Cross-tenant account takeover in ROOST Coop before 1.0.2 stems from the SAML verify callback resolving the authenticated user by email only, without binding that lookup to the organization named in the /saml/login/:orgId/callback path. An attacker who administers any organization on the instance and controls its SAML IdP signing key can post an assertion for a victim's email to a different organization's callback when the target organization has registered the same signing certificate, and receive a session as that victim. The attack is authenticated (PR:H in the vendor vector, since registering a SAML IdP requires organization administrator rights) and is only exploitable in the unusual shared-certificate configuration; there is no KEV listing and no public exploit identified at time of analysis, and the decision signal places remediation in the this-week window.
Stored cross-site scripting in Mythic 3.2.16 through 3.4.0.61 allows an attacker who controls files on an agent-reachable endpoint to plant HTML inside a .pdf-named file; when a Mythic operator downloads the file and clicks Preview Media, /direct/view content-sniffs the content and renders it in the Mythic UI origin. The injected JavaScript can read localStorage JWTs and use them for full GraphQL and REST access; CVSS 4.0 score is 7.4 (High), and the CVSS vectors list PR:N, though exploitation requires prior control of endpoint files and operator interaction. No CISA KEV status or EPSS score was provided; a public PoC reference exists but is not verified as working exploit code.
Unauthenticated privilege escalation in the Essential Addons for Elementor plugin (all versions up to and including 6.8.7) lets any site visitor register an account with a role as high as Editor - gaining edit_others_posts, publish_pages, upload_files and unfiltered_html - by passing a WordPress revision ID as the page_id parameter of the register_user() handler. The escalation is not a default-configuration drive-by: the site must have "Anyone can register" enabled, a page must carry the plugin's Login/Register widget, and an administrator must at some point have saved and then reverted a non-default "New User Role" on that page, leaving a revision whose _elementor_data still holds the elevated role. Those attacker-uncontrolled, non-default prerequisites are why the vector scores AC:H despite PR:N; no public exploit has been identified at time of analysis and the issue is not in CISA KEV. In practice this is a genuine but conditionally-gated issue rather than a drop-everything emergency, though the impact when the preconditions are met is full administrative-adjacent compromise of site content.
Unsafe deserialization in Metalama's Visual Studio design-time named-pipe channel lets a process already running as the same logged-in Windows user send a crafted message that instantiates arbitrary .NET types and achieves code execution inside devenv, the out-of-process Roslyn code-analysis process, or the CodeLens process. Affected are Metalama.Framework NuGet releases from 2023.0.101-rc onward (fixed in 2025.1.18, 2026.0.25 and 2026.1.19), the maintained 'Metalama + PostSharp' Visual Studio extension 2024.1 through 2026.1.4 (fixed in 2026.1.5), and all versions of the retired Metalama and Caravela extensions, which will not be patched. Because default Windows pipe ACLs only allow same-user writers, this is not a privilege escalation and command-line/MSBuild-only builds are unaffected; no public exploit code is identified at time of analysis and there is no CISA KEV entry.
A stack-based buffer overflow in Thales Sentinel Keys Server on Windows (versions lower than 1.3.8.0) lets a local attacker who can send a message to the service overflow a stack buffer by supplying an over-long first field, ultimately allowing code execution as NT AUTHORITY\SYSTEM. The flaw is local-only (CVSS:4.0 AV:L, PR:L, base 7.3 HIGH) and requires an additional, unspecified attack condition (AT:P), with no user interaction needed. It is a privilege-escalation primitive for a low-privileged local user rather than a remote entry point; no public exploit has been identified at time of analysis and no confirmed active exploitation is reported.
Symbolic-link, hard-link and FIFO race conditions in pgAdmin 4 server mode (versions before 9.19) allow an attacker who can write to a user's storage backend - for example a shared or network-mounted storage directory - to redirect servers.json export/import, the import file listing, and the files handed to pg_dump, pg_dumpall, pg_restore and psql during Backup, Restore and Import/Export Data outside the storage directory, so a file can be read, written or overwritten, or loaded into a table, as the operating-system account running pgAdmin. A FIFO planted at the path can instead hang the request, affecting availability. Successful exploitation requires authenticated pgAdmin access (PR:L) plus a non-default shared-storage layout and winning the time-of-check/time-of-use race, so this is a moderate, conditional risk rather than an urgent widespread threat; no public exploit code or confirmed active exploitation has been identified at time of analysis.
