Insecure direct object reference in WhaTicket Community (all builds through commit cf04795) lets any authenticated agent read another agent's complete WhatsApp ticket history and post replies into that conversation by simply substituting a victim's ticketId in GET /tickets/:ticketId or the /messages/:ticketId endpoints. The flaw is a missing authorization check rather than a memory-safety or injection issue, so it is trivially reachable by any logged-in agent account and requires no user interaction or unusual configuration; the assessed CVSS 3.1 vector (AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:N) is consistent with the published CVSS 4.0 score of 5.3, placing this at moderate severity rather than critical. No public exploit code was identified at time of analysis and the product does not appear in CISA KEV, although the disclosure is publicly documented in a HackMD writeup and a VulnCheck advisory, which raises the likelihood of opportunistic abuse on internet-exposed instances.
Missing authorization in the Gemini Labs Site Reviews WordPress plugin (versions through 8.3.4) allows unauthenticated remote attackers to perform actions that should require elevated privileges, because affected endpoints do not correctly enforce access-control checks. Per the assessed vector (AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:L/A:N), exploitation needs no authentication or user interaction against a default install, but impact is limited to low-integrity tampering with plugin-controlled data - no confidentiality or availability effect. No public exploit code and no confirmed active exploitation were identified at time of analysis; realistic risk is constrained to sites that both run the vulnerable plugin version and expose the affected functionality to the network.
Frappe Framework before 15.120.0 and 16.x before 16.33.0 does not enforce its per-user restrict_ip IP allowlist for non-session authentication methods, allowing an authenticated attacker with a valid API key, OAuth token, or header-based credential to authenticate from an IP address outside the configured allowlist. This bypasses network-based access restrictions but requires both a valid non-session credential and a targeted user account that has restrict_ip configured; it is a medium-severity access-control flaw (CVSS 4.0 base 5.3) with limited confidentiality and integrity impact. No public exploit identified at time of analysis.
Open WES through 1.0.0 does not enforce warehouse-scope authorization in GlobalGatewayFilter, so an authenticated user who holds the menu permission for the stock endpoints can freeze or unfreeze container stock belonging to any warehouse by enumerating sequential IDs on the freezeContainerStock and unfreezeContainerStock endpoints. This is a moderate-severity authorization bypass (CWE-639 IDOR; CVSS 5.3) with integrity and availability impact only, and no public exploit identified at time of analysis. Only installations up to and including version 1.0.0 are affected.
Missing authorization in Arkon 0.9.2 and earlier lets an authenticated employee holding only the default skill:read permission read another department's skill version metadata via GET /api/skills/{slug}/versions, exposing version numbers, hashes, storage paths and changelogs that should be scoped to the owning department. The root cause is an omitted can_access_skill check in the list_skill_versions handler, i.e. a CWE-639 authorization bypass through a user-controlled key, and no elevated role, user interaction, or non-default configuration is required beyond knowing or enumerating a target skill slug. Exploitation is limited to existing insiders with a valid account on a multi-department deployment, so single-department or single-tenant installs have no cross-boundary data to leak; no public exploit code was identified at time of analysis and the issue is assessed as a genuine but modest-severity information-disclosure weakness rather than a critical compromise vector.
Arkon deployments through version 0.9.2 allow authenticated, low-privileged employees to read and tamper with other departments' source ingestion records because the by-id routes in app/routers/sources.py (plan, progress, PATCH, retry, approve-extraction) omit the can_access_document authorization check. An attacker with a valid account only needs to know or enumerate a foreign source ID to disclose derived wiki titles and summaries and to change ingestion state; raw source content is not exposed and there is no availability impact. Per the independent assessment this is a moderate authorization-layer (IDOR) flaw requiring PR:L credentials, with no public exploit identified at time of analysis and no CISA KEV listing.
Insecure direct object reference in yshop-crm through 2.1.3 allows an authenticated sales user to read, overwrite, delete, or release into the public pool customer records that belong to other users, simply by enumerating the id parameter on the CRM customer get, update, delete and do-open endpoints. The assessed severity is medium (CVSS 3.1 5.3, vector AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:N) because the attacker must already hold a valid sales-privilege account in the admin CRM module; no special configuration or user interaction is required, and every deployment at or below version 2.1.3 with provisioned sales accounts is exposed. No public exploit identified at time of analysis, although a public technical writeup documents the vulnerable endpoints and the associated service-layer code.
