Subscriber-level users on WordPress sites running Image Hub 1.2.0 and earlier can extract six third-party image-service API credentials from admin-page HTML and then use an exposed AJAX nonce to overwrite those stored credentials. The plugin emits Giphy, Openverse, Pexels, Pixabay, and Unsplash key options via wp_localize_script, and the image_hub_update_settings AJAX handler lacks a capability check, allowing the same low-privilege attacker to modify settings through update_option(). No public exploit identified at time of analysis; the assessed risk is a moderate, patch-priority information-disclosure and integrity issue rather than an emergency.
Openwave Webtop Media Transcoder fails to enforce the administrator-defined file-type restrictions that are meant to limit which documents may be converted, so an authenticated webmail user can reach the transcoding function outside the normal Webmail UI and have unintended document types processed. The flaw carries a CVSS v3.1 score of 4.3 (AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:L), meaning any exploitation is confined to a low availability impact - no loss of confidentiality or integrity - and it requires a valid authenticated webmail account (PR:L). No public exploit code has been identified at time of analysis, KEV status is not indicated, and no EPSS score was supplied; this is a low-severity, authenticated-only issue that does not warrant priority remediation.
Authorization bypass in the LPagery (Bulk Page Generator) plugin for WordPress, affecting all versions up to and including 3.1.0, allows any authenticated user with subscriber-level privileges or above to create, modify, and delete LPagery-managed pages and bulk page-generation jobs that should be restricted to administrators. The plugin fails to enforce authorization checks (CWE-862) on its LPagery REST/OAuth endpoints, and on default fresh installations where the Freemius integration has not been opted into, the AES-256-CBC key used to validate Bearer tokens collapses to the publicly known SHA-256 hash of an empty string, making those tokens forgeable by any authenticated user. No public exploit code has been identified at time of analysis and the issue is not in CISA KEV; the CVSS 3.1 base score is 4.3 (AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:L/A:N), consistent with a low-integrity-impact flaw gated behind an authenticated account.
Authenticated vendor-level users of the Dokan multivendor marketplace plugin for WordPress (all versions up to and including 5.2.1) can reset the WordPress account password of any customer who has placed at least one order with their store, by supplying an arbitrary value to the 'password' parameter of the plugin's REST customers endpoint. The flaw is a CWE-639 insecure direct object reference: the endpoint's object-level guard only verifies that the target user is one of the requesting vendor's own customers, and never blocks changes to login-identity fields, so a vendor can overwrite the stored password hash and then log in as that customer for a limited account takeover. Vendor access required to exploit this can be self-obtained through the public 'I am a vendor' registration flow when that option is enabled, but the practical impact is bounded - the attacker must already hold a vendor or custom-level account and can only target customers with an existing order relationship; no confirmed active exploitation (not in CISA KEV) and no public exploit code identified at time of analysis.
Authenticated Mattermost users can degrade message post-processing latency for every member of the channels they share by storing an oversized list of custom mention keywords (notify_props.mention_keys) on their own profile, which the server then matches against each post. Because the matching routine scales quadratically and multibyte/Unicode keywords add per-character cost, and neither the number nor the size of stored keywords is capped, each post the attacker's channel peers send or receive is processed far more expensively on the attacker's behalf. Impact is confined to availability degradation (CVSS 4.3, AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:L; CWE-770); no data exposure or crash is involved, valid authentication plus at least one shared channel is required, and no public exploit has been identified at time of analysis.
An authorization bypass in Arcane before 2.11.0 lets anyone holding a valid environment-scoped access token reach administrator-only routes by presenting that token in the X-Arcane-Agent-Token header, which skips the WithAdminRequired check. The exposed surface is the manager's Go pprof diagnostic interface under /api/debug/pprof/*, so the practical impact is read-only disclosure of runtime and profiling data plus possible profiling overhead, not code execution or configuration change (CVSS 4.3, CWE-863). Exploitation is authenticated - the attacker must already possess a legitimate environment access token (CVSS PR:L) - and no public exploit code has been identified at time of analysis; the flaw is corrected in Arcane 2.11.0.
Authenticated WordPress users holding Contributor-level access or higher can publish, unpublish, or bulk-delete any map created with the Wp Map Block plugin, including maps owned by administrators, because the plugin's Maps_Controller REST endpoints through version 3.0.0 rely on a permission callback that only checks edit_posts and never validates object ownership or publish-level capability. Successful abuse either hides published maps from every public page site-wide or forces draft maps to become publicly visible, and the bulk branch additionally deletes wpmb_map posts outright. Impact is confined to the integrity of map post status with no code execution path; no public exploit code was identified at time of analysis, and this is a medium-severity broken-access-control issue (CVSS 4.3) rather than a critical compromise.