Information disclosure in Dell OpenManage Enterprise versions prior to 4.8.0 stems from a cryptographic key embedded in the application and shared across installations, allowing data protected by that key to be decrypted by anyone who can read it. Exploitation requires local access to the OpenManage Enterprise host plus a low-privileged authenticated account (CVSS AV:L/PR:L), so it is not remotely reachable; the impact is rated high for confidentiality and low for integrity, with a scope change. No public exploit code and no confirmed active exploitation were identified at time of analysis, and the vendor advisory (DSA-2026-467) is the reference for the fixed 4.8.0 release.
Stored cross-site scripting in the Food Store WordPress plugin through 1.5.5 allows unauthenticated attackers to inject script through the 'service_time' parameter when the 'Enable Checkout Service Fields' option is turned off and 'Enable Pickup/Delivery Service' remains enabled. The payload is stored without sufficient sanitization and later rendered without output escaping, so it executes in the browser of a user who accesses the injected page. The assessed severity is CVSS 7.2 with a changed-scope impact, but exploitation is limited by the required non-default configuration; no CISA KEV entry or public exploit code is identified in the supplied intelligence.
Stored cross-site scripting in the Gravity Forms WordPress plugin through 3.1.2 allows unauthenticated form submitters to inject script through List Field values; the payload can execute in an administrator's browser only after the admin opens the affected entry, clicks the injected link, and confirms a dialog. The interaction gate sharply limits practical exploitability, and no public exploit is identified at time of analysis. The reported severity is CVSS 7.2, with the assessed vector reflecting the required administrator interaction.
Stored cross-site scripting in the Gravity Forms plugin for WordPress, affecting all versions up to and including 3.1.1, lets unauthenticated attackers persist arbitrary script through Post Excerpt, Post Custom Field and Rich-Text-Editor-enabled Paragraph Text fields, which the plugin accepts as raw HTML via allow_html(). The payload is inert at save time - an entity-encoded value placed in a data-dialog-confirm attribute survives wp_kses_post() - and only executes when a user or administrator later opens the injected page or entry and the plugin's admin JavaScript passes that attribute value to insertAdjacentHTML(). Exposure is limited to sites that expose a publicly accessible form configured with one of those field types and where a victim subsequently views the stored entry; no public exploit was identified at time of analysis, and the current priority label is this_week.
Path traversal in pgAdmin 4 releases before 9.19 lets an authenticated user in server (multi-user web) mode point a server's password file, client certificate, key, CRL, SSH tunnel identity file, or the sslkeylogfile parameter outside their own storage directory, so a connection can consume another user's saved PostgreSQL passwords, client certificate or SSH key, or make libpq create/append a file as the pgAdmin service account. The flaw exists because get_complete_file_path() only stripped leading slashes before joining the configured path to the user's storage directory, and because make_conninfo() quoted connection parameter values but not their names, allowing a crafted parameter name to smuggle extra keyword=value pairs past that check. Only server-mode deployments are affected; exploitation requires an authenticated pgAdmin account and a target file that actually exists at a guessable path (or a writable directory for the sslkeylogfile vector), and no public exploit or active exploitation has been identified at time of analysis. A vendor fix is available in pgAdmin 4 9.19.
Authenticated users of AdminNeo's SQLite driver can escape the intended query blocklist and turn the SQL command console into an arbitrary file write, dropping a PHP web shell into a web-served directory and executing commands as the PHP process user. The flaw affects the Composer package adminneo-org/adminneo before version 5.5.1: the code only filters statements beginning with ATTACH, so a crafted VACUUM INTO statement (SQLite 3.27.0 or later) writes the whole database to any path with any extension, embedding a table name containing PHP code as raw bytes. Exploitation requires valid database credentials to reach the SQL console (CVSS PR:H), after which the only further conditions are a writable, PHP-executing web directory and a SQLite build that supports VACUUM INTO; no public exploit code was identified at time of analysis, and the issue is not in CISA KEV. The vendor addressed it in 5.5.1.