Information disclosure in Exrick XBoot through 3.3.4 allows any authenticated user to defeat server-side role data-scope restrictions by appending the parameter openDataFilter=false to the department getByParentId or search endpoints, causing the API to return the entire organization's department tree and names outside the caller's assigned scope. The underlying flaw is an improper-authorization condition (CWE-863) rated medium severity (CVSS 5.3, CVSS:4.0 AV:N/AC:L/AT:N/PR:L/UI:N/VC:L), and the impact is limited to organizational-structure disclosure - no credential exposure, data modification, or availability effect. Exploitation requires a valid logged-in XBoot account whose role carries data-scope restrictions (PR:L); no public exploit code was identified at time of analysis, although a public technical write-up of the bypass and the vulnerable source code are both referenced in the advisory data.
Broken access control in Faveo Helpdesk through 2.0.3 lets any authenticated agent read ticket conversations, customer personally identifiable information, and internal agent notes belonging to other departments by enumerating sequential ticketid values against the GET /ticket/tooltip endpoint (CWE-639, IDOR). The flaw is read-only, requires a low-privileged authenticated agent account (PR:L in the published CVSS vector), and works against default configurations because no special feature toggle or non-default setting is needed; impact scales with how many ticket IDs the attacker can walk. No CISA KEV entry exists for this CVE and there is no confirmed active exploitation, though a public technical writeup documents the attack path; the assigned CVSS score is 5.3.
Authenticated, low-privilege users of Cube-studio through 2026.09.01 can read or clone pipelines belonging to other project groups by enumerating sequential pipeline IDs through the web, log, monitoring, pod, and copy_pipeline endpoints, exposing DAG definitions, Kubernetes cluster names, namespaces, Argo workflow IDs, and pod names. Classified as an authorization bypass through a user-controlled key (CWE-639) and scored CVSS 5.3 (CVSS:4.0 AV:N/AC:L/PR:L/UI:N/VC:L), the issue requires a valid low-privilege account in a multi-tenant deployment where multiple project groups coexist, and its impact is confined to cross-group metadata disclosure and pipeline cloning - no code execution, system compromise, or modification of existing data is enabled. No public exploit identified at time of analysis; risk is highest in shared, multi-tenant installations and negligible in single-group deployments where there is nothing to enumerate.
Missing authorization in REBUILD through 4.5.0 allows any authenticated low-privileged user to read comments and attachment references from tasks in private projects by supplying a valid task ID to the /project/comments/list endpoint, bypassing project membership checks (IDOR, CWE-639). Exploitation requires only a valid account at any privilege level and knowledge of a task ID; the attacker need not be a member of the targeted private project. No public exploit code or confirmed active exploitation was identified at time of analysis, and EPSS data was not provided; impact is limited to confidentiality exposure of task comments and attachment references, with no modification or deletion possible.
Cross-tenant information disclosure in Steedos Platform through 3.0.15-beta.51 allows any authenticated low-privilege user to read other tenants' report definitions by issuing a single GET /api/v1/reports request. The root cause is that the 'reports' object is declared enable_space_global, and objectql's callAdapter skips space (tenant) filtering on find/findOne for such objects, so the API returns report names, descriptions, and object and field API names belonging to other tenants. The exposure is limited to metadata rather than underlying records, the attacker must already hold valid credentials on a genuinely multi-tenant instance, and no public exploit identified at time of analysis; exploitation status is not confirmed actively exploited (CISA KEV).
BeikeShop through 2.0.0.31 contains an insecure direct object reference (IDOR) in the markAsShipped handler of RmaController that allows authenticated customers to modify other customers' return shipping details. By submitting sequential RMA IDs to the account/rmas/{id}/mark_as_shipped endpoint, an attacker with a valid customer account can overwrite courier names and tracking numbers and mark approved returns as shipped. The vendor CVSS 4.0 base score is 5.3 (PR:L, VI:L), matching an independent assessment of CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:L/A:N; no public exploit code has been identified at the time of analysis, though a technical write-up is available.
Cross-tenant insecure direct object reference in ChatWiki 2.9.6 and earlier allows any authenticated tenant user to read another tenant's form filter rules and overwrite that tenant's filter sort order by supplying arbitrary or sequential form_id values to the form-filter management endpoints getFormFilterList, getFormFilterInfo and updateFormFilterSort. Exploitation requires a valid low-privilege ChatWiki account on a multi-tenant deployment (PR:L in the assessed vector), so the flaw is not reachable pre-authentication, and the impact is confined to form-filter configuration data rather than code execution or administrative takeover. No public exploit code identified at time of analysis, though a detailed third-party technical writeup and the vulnerable source code are publicly linked in the advisory references.
