SQL injection in Apache SIS's sql-storage module (versions 1.0 through 1.6) allows remote attackers to inject SQL through a crafted sort clause that is copied after ORDER BY without identifier quoting. Exploitation requires an application built on Apache SIS to use the sql-storage module and to derive the ORDER BY sort-clause identifier from attacker-controllable input; reachability is the key limiting factor. Other modules such as sis-referencing are not affected, and no public exploit identified at time of analysis.
Open redirect in Mautic's public form submission endpoint (/form/submit) causes the application to send a visitor to a return address supplied with the submission without confirming that it resolves to the Mautic instance or another trusted host, so a crafted form can bounce visitors from a legitimate Mautic domain to an attacker-chosen external site such as a phishing clone or a malware delivery page. Every Mautic release below 5.2.12, 6.0.10 and 7.2.2 is affected, and any instance with a published public form is exposed; the attacker needs no authentication (PR:N) but the victim must actually submit the attacker-prepared form (UI:R), which keeps this a moderate-priority, social-engineering-dependent issue rather than an emergency. No public exploit code was identified at time of analysis, and the vendor rates the impact as limited confidentiality and integrity only (CVSS 5.4).
Missing authorization in the Geo Mashup WordPress plugin through 1.13.22 lets a low-privileged authenticated user reach functionality that should be restricted, allowing limited reads and modifications of protected data (CVSS 5.4, AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:N). Any WordPress site running the plugin at or below 1.13.22 with self-registration or otherwise obtainable subscriber-level accounts is exposed; the assessed vector requires PR:L, so a valid low-privileged account is a prerequisite. There is no confirmed active exploitation (not in CISA KEV) and no public exploit identified at time of analysis, and the reported severity is capped at low confidentiality and integrity impact with no availability effect.
Tornado's HTTP/1.x header parser consumes quadratic CPU time when a header value is spread across many obsolete line-folding (obs-fold) continuation lines, allowing a client to stall the single IOLoop thread and delay every other connection that process serves. All Tornado HTTP servers running versions prior to 6.6 are affected, because header parsing happens before any application code, so no route, handler, or credentials are needed; only availability is impacted (C:N/I:N/A:L). At the default 64 KB max_header_size a single malicious request blocks the loop for tens of milliseconds, so meaningful degradation requires repeated requests, and the impact scales with the square of the configured header limit; EPSS is low (0.38%, 32nd percentile) and no public exploit has been identified at time of analysis, though the vendor has released a fix in 6.6.
Event-loop starvation in the Tornado Python web framework and async networking library before 6.6.0 allows a remote, unauthenticated attacker to stall every connection handled by a worker process by submitting an HTTP/1.1 chunked request body made of millions of one-byte chunks to any handler that reads the body. Because the decoder processes already-arrived body bytes without yielding to the event loop, a body near the default 100 MB max_body_size can block the process for close to a minute, and for handlers without @stream_request_body each retained chunk costs roughly twenty times the received bytes in memory. The same behaviour affects SimpleAsyncHTTPClient when it reads a response from a malicious server; no public exploit identified at time of analysis, and the issue is addressed in the 6.6.0 release rather than treated as an emergency by available triage signals.
Information exposure through error messages in a2aproject a2a-java up to 1.2.0.Final lets an attacker who can induce a server-side exception read raw internal exception text from the returned JSON-RPC or REST error payload. Affected deployments are those running the reference Quarkus JSON-RPC/REST routes or the gRPC transport, where the error handler built InternalError directly from the thrown exception's message; per the vendor vector the attacker holds low-level privileges (PR:L), although the real authentication bar depends on how the A2A server is exposed. Impact is limited to low confidentiality with no integrity or availability effect, there is no CISA KEV listing and no public exploit identified at time of analysis, and EPSS is 0.32% (24th percentile), consistent with a low-priority disclosure issue that is resolved by upgrading to 1.3.0.Final.