Authenticated SQL injection in the OpenRemote datapoint export API (versions prior to 1.26.0) lets a user who is permitted to create or rename an asset and export its datapoints store SQL metacharacters in the asset name and receive injected SELECT output inside the normal ZIP/CSV crosstab response, giving a practical database read primitive with no direct database, filesystem or token-forgery access. Because the injected query executes with the manager's PostgreSQL role, a multi-tenant deployment can expose other tenants' rows when that role can read shared manager tables, while integrity and availability effects were not established by the demonstrated read primitive. The vendor advisory describes a proof of concept that returned current_user, current_database() and an application-table row count, but no active exploitation is confirmed (not in CISA KEV); with the assessed vector CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:L/A:L this is a genuine, actionable vulnerability but not a drop-everything emergency.
PHP object injection in the WordPress plugin Product Feed Manager (best-woocommerce-feed) 7.14.1 and earlier allows an authenticated WooCommerce shop manager to pass serialized data into plugin code that unserializes it, scored CVSS 7.2 (CWE-502) with high confidentiality, integrity and availability impact. Because the vector is PR:H, only a high-privilege shop manager account can reach the vulnerable path, and the realistic worst case depends on whether a usable POP gadget chain is present in the plugin, WordPress core or another installed component; object injection alone may not yield code execution. No public exploit code is identified in the supplied sources and the issue is not listed as confirmed actively exploited (CISA KEV), so this is a conditional, privilege-gated risk prioritised for remediation this week rather than an internet-wide emergency.
Stored cross-site scripting in the Independent Analytics Pro WordPress plugin (all versions up to and including 2.15.5) lets unauthenticated attackers inject arbitrary web scripts through the 'utm_campaign' parameter, which execute when a victim later views the compromised page. The paid Pro build must be active - the free build nulls campaign parameters and is unaffected - and exploitation requires first fetching a frontend page to obtain a server-signed payload, then replaying it against the unauthenticated /iawp/search REST endpoint, because the signature check does not cover utm_* fields. Rated CVSS 7.2 (AV:N/AC:L/PR:N/UI:N/S:C/C:L/I:L/A:N); no confirmed active exploitation and no verified working exploit code from the available sources.
Stored cross-site scripting in the easy.jobs WordPress plugin (all versions up to and including 2.8.4) lets unauthenticated attackers plant arbitrary web scripts through the public job-application 'cover_letter' field, which the plugin later renders without adequate escaping inside its React-based admin interface. The injection point is the plugin's default, publicly reachable application flow and needs no credentials or non-default configuration, but execution is gated: a logged-in administrator must open the affected candidate's Application tab and click the attacker-crafted link, at which point WHATWG URL parsing collapses an encoded scheme back to javascript: and the payload fires in the admin's session. No public exploit code has been identified at time of analysis; the requirement for privileged, active administrator interaction is the main constraint on real-world exploitation, and the issue sits in this week's remediation queue rather than an emergency one.
Stored cross-site scripting in the WP ERP (ERP: Complete HR, Accounting & CRM Suite Built for WooCommerce) WordPress plugin affects all versions up to and including 1.17.10, where the inbound email router accepts a References header on a regex match alone without ever cryptographically validating the reply token's hash. That allows any external sender - no WordPress account required - to inject arbitrary script into the CRM timeline via an inbound email body, which then executes when an authenticated CRM agent or administrator opens the injected record. Exploitation depends on the plugin's inbound email routing and CRM email sync being enabled and processing mail, and the payload is passive; no public exploit code was identified at time of analysis and there is no CISA KEV confirmation of active exploitation.
OS command injection in the Tenda W18E web management server lets an attacker who can reach the 'diagnose' packet-capture feature inject shell metacharacters through the JSON 'protocol' or 'targetDevice' fields, which are concatenated into a tcpdump command executed with root privileges and therefore yield full compromise of confidentiality, integrity and availability. In a normally configured device the injection requires an authenticated session (CVSS PR:H), but the description states the same path is reachable without authentication while the unit is in a factory-reset, unconfigured state; the confirmed affected build is W18E V16.01.0.20. Exploitation is not confirmed active (CISA KEV status is absent and SSVC exploitation is 'none' with automatable 'no'), although a public exploit-development writeup is referenced for this target; that reference is not verified as working exploit code, and no EPSS score was supplied.