An insecure direct object reference (IDOR) in PickMall lilishop through 5.0.0 allows authenticated buyers to read other buyers' invoice records-including invoice titles, taxpayer identification numbers, order numbers, member IDs and member names-by supplying another buyer's receipt ID to GET /buyer/trade/receipt/{id}. The flaw resides in ReceiptServiceImpl.getDetail, which performs no ownership check, and can only be exploited by a user with a valid buyer account who can guess or enumerate valid receipt IDs. This is a moderate-severity information-disclosure issue (CVSS 5.3, CWE-639); it is not listed in CISA KEV and no public exploit code or active in-the-wild exploitation has been confirmed, though a public technical writeup exists.
Authenticated API users of Aureus ERP through 1.6.0 can create new records or move existing ones into companies they are not entitled to by submitting a foreign company_id to plugin REST endpoints, breaking the multi-company data segregation that the ERP relies on. The flaw is an IDOR/authorization bypass (CWE-639) in the Laravel Form Request validation of partners, sales, inventories, and accounts handlers, which checked only that the supplied company_id existed and never that it belonged to the caller's allowed companies; it is exploitable only by an authenticated account with API access and only on deployments that actually use more than one company. Rated 5.3 (CVSS 4.0) and independently assessed as CVSS 3.1 AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:N, its real-world priority is moderate-to-low, patch code is available upstream, and no public exploit was identified at time of analysis.
Missing authorization in the 10up Distributor WordPress plugin before 2.3.2 allows authenticated users with Contributor-level privileges to invoke AJAX handlers that were intended only for administrators, enabling them to send stored external-connection credentials to an attacker-controlled server via dt_verify_external_connection or to create new external connections via dt_begin_authorization. Exploitation requires a valid low-privilege login (CVSS vector PR:L, so authenticated, not anonymous), and the credential-theft path only works on sites that already have a configured external connection holding stored credentials; sites without pre-existing connections or without untrusted Contributor-level accounts have no exposed attack surface. No public exploit has been identified at time of analysis, and the vendor has released a fix in version 2.3.2.
Missing authorization in TDuck Survey Form through 5.3 allows any authenticated, low-privileged user to read another form owner's raw settings by calling GET /user/form/setting/{key} with the 8-character form key embedded in a public survey fill link. The endpoint returns the form write password, the owner's notification e-mail address and associated WeChat openIds, which can be chained into form tampering and targeted phishing; a valid account is required (CVSS PR:L, so risk is higher where self-service registration is open), no user interaction is needed, and no public exploit identified at time of analysis. Severity is modest rather than critical because disclosure is confined to per-form settings and the attacker must authenticate first.
Attackers holding any valid account on roncoo-education through 25.0.0-RELEASE can close other users' pending orders or read their course IDs and prices by supplying a victim's order number to the platform's payment 'cancel' and 'continue' endpoints, which never verify that the order belongs to the requesting user. Impact is limited to low confidentiality and integrity loss - pending order state changes and disclosure of course identifiers and prices - with no exposure of payment funds, credentials, or server control, since exploitation requires authentication and knowledge or enumeration of valid order numbers. No public exploit code has been confirmed at the time of analysis and there is no CISA KEV entry, although a public technical writeup of the flaw is referenced; no vendor patch was identified in the supplied data.
Remote, unauthenticated clients can crash libIEC61850-based MMS servers running versions prior to 1.6.2 by writing an RptID value longer than 129 bytes to a report control block and then triggering report serialization via General Interrogation (GI). The oversized identifier overflows a fixed 130-byte stack buffer that is later wrapped as a temporary MmsValue and handed to MmsValue_setVisibleString, whose resize path attempts to free non-heap (stack) memory - an invalid free that terminates the server process. Impact is availability-only (CVSS 5.3, A:L) with no code execution, data disclosure, or integrity effect, and no public exploit has been identified at time of analysis; the issue is fixed in upstream version 1.6.2.
