Stored cross-site scripting in the Drupal contributed module Menu Link Attributes (versions 0.0.0 through 1.8.0) lets an attacker persist a script payload inside a menu link attribute, which then executes in the browser of any user who renders that menu. Exploitation requires an authenticated account holding the Drupal permission to create or edit menu links and set their attributes (PR:H in the CVSS vector), so realistic abuse is limited to insider or compromised-privileged-account scenarios rather than anonymous visitors. Severity is low (CVSS 3.8, EPSS 0.16%, 5th percentile), CISA SSVC records no known exploitation and no automatable path, the project's priority label is monitor, and a vendor patch is available; no public exploit has been identified at time of analysis.
HTTP request smuggling in OpenWrt uhttpd allows remote attackers to have attacker-supplied body bytes parsed as a second request on the same keep-alive connection when an invalid chunk-length line, such as a value exceeding INT_MAX, causes client_poll_post_data() to reset content_length and transfer_chunked without draining the body or closing the connection. Affected are unpatched uhttpd repository snapshots and builds predating fix commits b78f518478794e16ba3568fc1258e40bf3e8eb3b and 34460b8211d676ef1b9f40d6b775c5e112933690. The assessed vector is AV:N/AC:H/PR:N/UI:N with low confidentiality and integrity impact; exploitation is constrained by the need to align the embedded request and by typical LAN-only exposure, and there is no CISA KEV entry or confirmed working exploit, although the public advisory provides reproduction steps.
Integer overflow in go-xrpl's internal MPT amount decoders causes a crafted 33-byte MPToken amount with an 8-byte magnitude of 2^63 or greater to be decoded without proper range validation: ParseMPTAmountBinary wraps the value into a negative int64, and readMPTAmount emits an over-limit decimal string. This affects go-xrpl through 0.0.0-20260618134614-cb77d640ee26 and can corrupt EscrowData.MPTAmount and cause divergence from rippled's maxMPTokenAmount (2^63-1) interpretation, while the public binary codec already rejects such values. Exploitation requires supplying malformed ledger or transaction data (AC:H) and yields integrity-only impact; no CISA KEV entry or confirmed public exploit code was identified, although the advisory includes a local reproducer, and the vuln.today priority label is monitor.
An out-of-bounds read exists in GCC's libstdc++ implementation of std::from_chars(), which is reached when a program passes a character buffer that is not null-terminated. Only programs that link libstdc++ and call std::from_chars() on attacker-influenced, unterminated input are exposed; the GCC compiler and standalone toolchain are not themselves triggerable. The consequence is limited to availability - a crash or an invalid parse result - and the high attack complexity reflects the dependence on that specific calling pattern; no public exploit code or active exploitation has been identified at time of analysis, and the priority decision is monitor.
Access bypass in the Drupal contrib module Permissions by Term (versions 0.0.0 through 3.1.41) allows remote visitors to read nodes that taxonomy-term-based access rules are meant to hide, a limited information disclosure scored CVSS 3.7 (C:L, no integrity or availability impact). Only Drupal sites that run the module and actually enforce term-based node access are exposed, and exploitation depends on reaching a specific exceptional condition in the access check whose trigger is not described in the available sources, which is why attack complexity is rated high. No public exploit code was identified at time of analysis, EPSS is 0.15% (4th percentile), SSVC reports Exploitation: none and Automatable: no, and the vuln.today decision label is monitor.
Missing authorization checks in the Drupal contributed module Easy Breadcrumb (versions 0.0.0 through 2.0.11) allow a remote unauthenticated visitor to reach resources by directly requesting them rather than following the intended navigation path, a forceful-browsing condition whose confirmed impact is limited to low-severity information disclosure (CVSS 3.7). There is no confirmed active exploitation (not in CISA KEV) and no public exploit identified at time of analysis, and the EPSS score of 0.14% (4th percentile) plus an SSVC posture of exploitation:none / automatable:no place this in a monitor-and-schedule band rather than an emergency. The vendor has released a patch, so sites running the affected range should update the module.
Out-of-bounds read in Eye of GNOME (eog) when parsing a crafted PNG with malformed ICC color-profile metadata allows an attacker who convinces a user to open the image to crash the viewer and potentially expose limited memory contents. Affected deployments are GNOME desktop systems, including the Red Hat Enterprise Linux 6 through 9 package scope identified in CPE data; headless servers without eog installed are not exposed. Exploitation requires user interaction via a local vector, and no public exploit or CISA KEV confirmation was identified at time of analysis.