Broken object-level access control in the Kirki plugin for WordPress before 6.4.1 lets any user who holds Kirki edit access - granted to the Editor role by default - rewrite the title, slug and post_status of arbitrary non-trashed posts, including custom post types and WooCommerce products they do not own, exposing private or draft content or silently unpublishing live content. The PUT /wp-json/kirki/v1/pages/{page_id} route and the save_page_data, duplicate and delete handlers in PageController.php apply only a coarse role-to-access middleware check and never perform a per-object WordPress capability check. Exploitation requires an authenticated account with Kirki edit access (CVSS vector PR:L); no public exploit code and no confirmed active exploitation (CISA KEV) were identified at time of analysis, and the vendor has released 6.4.1, so remediation is prioritized for this week rather than treated as an emergency.
Improper neutralization of argument delimiters (CWE-88 argument injection) in Dell OpenManage Enterprise consoles running any version prior to 4.8.0 allows a remote attacker who already holds a high-privileged account to inject command arguments and escalate privileges on the appliance. The CVSS 3.1 score is 7.2 (AV:N/AC:L/PR:H/UI:N/S:U/C:H/I:H/A:H), reflecting high confidentiality, integrity and availability impact but requiring a privileged foothold, which is the primary limiting factor and confines realistic exploitation to privileged insiders or attackers who have already compromised such an account. No public exploit code has been identified at time of analysis, and the issue is tracked for remediation this week rather than as an emergency; Dell has published a security update for the product, with 4.8.0 as the fixed release.
OS command injection in Dell OpenManage Enterprise versions prior to 4.8.0 allows a remote, high-privileged authenticated attacker to execute arbitrary operating-system commands, with high confidentiality, integrity, and availability impact. Exploitation requires an administrative account reachable over the network; no public exploit was identified at time of analysis. Dell has released 4.8.0 as the fixed version, and the high-privilege prerequisite keeps this a conditional priority rather than an internet-scale emergency.
OS command injection (CWE-78) in Dell OpenManage Enterprise affects all releases prior to 4.8.0 and lets a remote attacker who already holds high-privileged, administrative access to the management console execute arbitrary operating-system commands, with high impact on confidentiality, integrity and availability (CVSS 3.1 base 7.2, AV:N/AC:L/PR:H/UI:N/S:U/C:H/I:H/A:H). The hard prerequisite of existing admin-level access or valid high-privilege credentials sharply narrows realistic exploitation to insiders, credential-theft scenarios, or chaining from a separate authentication bypass rather than an unauthenticated remote attack. No public exploit was identified at time of analysis, and Dell has released a fixed version (4.8.0); the vendor-reported issue carries a this_week remediation priority.
A format-string flaw in netopeer2-server's logging path lets a low-privileged authenticated NETCONF user crash the server remotely by sending a `get` RPC containing an invalid subtree filter whose text includes format specifiers such as `%s%s%s%s`. The invalid value is formatted once with `vsnprintf()` and then passed unsafely as a format string to `vsyslog()`, which dereferences attacker-influenced percent sequences and segfaults the daemon; the reporter confirmed this on version 2.8.7 over NETCONF-over-SSH with AddressSanitizer, and states reliable remote code execution was not confirmed. Exploitation requires valid NETCONF credentials (CVSS PR:L, so authentication is mandatory) and network reach to port 830; the advisory identifies 2.8.8 as the patched release and repository comparisons verify the fix in 2.8.21, and a public PoC reference exists in the advisory but is not verified as working exploit code. Denial of service only - no confidentiality or integrity impact - makes this a bounded availability risk rather than a critical RCE emergency, and no active exploitation (CISA KEV) is indicated.
MyTube through 1.10.2, when deployed in multi-user mode with loginEnabled=true and a visitor password configured, lets an authenticated visitor retrieve hidden (visibility=0) video records and then download the associated media and subtitles because the backend video queries and static media routes do not enforce the visitor's intended public-only role restriction. The GitHub security advisory includes a public proof-of-concept request sequence; it is not independently verified as working exploit code, and no CISA KEV entry or confirmed active exploitation was reported. Deployments using the default loginEnabled=false single-user mode are unaffected.