Argument injection in Seanime v3.7.0 through v3.10.3 lets unauthenticated clients on the same LAN or a shared Tailnet smuggle attacker-controlled flags into the ffmpeg command line, achieving arbitrary file write and file truncation as the user running seanime. The flaw is reachable because the application's privileged-settings guard did not treat a switch of transcodeHwAccel to 'custom' (with attacker-supplied decodeFlags/encodeFlags in transcodeHwAccelCustomSettings) as a protected change, so a crafted mediastream or HLS playlist request is executed with the injected options appended verbatim to ffmpeg's argument vector. This is rated Medium (CVSS 4.0 5.3, AV:A/AC:L/PR:N/UI:N/VC:L/VI:L/VA:N); exploitation is bounded to adjacent-network reachability and requires a trusted LAN/localhost origin, no public exploit code has been identified at time of analysis, and no confirmed active exploitation (CISA KEV) has been reported.
Missing authorization in the WP Chill Passster WordPress plugin (content-protector) through version 4.3.18 allows unauthenticated remote attackers to bypass the plugin's gated-content access controls and read content that the site owner intended to restrict. According to the CVSS vector (AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:N/A:N), exploitation requires no authentication, no user interaction, and no non-default configuration, but the impact is confined to low-level confidentiality loss - no data modification, code execution, or site takeover. There is no public exploit identified at time of analysis, and the flaw only matters on sites that actively use Passster's protection feature to gate content; sites that do not gate content with the plugin are unaffected.
Unauthenticated attackers can exploit a missing authorization flaw in YITH PayPal Express Checkout for WooCommerce (versions up to and including 1.61.0) to gain unauthorized access to limited sensitive information. Tracked as CVE-2026-111409 and classified as CWE-862, the vulnerability has a CVSS 3.1 base score of 5.3 due to a low confidentiality impact only (C:L/I:N/A:N). No public exploit code or confirmed active exploitation has been identified at time of analysis, but the plugin's default configuration exposes an unauthenticated endpoint.
Missing authorization in the BookIt WordPress plugin (Liquid Web / StellarWP), in all versions through 2.6.0.5, allows remote unauthenticated attackers to reach a protected plugin action that should be restricted to privileged users, resulting in unauthorized modification of site data. The CVSS 3.1 vector (CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:L/A:N, score 5.3) confirms exploitation is network-based, unauthenticated, low complexity and requires no user interaction, but impact is confined to integrity - there is no confidentiality or availability loss and no scope change, so this is a moderate-priority broken access control issue rather than a full site takeover. No public exploit code and no confirmed active exploitation (CISA KEV) were identified at time of analysis, and the exact endpoint or feature lacking the capability check is not specified in the available description and should be confirmed against the Patchstack advisory.
Missing authorization in the Arraytics WP Event Solution WordPress plugin (versions through 4.1.26) allows unauthenticated remote attackers to bypass access controls and modify plugin-controlled data, as reflected by the CVSS 3.1 vector AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:L/A:N (low integrity impact only, no confidentiality or availability loss). This moderate-severity CWE-862 flaw requires only that the vulnerable plugin be installed and network-reachable on a WordPress site; no authentication or user interaction is needed. At time of analysis, no public exploit identified and the issue is not listed in CISA KEV.
Information disclosure in the MapSVG WordPress plugin (mapsvg-lite-interactive-vector-maps) through version 8.16.1 lets unauthenticated remote attackers retrieve sensitive data that a site has embedded in its MapSVG maps and content, because a missing authorization check (CWE-862) fails to properly gate access to that stored data. Per the assessed vector CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:N/A:N (5.3), no credentials or user interaction are needed and only network reachability to the affected site is required, while impact is confined to confidentiality with no integrity or availability effect and no public exploit identified at time of analysis. Practical risk therefore depends on what operators have placed inside MapSVG content: sites storing only public map data face negligible real impact, whereas those that embedded secrets or private records in map fields should treat this as a live exposure.
Cross-tenant server ID enumeration in the Nezha monitoring dashboard (versions 1.7.0 through 2.3.14) lets an authenticated low-privilege member map other tenants' server IDs by sweeping sequential identifiers against the getServerConfig and setServerConfig endpoints and distinguishing 'permission denied' responses from success payloads. According to the authoritative assessment, exploitation requires a valid member account plus a shared, multi-tenant deployment where more than one user's servers coexist; on a single-tenant self-hosted install there are no other tenants to enumerate, sharply reducing real impact. It yields server-ID mapping rather than configuration contents, secrets, or control, and stays confined to a low-severity information-disclosure outcome (C:L, I:N, A:N). No public exploit code was identified at time of analysis, and this CVE is not in CISA KEV; vendor patch version 2.3.15 is available.
