trigger is a single crafted GET request to the public search endpoint - but only after a site administrator has enabled the plugin's 'Optimize CSS' feature with the default 'Inline Google Fonts' option active, which is why the attack is rated AC:H and conditional rather than universally exploitable. Impact is read-only confidentiality loss (C:H, I:N, A:N), typically exposing wp-config.php and other sensitive files subject to web-server/PHP file permissions; no public exploit code has been identified at time of analysis and there is no KEV listing, making this a moderate, conditionally-exploitable issue rather than a top-tier emergency.
Resource exhaustion in FreeRDP's rdpecam camera-redirection channel affects client builds from 3.30.0 up to (but not including) 3.32.0, where a malicious or compromised RDP server can send a StartStreamsRequest for an unadvertised camera media type with oversized dimensions, forcing large pending-sample allocations and reconfiguring the local camera through the HAL. Impact is availability-only (client memory exhaustion leading to hang or crash) with no code execution or data disclosure, and exploitation requires the victim to have camera redirection enabled and to intentionally connect to an untrusted server. Independent assessment rates this a moderate-severity denial-of-service issue rather than a top-tier priority, with high attack complexity; no public exploit identified at time of analysis.