Generation-time server-side request forgery, remote file inclusion and out-of-tree local file inclusion in the Refitter .NET tool (versions through 2.0.0) allow an attacker who controls or can influence an OpenAPI description to make the build host fetch arbitrary http(s) URLs or read files outside the input directory, with the retrieved content inlined into the generated C# Refit interfaces. Rated CVSS 3.1 7.1 (High) with vector AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:L/A:N: unauthenticated and low complexity on the network side, but exploitable only when a developer or CI job actually runs refitter against the malicious spec, and the chain terminates at disclosure and SSRF - no code execution, because Refit sets the base URL at runtime, enum values are verbatim-string doubled and string defaults are escaped. A public reproduction script is attached to the upstream advisory and was verified by the reporter against 2.0.0; no CISA KEV entry appears in the supplied data, and the vendor fix ships in 2.1.1.
Kerberos credential misuse in pgAdmin 4 server mode (before 9.19) lets any authenticated pgAdmin web user connect to a PostgreSQL server using the Kerberos principal of the operating-system account that runs pgAdmin, regardless of the user name they supply and even with gssencmode=disable set. The problem requires a multi-user server-mode deployment where the pgAdmin service account holds usable Kerberos credentials (a default credential cache, a client keytab, or on Windows an SSPI logon) and the target server requests GSSAPI or SSPI authentication, which is why severity is tempered by high attack complexity and a race-like prerequisite of credential availability; Windows SSPI and gssproxy paths are also affected. No public exploit code has been identified at time of analysis and there is no CISA KEV listing; the vendor fix is pgAdmin 4 9.19, and full enforcement of the refusal logic requires libpq 16 or later.
Path traversal in pgAdmin 4 server mode before 9.19 lets an authenticated user escape their own storage directory and read, write, delete or move files elsewhere on the host - in the POSIX Query Tool and Google client-secret paths, any file the pgAdmin service account can read; on Windows, Import/Export Data, Backup, Restore and file-save can read or write outside storage via a drive-letter path. The defect is a check-versus-use mismatch: check_access_permission() validated a path built by get_abs_path(), while several callers stripped leading separators, double-decoded folder names, or concatenated without a separator, so download/delete/rename can reach another user's files where one storage directory name is a prefix of another's, and folder creation can land in another user's directory. Rated 7.1 High (CVSS 4.0) with high confidentiality and integrity impact in the independent CVSS 3.1 assessment; exploitation needs the multi-user server deployment and a valid account (PR:L), desktop mode is explicitly unaffected, no public exploit code has been identified at time of analysis, and the fix ships in 9.19.
Credential disclosure in pgAdmin 4 before 9.19 allows a database name controlled by a CREATEDB role to be interpreted by libpq as a connection string, redirecting the psql refresh process to an attacker-chosen host and exposing the connecting user's saved server password. Exploitation is conditional: it requires an authenticated role with CREATEDB, a pgAdmin user with a stored password who navigates into the malicious database and triggers a materialized view refresh, and an affected version before 9.19. No public exploit was identified at time of analysis, and the same issue class was previously fixed for Backup, Restore, Maintenance and Import/Export Data under CVE-2026-86862.
Server-mode pgAdmin 4 installations before 9.19 can be driven by any authenticated low-privileged pgAdmin user to open an SSH tunnel that authenticates with the pgAdmin service account's own ambient credentials, letting that user reach whatever SSH hosts trust the service account. The tunnel code passed neither host_pkey_directories nor ssh_config_file to the sshtunnel library, so sshtunnel fell back to the unencrypted private keys in the service account's ~/.ssh, its ~/.ssh/config (IdentityFile, User, HostName, ProxyCommand) and its ssh-agent; an attacker could succeed with no identity file of their own, or with an identity file that did not exist in their storage. Rated 7.1 under CVSS 4.0 (CVSS 3.1 assessed as AV:N/AC:H/PR:L/UI:N/S:C/C:H/I:H/A:N), prioritised 'this_week'; no public exploit code and no confirmed active exploitation were identified at time of analysis, and impact is bounded by whether the service account's credentials are actually trusted by a reachable host.
Cross-site scripting in the WordPress VikRentCar plugin through version 1.4.10 allows remote attackers to inject content that is rendered in a victim’s browser, but exploitation requires the victim to open attacker-controlled input or a crafted link rather than occurring automatically. No authentication is required (PR:N per the CVSS vector), yet impact stays within the victim’s browser/WordPress session context; the CVSS 7.1 score is elevated mainly by scope change (S:C) and network access, while confidentiality, integrity, and availability impacts are Low, with the independent assessment setting availability to None. Patchstack reported the issue and the vuln.today decision label is this_week; no CISA KEV entry is present and no public exploit identified at time of analysis.