Authorization bypass in eladmin through 2.7 allows an authenticated user who holds the delegated user:edit or user:del permission to modify, disable, reassign, or delete accounts that lie outside their assigned department data scope, because the PUT and DELETE handlers on /api/users omit the department check that the corresponding read endpoints enforce. The attacker must already possess a department-scoped user-manager role and be able to identify target user ids in other departments, so this is not an unauthenticated flaw (CVSS 4.0 vector rates privileges required as High; independent assessment scores CVSS 3.1 5.1 with PR:L). Impact is limited to account-state integrity and availability of other departments' accounts, with no confidentiality exposure and no code execution; there is no public exploit identified at time of analysis.
BuildAdmin through 2.3.9 contains an insecure direct object reference in the backend admin-profile edit endpoint (routine/AdminInfo/edit), allowing an authenticated administrator who holds that permission to overwrite another administrator's profile by manipulating the id parameter. The writable fields are avatar, nickname, email, mobile and motto; because altered email and mobile values commonly drive administrative account recovery and notification flows, the meaningful escalation path is takeover of a super administrator account rather than mass compromise. No public exploit code and no CISA KEV listing were identified at time of analysis, so this is a low-to-moderate priority issue whose exploitation requires an already-privileged backend session with the specific permission granted.
Site-scoped authorization bypass in SiteServer SSCMS through 7.4.0 allows an authenticated administrator who holds Channels permission on one site to read and overwrite another site's channel records by supplying a foreign channel ID to the ChannelsController Get and Update APIs. This CWE-639 IDOR breaks multi-site isolation for channel metadata such as names, file paths, templates, link URLs, and custom fields, but requires privileged tenant-admin access and a multi-site deployment; single-site installations and users without Channels permission are effectively unaffected. The provided CVSS 5.1 (CVSS 4.0) and the independent assessed CVSS:3.1/AV:N/AC:L/PR:H/UI:N/S:U/C:L/I:L/A:N both indicate a moderate, privilege-gated issue, and no public exploit identified at time of analysis.
An insecure direct object reference (IDOR) in the ThinkAdmin 'think-plugs-admin' plugin, versions through 1.0.81, allows an authenticated administrator who holds the admin/file.edit permission node - but who is not a super administrator - to read and overwrite other administrators' file records. By requesting or POSTing a target record id together with columns such as xurl, hash, or uuid, the attacker can disclose download URLs, tamper with stored file metadata, or reassign ownership of records belonging to other admins. Exploitation requires a valid privileged admin session and a deployment with multiple administrators, and there is no public exploit code identified at time of analysis and no CISA KEV listing, so real-world risk is a bounded, medium-severity authorization flaw (CWE-639) rather than a systemic compromise path.
PKP Open Journal Systems through 3.5.0-5 contains an insecure direct object reference (CWE-639) in its subscription management forms, letting an authenticated journal manager or subscription manager read subscriber details belonging to other journals via the editSubscription endpoint and reassign those subscriptions to their own journal via updateSubscription, which revokes the legitimate owners' access. Because subscription identifiers are sequential, the flaw is trivially enumerable, but exploitation requires a privileged account on at least one journal and a multi-journal installation where other journals' subscription records actually exist; single-journal deployments face negligible exposure and there is no public exploit identified at time of analysis.
A stored/DOM cross-site scripting flaw in tarteaucitron.js (CWE-79) affects all releases before 1.37.0 and allows an attacker who can place markup on a page - for example a CMS author or contributor - to inject a 'data-url' or 'data-src' attribute holding a javascript: URL next to a benign-looking 'url' attribute. Because getElemAttr resolves the 'data-' prefixed attribute into an iframe src while the https-only regex guard validates the unprefixed attribute, the javascript: scheme is never rejected and the injected script executes in the site's origin. The vendor rates this Medium (CVSS 5.1, 4.0 vector), our independent assessment scores it CVSS:3.1/AV:N/AC:L/PR:L/UI:R/S:C/C:L/I:L/A:N, and there is no public exploit identified at time of analysis; the authoring-privilege (PR:L) and victim-interaction (UI:P/UI:R) prerequisites limit realistic abuse to multi-author or user-generated-content environments.