Information exposure through error messages in the .NET A2A SDK (a2aproject a2a-dotnet) up to and including 1.0.0-preview2 causes the JSON-RPC endpoint to echo raw System.Text.Json parser output back to the caller. When a request body fails parsing, A2AJsonRpcProcessor.ProcessRequestAsync returns the -32700 ParseError response containing ex.Message, which discloses JSON paths, line and byte positions, exception type names and library details rather than a generic error. The exposure is diagnostic metadata only - no credentials or user data are implicated, and the independently assessed impact is low (CVSS 3.1 AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:N/A:N); no public exploit has been identified at time of analysis, and upgrading to 1.0.0-preview3 resolves the issue.
Code injection in the Seroval JavaScript serialization library 1.6.0 through 1.6.10 lets attacker-controlled text become executable JavaScript in the serializer's output when an application copies an untrusted own `constructor` property onto a typed array and then passes that array to serialize, serializeAsync, crossSerialize, crossSerializeAsync, crossSerializeStream, the Serializer class, or compileJSON. The vendor rates the result low severity because that property-copying pattern is described as unusual, and the injected script only executes once a victim loads a page that receives the serialized output (CVSS 4.0 AV:N/AC:H/AT:P/PR:N/UI:P with scope change to confidentiality and integrity). There is no confirmed active exploitation (not in CISA KEV; SSVC exploitation: none) and no public exploit identified at time of analysis, with EPSS at 0.27% (19th percentile), and the flaw is fixed in Seroval 1.6.11.
Parse Server's $relatedTo relation query does not validate the owning object reference, so an authenticated client can substitute a query operator for the object ID and slip past the owning-object ACL check introduced for GHSA-wmwx-jr2p-4j4r. The result is a limited information disclosure: the attacker learns which objects they can already read are linked to owning objects they are not permitted to read, in deployments that expose Relation fields to queries. Severity is moderate (CVSS 5.3, PR:L, confidentiality-only), EPSS is low at 0.25% (17th percentile), there is no CISA KEV listing, and no public exploit has been identified at time of analysis.
Information disclosure in the VMime C++ MIME library through 0.9.2: the uuEncoder::decode routine trusts the line-length byte of a uuencode/x-uuencode part, so a crafted part with an over-long length byte causes uninitialized stack memory to be written into the decoded output - up to 15 bytes of stack contents per over-long line. Affected are applications that use VMime (kisli:vmime, all versions up to and including 0.9.2) to parse attacker-supplied MIME messages; no authentication is required, but the victim must have the crafted message decoded by the application (UI:P in the CVSS 4.0 vector, UI:R in the assessed 3.1 vector), and the leaked bytes are only useful if they are observable in the resulting decoded attachment. There is no CISA KEV entry and EPSS is low (0.25%, 17th percentile), no public exploit code has been confirmed at time of analysis, and the vendor has published a fix upstream (PR #363, commit e85925ed); this is a genuine but low-to-moderate priority flaw rather than an emergency.
Missing authorization in OpenPanel's onboarding.project tRPC mutation lets any authenticated organization member - including a read-only one - create unlimited projects and mint plaintext write-type ingestion credentials inside that organization, bypassing the admin-only policy that the sibling procedures project.create and client.create enforce. All releases before 3.0.0 are affected (2.3.0 confirmed vulnerable), and the minted write-type secret cannot read existing data or manage resources, so impact is bounded to injecting fake analytics events, poisoning business metrics and consuming the organization's usage budget. EPSS is low (0.24%, 15th percentile), there is no CISA KEV entry, and the upstream GHSA advisory publishes a working PoC request (a single tRPC call over the dashboard httpLink with a non-admin member session cookie).
Improper escaping in OpenPanel's sankey.service.ts lets any authenticated member of a project - read access alone is sufficient - supply an events[].name value containing a backslash to the chart.sankey procedure and break out of the ClickHouse string literal, producing a parse error that takes the project's Sankey chart endpoint offline. The flaw affects OpenPanel releases before 3.0.0 and is a self-contained availability issue: only the one chart feature the attacker can already read is disrupted, with no confidentiality or integrity impact and no cross-tenant reach. The reporter could break the literal but not convert it into a working injection, so the practical effect is denial of a single feature; EPSS is only 0.23% (14th percentile), the issue is not in CISA KEV, and no confirmed exploitation has been reported. Authentication requirements are confirmed as low-privileged (PR:L in the CVSS vector). Vendor-released patch: 3.0.0.