Cross-origin CSS rule disclosure and mutation in Ladybird Browser allows an unauthenticated remote attacker page to read the parsed rules of a stylesheet it loads from another origin without CORS, and to insert or delete rules in that live victim stylesheet, bypassing the Same-Origin Policy. All builds through commit 7e8da6e7cf24 are affected, the vendor has fixed the flaw in commit c0d68bddf3e3616e5173835f0f4c155ee9700d91, and the advisory documents a confirmed live two-origin proof-of-concept; CISA KEV does not list this issue. This is a genuine but bounded Same-Origin-Policy bypass (CVSS 7.1, UI:R) rather than a memory-safety or code-execution issue - the advisory explicitly rules out memory corruption, RCE and persistence, and the intelligence tags claiming RCE or buffer overflow are not supported by the source evidence.
pgAdmin 4 server mode before 9.19 can let any authenticated pgAdmin user inherit libpq credentials belonging to the operating-system account that runs pgAdmin, by setting a server's Service or a service/servicefile connection parameter to a service entry that holds a password, password file, or client certificate. The same exposure applies to PGPASSWORD/PGSERVICE inherited by libpq and to pg_dumpall interpreting a maintenance database name such as service=... as a connection string; desktop mode is not affected. No public exploit was identified at time of analysis, and the issue is fixed in 9.19.
pgAdmin 4 in server (multi-user web) mode before 9.19 lets any already-authenticated pgAdmin user reach credential material belonging to the OS service account that runs pgAdmin. When a server definition has no password file, client certificate or client key configured - or the configured path cannot be found - pgAdmin omits the parameter and libpq falls back to its defaults (~/.pgpass, ~/.postgresql/postgresql.crt and postgresql.key, or PGPASSFILE/PGSSLCERT/PGSSLKEY), so a connection can be made with the service account's saved passwords or client certificate; supplying only a copy of the public certificate is enough for libpq to pair it with the service account's private key. The flaw also resurfaces through utility paths, where a pg_service.conf entry or a maintenance database name interpreted as a connection string by pg_dumpall can reintroduce the same files. Exploitation is conditional on server mode, prior authenticated access, and the non-default presence of those credential files on the service account; no public exploit was identified at time of analysis, and the vendor fix is released in 9.19 (CWE-522, assessed CVSS:3.1 AV:N/AC:H/PR:L/UI:N/S:C/C:H/I:H/A:N).
Missing authorization in RAGFlow 0.25.2 through 0.27.2 allows any authenticated user who knows another user's private agent UUID to execute that 'Only me' workflow and read its output by posting to /api/v1/agents/chat/completions with "openai-compatible": true, a branch that returns before the UserCanvasService.accessible() check is reached. The attacker needs a valid account or API key (PR:L) plus the non-sequential agent UUID, and no victim-side configuration change is required because agents default to 'Only me'. Demonstrated impact is unauthorized private-workflow execution and output disclosure; retrieval of private data or side effects from configured tools depends on the victim workflow and was not demonstrated, and a public proof-of-concept walkthrough of the bypass exists, with no confirmed active exploitation reported. Version 0.27.3 contains the fix and this is prioritized for remediation this week.
Unauthenticated cross-site scripting in the WordPress eCommerce Product Catalog plugin, affecting versions up to and including 3.6.5, allows remote attackers to inject script that executes in a victim's browser when that user loads or interacts with attacker-influenced content. The issue is rated CVSS 3.1 7.1 and requires user interaction (UI:R); no public exploit identified at time of analysis, and no CISA KEV listing is present in the provided data.
Unauthenticated cross-site scripting in Mang Board WP through version 2.4.6 allows a remote attacker to supply script-bearing content that executes when a victim interacts with a crafted request, affecting WordPress sites running the vulnerable plugin. The CVSS 7.1 vector reflects network reachability with no privileges required but user interaction, scope change, and limited confidentiality and integrity impact; the specific vulnerable parameter or endpoint is not described in the available sources. No public exploit was identified at time of analysis, and the issue is prioritized for remediation this week.