Improper verification of cryptographic signatures in IBM Db2 Genius Hub 1.0 allows local attackers to bypass JWT signature validation, leading to partial information disclosure and data integrity loss. The flaw requires local access (AV:L) with no authentication or user interaction, but its impact is limited to low confidentiality and integrity (C:L/I:L), making it a moderate-severity, lower-priority issue. No public exploit has been identified, and IBM has released a patch.
Mattermost versions 11.10.x through 11.10.0, 11.7.x through 11.7.9, 11.9.x through 11.9.1, and 11.8.x through 11.8.5 fail to validate the invited recipient's email address when redeeming team or guest invitation tokens, enabling an authenticated user who obtains a forwarded or leaked invite link to join a team or access private channels without authorization. The flaw is rated CVSS 5.0 (moderate severity) and requires an attacker with an existing authenticated account plus a valid, unredeemed invitation token originally issued to a different email address; no victim user interaction is needed. No public exploit code or confirmed active exploitation has been identified at time of analysis, and the vendor advisory is MMSA-2026-00747.
Local privilege escalation in the EA Javelin Anticheat standalone installer (builds 1.0.14329371, 1.0.14547437, 1.0.14789537 and 1.0.14940698) on Windows lets an authenticated, non-administrative user rewrite the EAAntiCheatService service configuration and executable path so that the service subsequently runs attacker-controlled code as NT AUTHORITY\SYSTEM. Exploitation is narrow: it only applies to hosts where the standalone installer was executed manually, standard EA App and Steam game installations are explicitly unaffected, and the weak service permissions self-heal as soon as any integrated EA game title is launched. EA rates the issue CVSS 5.0 (AV:L/AC:H/PR:L/UI:R), while independent assessment scores it higher (AV:L/AC:H/PR:L/UI:N with high confidentiality, integrity and availability impact) because a silent service-path hijack yields full host compromise; no public exploit was identified at time of analysis, although a third-party technical writeup on the flaw is referenced, and the vendor-released fix is version 1.0.14940699.
Authenticated SQL injection in the FluentCart WordPress eCommerce plugin, affecting all versions up to and including 1.7.0, lets users who hold shop-management privileges slip attacker-chosen keys into an unprepared bulk-product-update query and read arbitrary data out of the site's database. The vulnerable path is the product bulk-update flow, and on FluentCart Pro installs the non-administrator 'Manager' shop role carries the products/edit permission and can therefore trigger it without administrator access - which is why the effective privilege requirement is moderate rather than high. Reported by Wordfence, this is a genuine but constrained issue rather than an urgent fire-drill: no public exploit code and no confirmed active exploitation were identified at time of analysis, and impact is confined to confidentiality (database read/extraction) with no integrity or availability effect.
Authenticated SQL injection in the WP Booking System - Booking Calendar plugin for WordPress (all versions up to and including 2.1.0.1) lets a user with Administrator-level privileges or above inject arbitrary SQL via the AJAX 'data[orderby]' parameter and read sensitive records from the site database. The attack path is constrained: the orderby value only reaches the SQL query builder when the booking AJAX handler takes the use_ajax branch, which requires the targeted calendar to contain more than 200 bookings - a non-default state that an Administrator can deliberately create. This issue is not listed in CISA KEV and no public exploit code was identified at time of analysis; a confidentiality-only impact (CVSS 4.9, PR:H, no user interaction) keeps the practical risk low-to-moderate, and the vendor has released 2.1.0.2 to fix it.
Authenticated SQL injection in PeerBanHelper versions 5.0.0 through releases before 9.3.15 enables a high-privileged web API user to read arbitrary data from the configured SQLite, MySQL, or PostgreSQL database by supplying an unrecognized field value to the /api/statistic/analysis/field endpoint. Exploitation requires valid admin-panel credentials (CVSS PR:H) and reachable API access, with no user interaction needed. CVE-2026-62362 is rated medium severity (CVSS 4.9), is fixed in 9.3.15, and has no public exploit identified at time of analysis.
Debug-level logging in Mattermost Server can write authentication secrets - session tokens and OAuth access tokens - into server log entries without sanitization, allowing anyone who can read those logs to impersonate authenticated users. Affected releases are 11.7.x through 11.7.10, 11.8.x through 11.8.5, 11.9.x through 11.9.1, and 11.10.x through 11.10.1 (Mattermost advisory MMSA-2026-00763); the vendor scores it 4.5 (MEDIUM) under CVSS:3.1/AV:N/AC:L/PR:H/UI:R/S:U/C:H/I:N/A:N. Exploitation is prerequisite-heavy rather than opportunistic: the server must be running with debug-level logging (not the default production setting) and the attacker must already hold read access to those logs via filesystem access, a centralized log/SIEM viewer role, or a compromised log pipeline, and the assessed vector (AV:L/AC:H/PR:H/UI:N) reflects that local, high-privilege access barrier. There is no public exploit identified at time of analysis and no KEV listing; impact is confidentiality-only, as stolen tokens enable user impersonation while the flaw itself produces no direct integrity or availability loss.