Improper access control in Devolutions Server 2026.3.8.0 and earlier lets an authenticated remote user who has only the general Edit permission - but who was explicitly not granted the Edit asset permission - change an asset's stored contact password by issuing a crafted request straight to the partial-save endpoint, skipping the UI-level authorization check (CWE-863). The consequence is confined to partial integrity loss (CVSS 4.3, AV:N/AC:L/PR:L/UI:N, C:N/I:L/A:N): no data disclosure or availability impact, and the attacker must already hold a valid account with Edit rights, so this is not a privilege-escalation path to full administrative control. No KEV listing, no active exploitation and no public exploit code were identified at time of analysis, and SSVC/EPSS-style signals (Exploitation: none, Automatable: no, Technical Impact: partial) place this at low real-world priority; the flaw is fixed in 2026.3.10.
Authenticated PAM checkout users on Devolutions Server 2026.3.8.0 and earlier can manipulate the just-in-time (JIT) elevation settings embedded in a checkout update request to delete arbitrary Active Directory groups or preserve elevated privileges that should have been dropped at checkout. The flaw is a server-side trust issue: the application accepts client-supplied JIT elevation parameters without re-validating them, so an attacker only needs a provisioned account that already holds PAM checkout access. There is no public exploit identified at time of analysis, exploitation is not listed in CISA KEV, and the vendor has released a patched version; CVSS scores are modest (4.3 vendor base; independent assessment 4.3 with Scope:Changed, reflecting limited integrity and availability impact rather than confidentiality loss).
Information disclosure in Mattermost release lines 11.7.x, 11.8.x, 11.9.x and 11.10.x (up to 11.7.10, 11.8.5, 11.9.1 and 11.10.1 respectively) allows an authenticated user to learn the channel ID of a private channel that another user is actively viewing. The user status API endpoints return an unsanitized active-channel field, so any provisioned account on the same server - with no membership in the target private channel, which is precisely the flaw - can query another user's status and observe which private channel that user is currently reading. The exposure is limited to the channel identifier itself, not its contents, name or membership, but it still reveals the existence of private channels and correlates named users with sensitive collaboration spaces. Rated CVSS 4.3 (AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:N/A:N), this is a low-to-moderate urgency issue rather than an emergency, and no public exploit has been identified at time of analysis.
Flagged-post review status in Mattermost can be inferred by authenticated users who lack content-reviewer permissions, because the affected API performs other validations before checking the caller's permissions, producing a distinguishable response for flagged versus unflagged posts. The issue affects 11.10.x through 11.10.0, 11.9.x through 11.9.1, 11.8.x through 11.8.5, and 11.7.x through 11.7.9 (advisory MMSA-2026-00754), and exploitation requires a valid account on a deployment that actually uses the post-flagging/moderation feature, with attacker access limited to a single binary signal and no exposure of post content or reviewer details. Rated CVSS 4.3 (AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:N/A:N); no public exploit identified at time of analysis, and this is a genuinely low real-world priority.
Private team metadata in Mattermost Server can be enumerated by any authenticated user who lacks the permission to list private teams, because the team search functionality fails to exclude private teams from public search results. An attacker with a valid low-privileged account can send a modified team-search request and retrieve the ID, name, and display name of private teams they are not authorized to see. The issue affects 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.x through 11.10.0, and is rated low severity (CVSS 3.1 4.3, AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:N/A:N) with only metadata disclosure - no message content, membership lists, or credentials are exposed; no public exploit was identified at time of analysis.
Improper authorization in Apache Answer 2.0.2 and earlier allows holders of a valid global API key to retrieve answer content tied to questions that are hidden, deleted, or still pending review. The defect is in the MCP (Model Context Protocol) answer-retrieval tool, which returns content without honoring the per-question visibility state, so an authenticated key holder - an insider, a third-party integration, or anyone using a leaked key - can read material the Q&A platform intended to withhold. It is not reachable by anonymous remote attackers, impact is limited to information disclosure with no modification or denial of service, and no public exploit identified at time of analysis; CISA KEV does not list this issue.
Mattermost servers running 11.7.x through 11.7.9, 11.8.x through 11.8.5, 11.9.x through 11.9.1, and 11.10.x through 11.10.0 fail to restrict Slack import account binding to system-administrator-initiated imports, so an authenticated attacker holding team admin privileges can force-add existing Mattermost users to their own team and then forge posts attributed to those users' identities. The attacker only needs to be an authenticated insider with the team-admin role and to supply a crafted Slack export archive containing the target users' email addresses; no victim interaction is required (CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:L/A:N, base score 4.3). Impact is confined to integrity - unauthorized team membership and identity-forged content - with no confidentiality loss or availability impact, and there is no public exploit identified at time of analysis and no CISA KEV listing, making this a moderate-priority insider-abuse issue rather than an urgent remote-code-execution or data-breach emergency.