Unauthenticated cross-site scripting in the VikRentCar WordPress plugin affects versions up to and including 1.4.9, letting a remote attacker get script executed inside a victim's browser session in the context of the affected site. The CVSS 3.1 vector (AV:N/AC:L/PR:N/UI:R/S:C/C:L/I:L/A:L, 7.1) confirms no authentication is required, but the attacker must induce a victim to load a crafted page or link, and the resulting impact is bounded to low confidentiality, integrity and availability effects with a scope change out of the vulnerable component. No CISA KEV listing and no public exploit reference appear in the supplied data, and the priority signal is this_week rather than an emergency.
Unauthenticated cross-site scripting in the Ad Inserter WordPress plugin affects versions through 2.8.19 and lets a remote attacker inject script that executes in a victim's browser after the victim is induced to load or click crafted content. The CVSS 3.1 vector (AV:N/AC:L/PR:N/UI:R/S:C/C:L/I:L/A:L) confirms no authentication is required but user interaction is decisive, and impact is bounded to the victim's browser/session context rather than server takeover. The signals point to a moderate, not emergency, priority, and no public exploit has been identified at time of analysis.
Reflected cross-site scripting in the All In One Login WordPress plugin through version 2.4.1 allows an unauthenticated remote attacker to run arbitrary script in a victim's browser within the context of the affected WordPress site, after convincing that victim to follow a crafted link or load an attacker-supplied page (CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:C/C:L/I:L/A:N, scope changed). Impact is limited to confidentiality and integrity of data reachable from the victim's session, with no availability effect assessed; the exact vulnerable parameter and endpoint are not described in the available sources. No public exploit code has been identified at time of analysis, and priority is assessed as 'this_week' - mixed signals leaning toward moderate rather than critical, mainly because successful exploitation depends on victim interaction and on the site running the plugin at an affected version.
Unauthenticated cross-site scripting in the Profile Builder WordPress plugin (Cozmoslabs) affects versions up to and including 4.0.5, letting remote attackers inject script that executes in a victim's browser. Per the CVSS vector (AV:N/AC:L/PR:N/UI:R/S:C), no attacker credentials are needed, but exploitation requires the victim to interact - clicking a crafted link or viewing attacker-supplied content - and impact is confined to the victim's browser session with low confidentiality, integrity and availability effect. The supplied base score is 7.1; no public exploit code or confirmed active exploitation was identified at time of analysis, and the specific vulnerable parameter and injection point are not described in the available sources.
Cross-site scripting in the Fast Velocity Minify WordPress plugin through version 3.5.7 allows unauthenticated attackers to have script run in a victim's browser when the victim interacts with attacker-supplied content, with scope change and low confidentiality and integrity impact. The user-interaction requirement is the primary limiting factor, and the specific vulnerable parameter, input field, or feature is not described in the available sources. No CISA KEV confirmation or public exploit was reported in the supplied intelligence; no public exploit identified at time of analysis, and the vuln.today decision label is this_week.
Unauthenticated cross-site scripting in BA Book Everything 1.8.33 and earlier lets an attacker induce a victim to open a crafted request so script runs in the victim's browser in the context of the affected WordPress site (CWE-79; reporting source CVSS 7.1). No attacker authentication is required (PR:N), but successful exploitation depends on user interaction (UI:R), which limits unattended mass exploitation. No public exploit was identified at time of analysis, and vuln.today places remediation in the this_week priority tier.
Release-gate substitution in the Zuke GitHub orchestration package @zuke/gh before 1.10.0 lets a party who can dispatch the external workflow satisfy or block a release gate without the gated suite ever validating the shipped artifact. The githubWorkflow(...) gate located the run it had just dispatched purely by a display-title marker echoed into the run list, so anyone able to read that marker and dispatch the same workflow could raise a substitute run whose conclusion the gate adopted; a passing run authorizes an unvalidated release and a failed or cancelled run holds a valid release closed, and the persisted run ID keeps the substitution across build resume. Exploitation requires marker-mode use plus a separate repository whose workflow_dispatch / actions:write holders are not equally trusted by the releasing repository; no public exploit code was identified at time of analysis, and the issue is fixed in 1.10.0. CVSS 3.1 base score 7.1 (High) with integrity as the dominant impact.