Read-level project members in OpenPanel before 3.0.0 can retrieve the argon2id hashes of every ingestion credential in a project by calling the client.list tRPC query, which returns full Prisma Client rows because the findMany call carries neither a select nor an omit clause. The vendor advisory includes a PoC reference request showing the expected response shape; it is not verified as working exploit code, there is no CISA KEV entry, and EPSS is 0.22% (12th percentile). Realistic risk is limited: the underlying secrets are roughly 80-bit random values and the stored hashes use argon2id (m=19456, t=2, p=1), so the leak chiefly enables offline confirmation of secrets obtained elsewhere without using the credential against the live API and without leaving an audit trail.
Race condition in a2a-dotnet through 1.0.0-preview3 allows a low-privilege authenticated remote client to disturb task status and background-cancellation bookkeeping by submitting concurrent SendMessage or SendStreamingMessage requests for the same TaskId, producing limited integrity and availability impact. The trigger is a CWE-362 race in TaskProjection.ApplyStatus and the A2AServer cancellation-registration path, which the proposed fix in PR #453 replaces with lock-guarded reservation and release. Severity is moderate (CVSS 5.3), no public exploit code has been identified at time of analysis, and the fix has not yet been accepted or released.
Stored cross-site scripting in YesWiki before 4.7.0 lets unauthenticated remote attackers persist JavaScript into the wiki's referrer log by sending a single request whose Referer header carries HTML metacharacters; the payload executes for any user who subsequently opens the '?PageName/referrers_sites' handler view, giving the attacker same-origin fetch/XHR capability in that visitor's session. Because session cookies are HttpOnly and the payload stays dormant until the handler page is viewed, impact is constrained to content exfiltration and actions taken as the victim rather than direct cookie theft; EPSS is low at 0.18% (7th percentile), there is no KEV entry, and the upstream GitHub advisory includes a curl-based PoC reproduction. The vendor-released fix is YesWiki 4.7.0, and per the vuln.today decision this is a this-month priority rather than an emergency.
Reflected cross-site scripting in YesWiki before 4.7.0 lets unauthenticated remote attackers run JavaScript in a victim’s browser when the victim opens a crafted link targeting a Bazar entry that lacks a stored field key. The attacker supplies the missing field name as a query parameter, and BazarField::getValue() falls back to unsanitized $_REQUEST data rendered through Twig |raw templates; exploitation requires victim interaction and a form entry created before the field was added. Overall risk signals converge on a genuine but moderate-priority issue rather than an emergency: EPSS is 0.18% (7th percentile), no KEV listing is present, a public PoC reference exists but is not verified as working exploit code, and vendor patch 4.7.0 is available.
An app already running on an affected Mac can win a race condition to corrupt process memory or cause unexpected system termination; Apple corrected the locking defect in macOS Golden Gate 27.2, macOS Sequoia 15.8.2, and macOS Tahoe 26.7.2. Reachability is local only (CVSS 5.3, AV:L/AC:H/PR:L/UI:N), so an attacker needs existing low-privilege code execution on the host and must then win a timing window, making success non-deterministic and matching the SSVC 'Automatable: no' rating. No public exploit code or active exploitation has been identified at time of analysis, and EPSS is 0.15% (4th percentile), which supports scheduling the update in the current monthly cycle rather than treating it as an emergency.
Service metadata tampering in Esri ArcGIS Server 11.3 through 12.0 allows a remote, unauthenticated attacker with network reachability to update service metadata without user interaction, affecting integrity only (CVSS 5.3). The vendor restricts the affected window to 11.3-12.0 and provides patches for 12.0, 11.5, and 11.3; no public exploit code or CISA KEV entry was identified at time of analysis. This is a moderate, this-month remediation priority rather than an emergency.