Topic-title disclosure in the Discourse Doc Categories plugin lets an authenticated user who holds edit rights on a documentation index topic learn the titles of private messages and topics in read-restricted categories that the user is not authorized to view. The index refresh routine resolves linked topic titles without enforcing viewer authorization, so the leaked titles appear in the generated index structure even though the underlying topic contents remain inaccessible. Rated CVSS 4.3 (AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:N/A:N) with no public exploit identified at time of analysis and no confirmed active exploitation; the issue is largely confined to non-default Discourse instances that have this plugin installed and actively used, and the fix is a targeted authorization filter in the upstream commit b15d4de02ad29d7c9525346afee51fef4a2c605a.
Apache Answer through 2.0.2 can continue to honor sessions that should have been invalidated after a central account was suspended or deleted, because the User Center login integration creates first-login sessions that omit the user's external identity. An authenticated user (CVSS vector PR:L) who already holds such a session can keep issuing requests that bypass the central suspension/deletion check until the session expires or is manually revoked. Exploitation requires the User Center integration to be configured and the attacker to hold a valid pre-suspension session; no public exploit code or confirmed active exploitation was identified at time of analysis, and deployments not using User Center are unaffected. The vendor has released version 2.0.3, which fixes the issue.
Apache Answer through 2.0.2 fails to invalidate visit tokens after a user's access is revoked, allowing an authenticated user to continue accessing previously known private uploaded files using a stale token. This insufficient session expiration flaw (CWE-613) requires the attacker to have been a legitimate user who obtained a valid token before revocation and to already know the specific file locations; it does not allow discovery of unknown private content. No public exploit code or active exploitation has been identified at time of analysis, and upgrading to version 2.0.3 remediates the issue.
Improper authorization in the similar-question search of Apache Answer through 2.0.2 allows an authenticated user to retrieve metadata belonging to non-public questions they are not entitled to see. The issue is remotely reachable, low complexity, and requires only a valid account plus use of the similar-question search feature, matching the assessed vector CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:N/A:N, so impact is confined to information disclosure of question metadata with no effect on integrity or availability. No public exploit code has been identified at time of analysis and there is no confirmation of active exploitation (not in CISA KEV); the vendor-stated fix is Apache Answer 2.0.3.
Unauthenticated information disclosure in Apache Answer through 2.0.2 allows remote attackers to read comments tied to non-public (restricted) content by browsing public user profiles, which are rendered without re-checking the visibility of the parent content. The issue is network-reachable, requires no authentication or user interaction (CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:N/A:N), and is bounded to comment disclosure - no content bodies, credentials, or write capability are exposed, and instances hosting only public Q&A material have little to leak. No public exploit code has been identified at time of analysis and there is no indication of active exploitation; users should upgrade to version 2.0.3, which fixes the issue.
Apache Answer through 2.0.2 leaks metadata belonging to hidden questions to unauthenticated visitors through publicly reachable answer-profile pages, letting anyone who can reach the instance learn about hidden questions and read metadata that was never meant to be public. The disclosure is metadata only, not the full hidden-question content, so real-world sensitivity depends on whether the deployment treats the existence of hidden questions as confidential; no public exploit code has been identified at time of analysis and the issue is not listed in CISA KEV. The assessed severity is low-to-moderate (CVSS 3.1 5.3, AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:N/A:N), with the vendor-released fix in Apache Answer 2.0.3.
Missing visibility authorization checks in Apache Answer through 2.0.2 let remote, unauthenticated attackers retrieve the invited-user list attached to non-public questions that they should not be able to see. The flaw is limited to information disclosure - an attacker who can reach the invite-lookup functionality and reference a private question identifier can enumerate the users invited to that question, but cannot modify content or disrupt the service. The vendor rates this low severity, our independent assessment concurs (CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:N/A:N), no public exploit code has been identified at time of analysis, there is no indication of confirmed active exploitation, and no EPSS score was supplied; upgrading to Apache Answer 2.0.3 resolves the issue.
Incorrect permission mapping in Apache Answer through 2.0.2 allows an authenticated user who holds only the question.unpin permission to hide or restore questions without holding question.hide or question.show. The flaw is only reachable in non-default installations where an administrator has built a custom role that grants question.unpin while separately withholding question.hide and question.show; where Apache Answer ships its stock bundled permissions, the mis-gated check is never satisfied and the issue is not exploitable. No public exploit code has been identified at time of analysis and this is not confirmed actively exploited (not in CISA KEV); the assessed impact is bounded to integrity of question visibility state (CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:U/C:N/I:L/A:N), and upgrading to 2.0.3 resolves it.