Mattermost 11.7.x through 11.10.x fails to invalidate the stored AD/LDAP bind password whenever other connection settings - such as the directory server address - are changed in the System Console, so the previously saved plaintext credential is carried over to the new configuration. An authenticated System Console administrator (CVSS:3.1/AV:N/AC:L/PR:H/UI:N/S:U/C:H/I:N/A:N, base 4.9) can then trigger the LDAP Test Connection / diagnostics feature, which transmits that retained plaintext bind password to the configured server address, potentially an attacker-controlled host set by the same admin. There is confirmed no CISA KEV entry and no public exploit identified at time of analysis, and no EPSS score was supplied; exploitation is gated behind high-privilege System Console access, an existing AD/LDAP integration with a stored bind password, and the deliberate admin action of pointing the connection at a hostile host, making this a genuine but moderate-priority confidentiality issue rather than an emergency.
Account takeover in Mattermost 11.7.x through 11.10.0 lets an authenticated attacker who already holds a valid OAuth bearer token for a specific target user permanently hijack that user's SSO-backed account: the POST /api/v4/users/login/switch endpoint fails to reject OAuth application sessions, so the attacker can switch the account to password authentication, set an attacker-controlled password, and detach the account from its identity provider (CWE-863 Incorrect Authorization; CVSS 4.4, CVSS:3.1/AV:N/AC:H/PR:H/UI:N/S:U/C:N/I:H/A:N). Exploitation is heavily constrained - the victim must have an identity-provider-backed account, the attacker must obtain a victim-scoped OAuth token (typically by getting the user to authorize a malicious OAuth application), attack complexity is assessed as high, and deployments without SSO or OAuth applications are not exposed. No public exploit identified at time of analysis, and there is no confirmed active exploitation (Mattermost advisory MMSA-2026-00758).
Unauthenticated privilege escalation in the WP Compress - Instant Performance & Speed Optimization plugin for WordPress (all versions up to and including 7.25.00) allows remote attackers with no WordPress account to gain plugin-administrator-equivalent control over the plugin's optimization, CDN and critical-CSS settings. The chain is two-stage: the stored plugin API key is first leaked by the unauthenticated `wpc_send_critical_remote` nopriv AJAX action, but only when the external critical-CSS service returns an `apiResults` field that causes `initCritical()` to echo the key via `wp_send_json_success`; the harvested key is then replayed as the `apikey` parameter to the `start_comms` dispatcher, which performs no capability check and no nonce verification before passing attacker-controlled `group`, `setting` and `value` parameters into `change_setting()`, writing them into the `WPS_IC_SETTINGS` option without an allow-list and causing `checkAndAddCaps()` to grant the `manage_wpc_settings` capability to arbitrary roles, including Subscriber. The assessed severity is moderate - CVSS 4.8 (AV:N/AC:H/PR:N/UI:N/S:U/C:L/I:L/A:N) - unauthenticated per the vector (PR:N) but rated high complexity because the API-key disclosure depends on the external critical-CSS service response; impact is scoped to the plugin's settings surfaces rather than full WordPress site administration, and no public exploit code and no confirmed active exploitation (not in CISA KEV) were identified at time of analysis.
Arbitrary shortcode execution in the Popup Box plugin for WordPress (ays-pro) versions 6.4.1 through 6.4.2 lets unauthenticated remote attackers invoke any shortcode registered on the site with attacker-chosen attributes, because client-controlled IP-related HTTP headers are folded into a popup body and then run through do_shortcode(). The attack is only reachable when a site administrator has manually inserted the %%user_ip_address%% token into a popup's Custom Content body, so default installations and sites that never use that token are not affected; impact is bounded by the shortcodes already installed (no direct PHP injection), which is why the assessed vector scores only 4.8 with high attack complexity. No public exploit code and no confirmed active exploitation were identified at time of analysis, and the vendor reference points to a fix in 6.4.3.