Out-of-bounds read in the Box-Average Downsampler engine of FastStone Image Viewer (function FUN_0062d96c) affects versions up to and including 8.5 and can be triggered when a user opens a crafted image file in the application. Exploitation requires the victim to open the malicious file (CVSS 4.0 base score 5.3; assessed CVSS 3.1 AV:N/AC:L/PR:N/UI:R/S:U/C:L/I:N/A:L), and because the trigger is confined to the downsampling path there is no remotely reachable service and no unattended exploitation route. Impact is bounded to partial memory disclosure or an application crash rather than code execution; no public exploit code was identified at the time of analysis, and the vendor was contacted about the disclosure but did not respond.
Resource exhaustion in NLnet Labs Routinator lets an attacker make the validator's HTTP server unavailable by opening a large number of connections that are never closed, consuming its finite open-connection capacity (CVSS 5.3, availability-only). Only operators who deliberately expose Routinator's HTTP or RTR server to untrusted networks are reachable; in a standard deployment bound to localhost or trusted internal networks the flaw is not remotely exposed. No KEV listing and no public exploit identified at time of analysis, so urgency is driven by deployment topology rather than an active campaign.
LearnPress, the WordPress learning-management-system plugin from ThimPress, leaves access-controlled functionality reachable without proper authorization in all versions through 4.4.11, letting remote attackers retrieve restricted information from affected sites. Per the assessed vector (AV:N/AC:L/PR:N/UI:N) no credentials or user interaction are needed, but the exposure is bounded: confidentiality impact is LOW while integrity and availability are unaffected, so this is information disclosure rather than account takeover or site compromise. No public exploit code and no confirmed active exploitation were identified at time of analysis, and the priority decision places it in this-month remediation rather than emergency response.
Missing authorization (CWE-862) in the TMS Amelia WordPress booking plugin, affecting versions through 2.4.10, lets remote unauthenticated attackers reach functionality that should require proper access-control checks, producing a low integrity impact with no confidentiality or availability loss (CVSS 5.3, PR:N/UI:N). Any WordPress site running a vulnerable Amelia version is potentially exposed, though the sources do not identify the specific endpoint or configuration that must be reachable, and whether a non-default setup is required is not described. A public PoC reference exists in the Patchstack advisory; it is not verified as working exploit code, and there is no CISA KEV confirmation of active exploitation.
Improper authentication in the Akubela HyPanel PS51 web login handler (firmware 51.130.40.26) lets a remote attacker reach an authenticated state by sending a request to /login with the engrmode argument set to 'admin'. Severity is moderate - CVSS v4.0 5.3 from the reporter, with an independent 3.1 assessment of AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:L - and all three impact axes are rated low, so the resulting access may be limited rather than full device administration. A public proof-of-concept reference exists (a GitHub repository titled 'Authentication-Bypass-Vulnerability-in-akubela-HyPanel-PS51'), but it is not verified as working exploit code; no CISA KEV listing or EPSS score was supplied, and the vendor was contacted about the disclosure without responding, so no vendor patch is available.
Kantech Entrapass before 9.20 stores embedded sensitive information in cleartext in memory, so an actor with local access, high-level host privileges, and the required user interaction can retrieve that data. The issue is rated CVSS 5.3 and assessed as a moderate, low-urgency information-disclosure risk rather than a remote compromise path, because exploitation depends on pre-existing privileged local access and realistic exposure is confined to insiders or post-compromise situations. Vendor patching is available (upgrade to 9.20 or later); no public exploit identified at time of analysis.