Stored cross-site scripting in the Drupal Data field contrib module (versions 0.0.0 through 2.0.14) lets an attacker with a privileged content-editing role persist markup inside a Data field that executes in the browser of another user who later views the rendered page. Only sites that have installed and enabled this optional contrib module are exposed; Drupal core is unaffected, and the vendor CVSS score is 3.5 (AV:N/AC:L/PR:H/UI:R/S:U/C:L/I:L/A:N), with our assessment corroborating low severity. EPSS is 0.17% (6th percentile), no CISA KEV listing, and no public exploit identified at time of analysis; a vendor patch is available per the Drupal advisory and the priority decision is monitor.
Immich before 3.2.0 lets an authenticated regular user attach another user's asset to a shared link they legitimately own: by submitting type=ALBUM with an attacker-owned albumId and the victim's UUID in the assetIds field, the server authorizes only the album yet persists the asset association, and the resulting anonymous shared-link URL then serves the victim asset's metadata, thumbnail and byte-identical original file to anyone who receives the link. The attack is confined to assets whose 128-bit UUID the attacker already knows out of band, so it is a targeted single-asset disclosure rather than bulk harvesting, and it requires a valid account on the affected server. Scored CVSS 3.5 (confidentiality-only, PR:L on the source vector), classified CWE-285 and triaged as monitor; no CISA KEV entry or other confirmation of active exploitation is present, and the vendor advisory (GHSA-g4w3-c888-cr93) includes a full step-by-step curl reproduction of the request sequence.
An out-of-bounds memory read in Ghostscript's mesh shading renderer lets an attacker who supplies a crafted PDF or PostScript document cause the interpreter to crash or leak a small amount of adjacent memory. The flaw sits in the handling of indexed color values inside mesh shading data, where bounds validation is incomplete; Ghostscript builds shipped with Red Hat Enterprise Linux 6, 7, 8, 9 and 10 are listed as the affected packages, so any user or automated pipeline rendering untrusted documents with those builds is exposed to the crash path. The assessed vector is local with user interaction (CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:U/C:L/I:N/A:L, base 3.3), meaning a victim must actually open or process the malicious file; there is no network-service exposure and no privilege escalation. No public exploit code has been identified at time of analysis and the issue is not in CISA KEV, so prioritization is monitor.
LibreDWG before 0.14 contains an error-path flaw in bit_TV_to_utf8() (src/bits.c) where a failure inside iconv() with an errno other than E2BIG frees the conversion buffer but still routes the original const input through bit_u_expand(), which rewrites \U+XXXX and \M+... sequences in place. The result is a SIGSEGV denial of service when the caller passed a string literal, read-only mapping, or other immutable buffer, and an unexpected mutation of the caller's input when the buffer is writable. The library is a C DWG read/write library, and practical reachability is limited: normal DWG decode/export paths heap-allocate most TV strings via bit_read_TV(), so an application must itself expose this conversion API to untrusted input and pass it non-writable storage - the assessed vector is local (AV:L), high attack complexity, requiring user interaction, with only low availability impact, and no public exploit identified at time of analysis; LibreDWG 0.14 contains the fix.
Cross-site scripting in the Drupal Country contributed module lets an attacker with high privileges on a Drupal site place script into pages that a victim then views after being induced to interact with the crafted content. The CVE description covers Country 0.0.0 through 2.1.3 and 2.2.0 through 2.2.1; the independent assessment rates the issue low severity (CVSS 3.1) and not an urgent priority because exploitation requires high privileges, user interaction and high attack complexity. No public exploit or active exploitation is identified in the supplied data (SSVC exploitation: none; EPSS 0.16%), and a vendor patch is available.