A stack out-of-bounds read in Antiword 0.37 and earlier (prop8.c, vGet8PapInfo and vGet8ChrInfo) allows a crafted .doc file containing malformed FKP pages to drive run counts and grpprl offsets past the end of a 512-byte stack buffer, crashing the parser or corrupting style parsing. Exploitation is local and gated by user interaction: a victim must be given the file and invoke antiword on it (AV:L/UI:P), and the assessed impact is availability-only with no confidentiality, integrity, or code-execution effect. The attack is unauthenticated per the CVSS 4.0 vector (PR:N), no public exploit is identified at time of analysis, and the CVSS 4.0 base score of 4.8 (MEDIUM) likely overstates its operational urgency given the limited impact and delivery prerequisites.
Code injection in Ultralytics FastSAM's AutoBackend model-loading path lets a crafted model file execute arbitrary Python when its metadata entries (imgsz, names, kpt_shape) are passed into a dynamically evaluated expression from session.get_modelmeta().custom_metadata_map in ultralytics/nn/autobackend.py. The attack is bounded to the local model-loading surface - there is no network-facing exposure - and requires that a user or process loads an attacker-supplied model artifact, so the practical prerequisite is delivering a malicious model to a victim rather than obtaining privileges on the host. VulDB rates it 4.8 (medium) under CVSS 4.0 with low confidentiality/integrity/availability impact and PR:L/UI:N, while our independent assessment scores it CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H, i.e. higher impact and requiring user interaction instead of prior authentication; no CISA KEV listing or public exploit code is present in the supplied data, and the maintainers were notified via a GitHub issue but have not responded, so no vendor-released patch is identified at time of analysis.
Safing Portmaster before 2.2.5 exposes a local update endpoint, POST /api/v1/updates/from-url, that is registered with api.PermitAnyone and therefore performs no authorization check, allowing any unprivileged local process to point the updater at an attacker-controlled index URL and bypass both version checks and auto-apply settings. The result is a forced downgrade that can only install releases legitimately signed by Safing, so the practical impact is the reintroduction of previously patched vulnerabilities rather than execution of arbitrary attacker-supplied code, and exploitation requires low-privileged local access (CVSS PR:L) with no user interaction and cannot be performed remotely. Fixed in 2.2.5 by removing the endpoint entirely; no public exploit identified at time of analysis.
In ikus060/ikus-soft rdiffweb 7.1.1, the repository containment check relies on a string prefix comparison (startswith) instead of a true filesystem path-boundary check, so a percent-encoded traversal sequence (%2e%2e) in the request path lets the browse, restore, history, and delete operations resolve to a path outside the intended repository. The attack surface is narrow: the actor must already be an authenticated user with permissions on the target, the target must be a sibling directory whose name begins with the operated repository's path string, and the victim must be socially induced to issue the crafted request (CVSS 3.1 rating 4.6; our independent assessment is more conservative at 3.1 given AC:H). The realistic outcome is a user operating on - or deleting - data other than the repository being displayed; no confidentiality impact is involved because the user only reaches data they already have rights to. No public exploit code was identified at time of analysis, and there is no KEV entry.
Denial of service in IBM i's Digital Certificate Manager (DCM) for System TLS allows a local, high-privileged attacker to prevent secure network connections from starting or cause fallback to non-secure connections. Affected versions are IBM i 7.3, 7.4, 7.5, and 7.6. No public exploit code or active exploitation has been identified at time of analysis, and the requirement for authenticated, high-privilege local access makes this a moderate-severity, low-urgency issue.
An out-of-bounds memory read in MariaDB Server's spatial (GIS) functions lets an authenticated but low-privileged database user pass a truncated WKB MultiPolygon or MultiLineString to ST_Area, ST_Centroid, or GLength and have the parser read past the end of the allocated buffer, returning adjacent heap bytes as a geometry result or error path. Affected releases are MariaDB 10.6.1 through the fixes in 10.6.28, 10.11.19, 11.4.13, 11.8.9, 12.3.3, and 13.0.2; exploitation requires valid credentials and the ability to submit attacker-controlled spatial data (CVSS 4.3, PR:L), and the leak is primarily useful as an ASLR-bypass primitive inside a larger exploit chain rather than as a standalone compromise. No public exploit code has been identified at time of analysis and the flaw is not listed in CISA KEV, so it should be treated as low-to-moderate priority and scheduled with routine patching; the vendor advisory GHSA-p7v9-24h4-mch8 and MDEV-40328 document the issue.