A use-after-free in the Zephyr RTOS Bluetooth Classic host lets an adjacent, unauthenticated remote Audio Gateway crash or permanently hang the Bluetooth host thread on devices running 4.4.0 through 4.4.2 as an HFP Hands-Free unit (CONFIG_BT_CLASSIC and CONFIG_BT_HFP_HF enabled). When bt_hfp_hf_connect() starts an SDP discovery and the following bt_rfcomm_dlc_connect() fails, the old code wiped the live HFP HF object with memset(), destroying a node still linked into the SDP client's request list and a possibly-queued work item; the host later dereferences the NULL pool and function pointers or walks a self-referential list. Impact is denial of service only (CVSS 5.3, AV:A/AC:H/PR:N/UI:N/C:N/I:N/A:H, CWE-416) - the stray indirect call goes through an all-zero pointer the attacker does not control, so there is no code execution. No public exploit code was identified at time of analysis, but reachability depends on product application code that feeds a peer-discovered RFCOMM channel into bt_hfp_hf_connect(); the only in-tree caller is the shell sample, which limits exposure in default builds.
OS command injection in Sunwood-ai-labs github-kanban-mcp-server through version 0.4.0 lets an attacker who can influence the state or labels arguments of the handleListIssues tool inject shell metacharacters that reach an operating-system command (CWE-78). The flaw is network-reachable and unauthenticated per the assessed vector (CVSS:3.1/AV:N/AC:L/PR:N/UI:R), but it requires the operator or driving agent to invoke the issue-listing operation before the injected input executes, so it is not a fully autonomous remote trigger. Exploitation has not been confirmed in the wild and no public exploit code was identified at time of analysis; the vendor released version 0.4.1 plus commit 253d93bbdf7ee9fe42bd8fdf15feea93ae59cbae as the fix.
Cross-site scripting (CWE-80) in the WpEvently WordPress plugin from Magepeople Inc. affects all versions through 5.7.9 and lets an attacker place script-related HTML that executes in a victim's browser once the injected content is rendered. The supplied vector scores 5.3 with PR:N, but real-world exploitation still hinges on a victim loading the affected page (our assessment uses UI:R rather than UI:N), so impact is bounded to visitor-side script execution such as content or session manipulation rather than server compromise; the vulnerable parameter and whether the payload is stored or reflected are not described in the sources. Only a Patchstack database entry is available, no public exploit code has been identified at time of analysis, and the issue is prioritized as this-month rather than immediate action.
A pre-authentication out-of-bounds read in the SNMP component of Arm Keil MDK-Middleware 8.0.0 through 8.3.0 lets a remote, unauthenticated attacker reach SNMP parsing code before any credential check and read past the bounds of an internal buffer, with the assessed impact limited to availability (CVSS 5.3, AV:N/AC:L/PR:N/UI:N/C:N/I:N/A:L). Only devices whose firmware is built on the affected middleware with the SNMP component compiled in, enabled and network-reachable (typically SNMP on UDP 161) are exposed; devices that do not ship or expose this stack are not exploitable. Exploitation has not been observed (CISA SSVC Exploitation: none) and no public exploit was identified at time of analysis, so this is a medium-severity denial-of-service-class issue rather than an emergency, though the automatable nature of the flaw warrants patching within the vendor's normal cycle.
Sandbox isolation in DeepSeek Harness 0.1.0-rc.0 through 0.1.0-rc.5 can be manipulated through the createContext function of the cordis-host-runner extension (packages/extensions/cordis-host-runner/src/sandbox.ts), giving remote, unauthenticated attackers limited confidentiality, integrity and availability impact - but only where a user of the affected deployment takes an interaction step, as reflected in the UI:P metric. Impact is capped at LOW on all three pillars with no scope change, so the realistic outcome is a weakened sandbox boundary rather than host takeover. No CISA KEV entry and no public exploit code were identified at time of analysis, no vendor patch or fixed version is listed, and the vendor was contacted about the disclosure without responding; VulDB-derived triage places the issue in the 'this_month' priority band.
Missing authentication on several boneIO Black API endpoints lets any unauthenticated caller who can reach the device on the local network read system logs, the firmware version, the device serial number, the configured MQTT username, configuration file listings and the interactive API documentation. The exposure is confined to information disclosure - there is no integrity or availability impact - but the harvested details are directly useful for reconnaissance and for planning follow-on attacks against the device or the MQTT broker it talks to. All versions prior to 1.6.0 are affected; 1.6.0 requires an authenticated role on every API route, disables the interactive API documentation and withholds the serial number from anonymous callers. No public exploit code was identified at time of analysis and there is no CISA KEV entry, so the practical risk is bounded by the adjacent-network requirement.