Server-side request forgery in Campfire 1.4.4 and earlier allows an authenticated user who can post a chat message to steer the server's OpenGraph link-unfurl fetcher toward internal network resources by supplying a link whose hostname resolves into special-use ranges the guard did not recognise - CGNAT 100.64.0.0/10, NAT64 64:ff9b::/96 or 64:ff9b:1::/48, and 6to4 2002::/16, the last two of which can embed an internal IPv4 target in their low bits. Exploitation requires authentication (CVSS PR:L), attacker-controlled DNS for the posted link, and the unfurl preview path being active, and the practical payoff is bounded because an unfurl returns only OpenGraph metadata and most internal endpoints return none - there is no integrity or availability impact. No public exploit code has been identified at time of analysis, and the vuln.today decision label is monitor, consistent with the low assessed severity (CVSS 3.1, availability and integrity unaffected).
Cloud Foundry CAPI versions before v1.248.0 and CF Deployment versions before 60.9.0 allow an authenticated SpaceDeveloper to redirect a space-scoped service broker's stored credentials to an attacker-controlled endpoint by changing the broker URL while leaving its credentials unchanged. This is an indirect credential-disclosure path because the CF API does not permit direct reading of broker credentials; the assessed vector is CVSS:3.1/AV:N/AC:L/PR:H/UI:N/S:U/C:L/I:N/A:N (2.7), with impact limited to confidentiality. No public exploit was identified at time of analysis, and vendor patches are available in CAPI v1.248.0 and CF Deployment 60.9.0.
Dell OpenManage Enterprise versions prior to 4.8.0 contain a server-side request forgery flaw (CWE-918) that a high-privileged, authenticated remote user could use to make the appliance issue requests on their behalf, resulting in limited information exposure. The CVSS base score is only 2.7 (Low) with confidentiality impact limited to Low and no integrity or availability effect, and exploitation requires an administrator-level account, so an attacker must already hold trusted credentials. No public exploit code, PoC reference, or CISA KEV entry is associated with this CVE, and the vendor priority label for it is monitor.
Płatnik, the ZUS (Polish Social Insurance Institution) client application, retrieves updates, libraries and support files over plain HTTP, so an attacker who occupies a man-in-the-middle position on the client's network path can read the transferred data and swap the legitimate files for their own. Successful abuse requires that network position plus the user initiating or accepting a download (UI:P/passive interaction), and impact is bounded to low confidentiality and low integrity with no availability effect, giving a CVSS 3.1 score of 2.3 (AV:N/AC:H/PR:N/UI:R/S:U/C:L/I:L/A:N). There is no public exploit identified at time of analysis and no KEV listing; the vuln.today decision label is monitor, consistent with a low real-world risk that is resolved by updating to version 10.02.002 (metric 242).
Coop, the open source review and moderation tool from ROOST, fails to invalidate database-backed sessions when a user resets their password, so any session an attacker had already established survives the reset for up to the 30-day session maximum age. The flaw requires a prerequisite compromise - an attacker must already hold the victim's password and have created an authenticated session before the legitimate user completes a reset (PR:L, AT:P), meaning it grants no initial access on its own. It affects versions prior to 1.0.2 and is fixed in 1.0.2; this is a low-severity issue (vendor CVSS 2.3) that only manifests after prior credential compromise and only when the victim performs the reset. No public exploit code is identified at time of analysis, and the priority label is monitor.
pgAdmin 4 before 9.19 fails to invalidate a user's existing sessions when that user's password is changed, whether through the user's own change-password dialog or through an administrator reset in User Management. Authenticated attackers who already hold the victim's pre-change password or a stolen pre-change session cookie retain a usable session from the moment of the change until that session naturally expires, so the very action intended to cut off access does not do so. The assessed impact is limited (CVSS 3.1: AV:N/AC:H/PR:L/UI:N/S:U/C:L/I:L/A:N) and exploitation is conditional on prior credential or cookie compromise plus a password-change event; no public exploit has been identified at time of analysis, and the vendor has released a fix in 9.19 that rotates the session identifier on password change.