PipesHub 0.3.0 through 0.9.1 allows an authenticated user holding only the crawl:write OAuth scope to permanently destroy a connector instance and purge its graph records and embeddings by sending a crafted :connector path segment to the crawling manager schedule route. A value such as web.delete is concatenated into the published event type (web.delete.resync), which the Python connector service splits on the dot and dispatches to its delete handler, bypassing both the connector:delete scope check and the connector-lock guard enforced by the sanctioned DELETE /api/v1/connectors/:id endpoint. The privilege escalation is bounded to connectors the caller can already pass validateConnectorAccess for (their own personal connector, or an org admin's team connector); the vendor fixed the flaw in 0.10.0 by deriving the event action from the stored connector instance, and the advisory includes a verified, executed proof-of-concept walkthrough. The vulnerability is not listed in CISA KEV, and no EPSS score was supplied.
Cross-tenant file disclosure in abp-next-admin 10.4.0 through 10.7.0 lets any authenticated tenant user read blobs belonging to another tenant by supplying the victim's tenantId in the BlobManagement download-by-name request, which causes DownloadByNameAsync to switch tenant context and return the other tenant's stored files. The CVSS 4.0 vector (AV:N/AC:L/AT:N/PR:L/UI:N/VC:H/VI:N/VA:N) matches the description: low-complexity, network-reachable, requires a valid logged-in account, and impacts confidentiality only. Exploitation is limited to deployments running in multi-tenant mode with more than one tenant storing blobs, and the attacker must know or guess both the victim tenantId and the blob name; a public PoC reference exists (a walkthrough write-up), and the vendor has addressed the issue in available patches.
Authenticated users with author or editor roles in SonicJS through 3.0.0 can create, modify, publish, or soft-delete documents they do not own because the content write routes check only coarse role labels and fail to enforce per-document ACLs or administrator deny overrides. The flaw requires a valid bearer JWT or API key and is bounded to content integrity and availability rather than data disclosure. A vendor fix is referenced by PR #1147, and no public exploit identified at time of analysis is reported in the supplied data, with no CISA KEV entry listed.
Improper authorization in Orchard CMS 1.10.3 through 1.11 allows authenticated users holding the Author or Contributor role to display, clone and delete other users' media items by referencing their content ids, and to delete arbitrary media folders through FolderController.Delete(), because the media folder permission check never validates ownership. Exploitation requires a valid low-privilege account (CVSS 4.0 score 7.1 with PR:L), so anonymous or unauthenticated visitors cannot trigger it; risk is confined to multi-user sites that provision Author/Contributor accounts for semi-trusted people. A public technical write-up describing the issue exists but is not verified as working exploit code, and no CISA KEV entry or vendor-released fixed version appears in the available data.
Missing authorization in the doc::ajaxGetDoc() handler in ZenTao PMS through 22.6 lets any authenticated user - including a low-privilege account - read the body and draft content of private and access-restricted documents by enumerating sequential document IDs, because the handler returns content without calling checkPrivDoc(). The vendor issue tracker entry (GitHub issue #215) and a public technical writeup describe the flaw, and no fixed version is identified in the available data; exploitation is limited to confidentiality impact and requires only a valid login (PR:L, UI:N). No public exploit code has been confirmed and there is no CISA KEV entry, though a public PoC reference exists in the form of a third-party technical writeup that has not been verified as working exploit code.
Vatos DLP 1.1.176 through 1.1.242 stores domain credentials in plaintext, so an authenticated local low-privileged user who can read the product's stored configuration can recover reusable domain logins. Successful exploitation yields high confidentiality and integrity impact with no availability impact (CVSS 7.1, CWE-256); there is no remote or unauthenticated path in the assessed vector. No public exploit was identified at time of analysis, no active-exploitation signal is present in the provided intelligence, and the vendor fixed the issue in Vatos DLP 1.1.243.
Authentication bypass by spoofing in Dell OpenManage Enterprise versions prior to 4.8.0 lets an unauthenticated attacker who has adjacent network access hijack existing user sessions, scoring CVSS 3.1 base 7.1 (AV:A/AC:L/PR:N/UI:N/S:U/C:L/I:H/A:N). The attack is low in complexity and needs no credentials or user interaction, but it is bounded to attackers positioned on the same local or management network segment as the OpenManage Enterprise console rather than the open internet. No public exploit code or confirmed active exploitation was identified at time of analysis, and the vendor's stated complete fix is upgrading to version 4.8.0 or later.