Information disclosure in Kameleo through version 5.3.1 lets an unauthenticated remote attacker read data exposed by the /general/user-info endpoint of the Local API component, provided the victim is induced to perform an interaction that causes the request to be issued (CVSS 4.0 base 5.3; assessed CVSS 3.1 AV:N/AC:L/PR:N/UI:R, C:L only). Only low-confidentiality data is at risk - there is no integrity or availability impact - and the Local API must be active and reachable along the attacker's request path, which typically means a browser-mediated or cross-origin trigger against a normally localhost-bound service. No public exploit code was identified at time of analysis, and the vendor did not respond to the disclosure, so no vendor-released patch has been identified.
Information disclosure in the Local API of Kameleo (versions up to and including 5.0.1.26) lets a remote, unauthenticated attacker read data exposed by the /profiles endpoint, with the assessed CVSS 3.1 score of 5.3 reflecting confidentiality-only impact (C:L) and a required victim-side action (UI:R). Exploitation is opportunistic rather than mass-scale: Kameleo must be running with its Local API reachable along the attacker's request path, and the exact mechanism by which /profiles leaks data is not described in the available sources. The vendor addressed the issue in release 5.2.0, whose security note states that proxy passwords are now redacted in API responses; no confirmed active exploitation (CISA KEV) or verified public exploit code was identified at time of analysis, and the vulnerability is triaged as a plan-and-patch item rather than an emergency.
Information disclosure in Kameleo up to and including 5.3.1 allows remote attackers to read cookie data held by the client engine by manipulating the 'guid' argument passed to CookieController.ListCookies in client-engine-application/w.cs. The assessed CVSS 3.1 vector (AV:N/AC:L/PR:N/UI:R/S:U/C:L/I:N/A:N) indicates no privileges are needed and the attack is network-reachable, but exploitation depends on passive user interaction, and impact is confined to low confidentiality with no integrity or availability effect. No vendor patch or fixed version is identified - the vendor was contacted and did not respond - and there is no CISA KEV entry or confirmed public exploit code; the references include an external Google Drive link whose contents are not characterized in the available data.
Reflected cross-site scripting in RosarioSIS before 12.9.2 allows an unauthenticated attacker to break out of the double-quoted form action attribute on the Food_Service MenuReports page by supplying crafted menu_id (or modname) values through Modules.php, injecting an onmouseover handler that can steal session and CSRF tokens and let the attacker act as the administrator. Exploitation is gated by a hard prerequisite: a logged-in administrator must open the attacker's crafted link and interact with the injected element for the payload to execute in their session, which is why the CVSS 4.0 base score is 5.3 despite PR:N (user interaction is passive). The vendor addressed the flaw in version 12.9.2 (commit 08c871d wraps the reflected values in URLEscape()); no public exploit identified at time of analysis, and priority is triaged as this_month rather than emergency.
Frappe Builder before 1.36.1 lets any authenticated user bypass its SSRF guard by supplying a public URL that redirects, or DNS-rebinds, to internal or link-local addresses. The convert_to_webp endpoint lacks a permission decorator, so any authenticated Builder user can reach cloud metadata services, scan internal ports, or hang requests because no timeout is set; import_remote_assets and import_remote_fonts require Builder Page write permission. This is a moderate, non-emergency issue, and no public exploit was identified at time of analysis.
Missing authorization in Frappe Builder's builder.ai.api.improve_prompt endpoint lets any logged-in non-Guest session - including a Website User with no Builder access whatsoever - run LLM completions on the AI provider key the site owner configured for Builder, spending the owner's paid credits and returning the model output to the caller. Versions 1.33.0 through 1.35.x are affected; 1.36.0 removes the endpoint entirely. Exploitation still requires that an AI provider key (OpenRouter per the report) is actually stored in Builder's AI settings, and the impact is limited to resource consumption - no site data read or write, no system compromise. No public exploit code or CISA KEV entry was identified at time of analysis, and the vulnerability is scheduled for routine (this-month) handling rather than emergency response.