Server-side request forgery in the open-source Discord bot Quest Bot (versions up to and including 1.1.8) lets any ordinary member of a server the bot has joined drive the bot's 'image to GIF' command against HTTPS services bound to loopback or internal addresses. The command accepts a user-supplied URL, and the helper that is supposed to block internal destinations validates one DNS answer but then re-resolves the hostname when the request is actually sent, so an attacker who controls a domain with a short TTL can answer public during validation and internal during connection (DNS rebinding, CWE-918). Impact is deliberately narrow: the request is pinned to HTTPS on port 443 and content is only returned when the response is a genuine image, so the realistic outcome is blind internal probing with limited retrieval of image responses, not arbitrary data disclosure, and plain-HTTP cloud metadata services are not reachable by this path; severity is low (CVSS 4.0 score 2.3, assessed CVSS 3.1 AV:N/AC:H/PR:L/UI:N/S:C/C:L/I:N/A:N) and the priority label is 'monitor'. The advisory publishes PoC reproduction steps (a loopback-only HTTPS server, a rebinding domain, and repeated command invocations), though no packaged working exploit is confirmed; the flaw is fixed in 1.1.9, and there is no KEV listing or evidence of active exploitation.
ShowDoc through 3.9.5 lets an authenticated user with edit rights on any single parameter-description library delete parameter-description entries belonging to other projects, because the deleteEntries API only verified edit permission for the first entry id in the submitted list. An attacker supplies one entry id they legitimately control as the first element and appends victim entry ids; the permission check passes on the first entry and the delete then runs across the whole array, permanently removing the appended records. Confidentiality is unaffected and the assessed severity is low (CVSS 3.1: 5.4-equivalent vector scored as AV:N/AC:L/PR:L/UI:N/C:N/I:L/A:L, vendor CVSS 4.0 base 2.3); no CISA KEV entry exists, and a public technical writeup of the bypass is referenced, though it is not verified as working exploit code. Priority is 'monitor'.
Authenticated FilePress users can delete, rename or read folders belonging to other libraries in versions through 3.0.1, because the pichome permission check validates membership of any library instead of the specific library being operated on. The attacker must supply a foreign library appid or folder id to folder.php (deletefolder, rename) or rightdata.php (getfolderdata), so practical impact is bounded by how easily those identifiers can be learned; scoring reflects that prerequisite, with the input CVSS 4.0 vector at 2.3 (AT:P) and our assessment at CVSS 3.1 AV:N/AC:L/PR:L/UI:N with low confidentiality, integrity and availability impact. No public exploit code is confirmed at time of analysis, though a public technical write-up describing the IDOR is referenced, and the issue is a monitor-priority item rather than an active emergency.
Parse Server 9.7.0 through 9.10.5-alpha.6 can return protected field values to an authenticated client holding create or update permission on an object, when the deployment sets the protectedFieldsSaveResponseExempt option to false and a beforeSave trigger writes beneath a protected field using a dot notation key. The leak is confined to values that the trigger writes in the same create or update request; the save-response field stripping simply failed to remove those nested keys. Exploitation requires the non-default option, a specific trigger-implementation pattern and existing write permission on the object, and no public exploit was identified at time of analysis; a fix is available in 9.10.5-alpha.7 and Parse Server 8 is not affected because it lacks the option.
Authenticated FTP users can bypass vsftpd’s cmds_denied command filter by sending RFC 775 X-prefixed aliases such as XRMD, XMKD, XCWD, XPWD, and XCUP, allowing operations the administrator intended to block, including directory deletion. The issue affects vsftpd through 3.0.5 and only matters when the server is explicitly configured with cmds_denied or an equivalent command restriction; default installations without such filtering have nothing to bypass. It is low severity with a CVSS 4.0 base score of 2.3 and integrity-only impact, no CISA KEV entry or public exploit code has been identified, and the vuln.today priority is monitor; the supplied vector indicates authenticated access (PR:L).
SQL injection in phpMyAdmin through 5.2.3 allows an authenticated user to inject SQL through the unescaped col_extra and col_attribute fields submitted to the /database/central-columns endpoint, reaching INSERT and UPDATE statements run against the pma__central_columns table with controluser privileges. Exploitation requires a working phpMyAdmin login plus a deployment where configuration storage (pmadb) is configured and the Central Columns feature is in use, so impact is bounded to reading or altering other users' configuration-storage data rather than arbitrary application databases. The vendor priority is 'monitor' and CISA KEV does not list this CVE; a public research writeup documents the flaw, but no working exploit code is confirmed, and the assessed severity is low (CVSS 3.1 AV:N/AC:H/PR:L/UI:N/S:U/C:L/I:L/A:N).