Stored cross-site scripting in Apache Answer through 2.0.2 allows an attacker who can control the contents of a data source consumed by the question importer to persist raw, unsanitized HTML that is later rendered to every viewer of the imported question, including privileged users. The importer bypasses the normal Markdown-to-HTML sanitization path applied to ordinary user content, so the injected markup survives storage and executes in the browser of anyone who opens the affected question - no additional victim interaction beyond viewing is required, but the attacker must first reach the administrative/trusted import workflow. The issue is rated low severity by the vendor and carries an assessed vector of CVSS:3.1/AV:N/AC:L/PR:H/UI:N/S:C/C:L/I:L/A:N; no public exploit code has been identified at time of analysis and there is no indication of active exploitation. A vendor-released patch is available in Apache Answer 2.0.3.
Apache Kylin 4.0.0 through 5.0.4 ship a default startup profile that leaves the security (authentication) settings disabled, so a Kylin instance launched with that default profile may be reachable over the network without the intended access controls; the advisory explicitly states that any deployment with the security settings profile turned on is not impacted. The assessed vector (CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:N) makes this a low-severity configuration-hardening flaw with limited confidentiality and integrity impact and no availability impact, and no public exploit code was identified at time of analysis. The vendor has fixed the issue in Apache Kylin 5.0.5, tracked as KYLIN-6113 (resolution: Fixed).
Metadata tampering and sensitive-information exposure in Apache Kylin 4.0.0 through 5.0.4 let a network-reachable attacker replace job metadata through the job service/job-metadata interface, with the assessed CVSS vector (AV:N/AC:H/PR:N/UI:N/S:U/C:L/I:L/A:N) indicating unauthenticated access and low confidentiality and integrity impact. Exploitation is gated by deployment exposure: the vendor explicitly scopes impact as minor unless job services are exposed to the public internet, which it describes as uncommon. The independent assessment rates real-world priority as LOW, and no public exploit code or CISA KEV listing was identified at time of analysis; upgrading to 5.0.5 remediates the issue.
Insufficient validation of JDBC data source configuration in Apache Kylin 4.0.0 through 5.0.4 allows an authenticated user who can create or modify a JDBC source definition to influence how Kylin resolves and connects to external databases, yielding limited confidentiality and integrity impact (assessed CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:N). The vendor explicitly characterizes the impact as minor unless Kylin backend services are exposed to a public network, making network exposure the primary risk amplifier rather than the flaw itself. No public exploit code or confirmed active exploitation has been identified at time of analysis, and the fix is available in Apache Kylin 5.0.5 (tracked as KYLIN-6111).
SQL injection in Apache Kylin 4.0.0 through 5.0.4 lets an authenticated user who can reach the table metadata API pass unsanitized 'tables'/'names' input that is concatenated into a SparkSQL command, yielding limited data exposure and limited integrity impact (assessed CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:N). The vendor explicitly scopes real-world impact as moderate because the default and most common deployment is query-only mode, which constrains attackers largely to reading data rather than writing or taking over the cluster; non-default modes would likely increase impact. No public exploit code or confirmed active exploitation has been identified at time of analysis, and the issue is resolved by upgrading to Apache Kylin 5.0.5.
Server-side request forgery in the Apache Kylin job-driver UI proxy (versions 4.0.0 through 5.0.4) allows an attacker to induce the Kylin job service to issue requests to attacker-chosen destinations, potentially reaching internal services or cloud metadata endpoints that the Kylin host can access. The vendor explicitly scopes real-world risk to deployments where the job service and its UI proxy are directly exposed to the public network, describing that configuration as uncommon and stating that in-house, segmented Kylin deployments are 'mostly not impacted'; exploitation prerequisites are therefore dominated by network exposure rather than by any inherent weakness in default internal deployments. No public exploit code, EPSS score, KEV listing, or vendor-published CVSS vector was available at time of analysis, so there is no public exploit identified at time of analysis and the issue carries a moderate severity rating; upgrading to 5.0.5 removes the condition (tracked as KYLIN-6104).
Arbitrary file deletion outside the intended backup directory is possible in Apache Kylin 4.0.0 through 5.0.4 via a path-confinement bypass in the metadata backup deletion feature, allowing an attacker who can reach that function to delete files elsewhere on the Kylin host (assessed CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:H/A:H - integrity and availability impact, no confidentiality impact; PR:L indicates low-privilege authenticated access). The dominant limiting factor is exposure: Apache states the issue 'most' affects Kylin services directly reachable from the public network, which the vendor describes as uncommon, while in-house, non-exposed deployments are mostly not impacted. No public exploit code, EPSS score, or CISA KEV listing was identified at time of analysis, and a vendor fix is available in Apache Kylin 5.0.5 (tracked as KYLIN-6105).