Logged-in OpenCart customers can place unpaid full-price orders on releases through 4.1.0.4 by manipulating the reward-points and Free Checkout workflow: apply points, save Free Checkout, then submit zero points via reward.save so the confirm handler marks the order as Free Checkout without payment. The flaw requires an authenticated customer account and a store with both the reward points extension and Free Checkout enabled, and it only affects order integrity and merchant revenue, not confidentiality or availability (CVSS 2.3; CWE-841). No public exploit code is confirmed at time of analysis, though a public research write-up describes the sequence; no KEV entry is present in the supplied data, and the priority is monitor.
Cross-organization invoice payment manipulation in Lago through 1.55.0 allows an authenticated user who controls their own organization's Stripe integration to mark another organization's invoice as paid by signing a payment_intent.succeeded webhook containing the victim invoice UUID. The flaw is an unscoped invoice lookup in Invoices::Payments::StripeService; impact is limited to integrity, with no confidentiality or availability effect, and exploitation requires prior knowledge of a non-enumerable UUID (assessed CVSS 2.3). No working public exploit is confirmed in the supplied sources and no fixed version is listed; the vuln.today decision is monitor.
SQL injection in phpMyAdmin's Designer saveSetting() handler (all versions through 5.2.3) lets an authenticated user who can reach the Designer feature inject SQL through the index and value parameters of operation=save_setting_value on /database/designer, allowing other users' stored Designer settings to be overwritten or subqueries to be run over the controluser connection. The bar is high: an account is required, and the phpMyAdmin configuration storage (controluser / pmadb with the designer settings control table) must be configured, since the injection lands in control-database tables rather than user data or the host. Upstream fixes exist as commits and pull request 18698; severity is scored low (CVSS 4.0 base 2.3; assessed CVSS 3.1 AV:N/AC:H/PR:L/UI:N/S:U/C:L/I:L/A:N), the priority label is monitor, and there is no CISA KEV entry - a public research writeup describing the flaw exists, but it is not verified as working exploit code. This is a low-priority, high-barrier issue rather than an urgent threat.
Local file disclosure in Pi-hole 6.4.3 and earlier lets an authenticated Pi-hole administrator who can add adlists - and who can pre-position a symlink plus a pihole-readable decoy on the host filesystem - cause gravity.sh to read a root-only file through a crafted file:// adlist. The flaw is a check-versus-read path resolution mismatch: the readability test runs as the pihole user via the kernel (symlinks first, then dot segments) while curl strips dot segments from the URL text before opening the file, so the two operations can land on different files; the root-only read only occurs when gravity is launched as root via 'pihole -g' or cron. With a vendor severity of 2.2 (CWE-59), mandatory admin privileges and mandatory local filesystem write access, this is a low-priority issue: no public exploit code was identified at time of analysis, no CISA KEV confirmation, and vuln.today priority is 'monitor'. A vendor fix ships in Pi-hole 6.5.
Cross-namespace authorization bypass in Traefik's Kubernetes Ingress NGINX provider (v3.7.11 through v3.7.14) lets a user who can create Ingress resources have a CA secret belonging to another namespace enforced as the client certificate authority on their own routes, and infer whether that secret exists and is in use, although its contents are never disclosed. The bypass only occurs when a client-authentication TLS option is already cached from another Ingress referencing that same CA secret, because the cross-namespace check runs only on the first build of the option while allowCrossNamespaceResources is disabled (the default); the flaw was introduced in v3.7.11 when client-auth TLS options began being shared across Ingresses, and Traefik v2 and pre-3.7.11 releases are unaffected. Severity is low (CVSS 2.1 v4.0; independently assessed CVSS 3.1 5.0-equivalent vector AV:N/AC:H/PR:H/S:C/C:L/I:L/A:N), the priority decision is monitor, and no public exploit or confirmed active exploitation was identified at time of analysis.
SQLite before 3.54.0 can disclose column values that an application explicitly denied through sqlite3_set_authorizer() SQLITE_READ restrictions when a RIGHT or FULL JOIN is used with a USING clause naming the restricted column. Only applications that register an authorizer callback and use it to block specific column reads are exposed, and the attacker must be able to submit SQL to that connection; the impact is limited to confidentiality of the shared USING column, with no write or denial-of-service effect. The vendor-released fix is SQLite 3.54.0, and this is a low-priority issue across every available signal; exploit code references exist in the public advisories and research writeup, but active exploitation is not indicated.