Missing authorization (CWE-862) in the WordPress plugin Contact Form by Supsystic, affecting all versions through 1.10.7, lets a remote attacker reach functionality that should have been gated by an access-control check without authenticating. The assessed vector (CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:L/A:N, 5.3) confines the impact to low integrity - an attacker may be able to make some unauthorized modification, but cannot disclose data or disrupt availability through this flaw. The precise plugin feature or action that lacks the authorization check is not described in the available sources, and no public exploit has been identified at time of analysis; priority is rated this_month.
Missing authorization (CWE-862) in the WPExperts APIExperts Square for WooCommerce plugin, affecting all versions up to and including 4.7.5, allows unauthenticated remote attackers to reach functionality that should sit behind an access-control check, with a partial integrity-only impact and no effect on confidentiality or availability (CVSS 5.3). The CVSS vector (AV:N/AC:L/PR:N/UI:N) indicates the flaw is reachable over the network against default configurations with no credentials and no user interaction, although the concrete protected function or data exposed is not described in the available sources. There is no KEV entry and no public exploit identified at time of analysis, and the practical exposure is bounded by the plugin's limited deployment footprint and the medium-severity, integrity-only rating.
Unauthenticated remote attackers can make limited unauthorized data modifications in Magepeople's Booking and Rental Manager for WooCommerce through version 2.8.2 because an access-control check is missing. The CVSS 5.3 rating reflects network reachability with no privileges or user interaction, but only low integrity impact and no confidentiality or availability impact. No public exploit code or CISA KEV entry was identified at time of analysis.
Authorization bypass in the Royal Elementor Addons WordPress plugin through 1.7.1069 lets remote attackers request data belonging to other users or objects by supplying an attacker-chosen identifier to a plugin endpoint. The assessed impact is limited to low confidentiality disclosure (C:L, I:N, A:N), with no data modification or denial of service, and no public exploit was identified at time of analysis. The CVSS 3.1 score is 5.3; the vector indicates PR:N, though the exact vulnerable endpoint and trigger are not described in the available sources.
Missing authorization (CWE-862) in the VillaTheme 'Bopo - WooCommerce Product Bundle Builder' WordPress plugin, affecting versions up to and including 1.2.3, lets remote attackers reach an incorrectly protected plugin function without any credentials - the CVSS 3.1 vector AV:N/AC:L/PR:N/UI:N scores 5.3, with integrity-only impact (C:N/I:L/A:N). Because no privilege check is enforced on the affected code path, an unauthenticated visitor can perform an action that should require elevated permissions, though the sources do not specify which action or endpoint. No public exploit has been identified at time of analysis and the flaw is not in CISA KEV; Patchstack's report places it in the 'this month' remediation tier rather than an emergency tier, reflecting the narrow plugin footprint and the limited integrity-only consequence.
WP Discussion Board, a WordPress plugin, fails to enforce an authorization check on one of its functions in versions up to and including 2.6.0, letting remote visitors reach data or actions that should have been restricted by access-control configuration. The assessed vector (CVSS 5.3, AV:N/AC:L/PR:N/UI:N) means the flaw is exploitable over the network without credentials or user interaction against a default-exposed WordPress site, but the impact is bounded to low confidentiality exposure (C:L, I:N, A:N) rather than data modification or code execution. No public exploit code is identified at time of analysis, there is no CISA KEV listing, and the priority label is 'this_month'; the exact vulnerable endpoint, parameter or plugin configuration that triggers the missing check is not described in the available sources.
Heap-based buffer overflow in the OLE2 parser of GNOME libgsf (function ole_info_read_metabat in gsf/gsf-infile-msole.c) affects versions up to 1.14.60 and allows remote attackers to corrupt heap memory when a victim opens, previews, or otherwise processes a crafted OLE2 compound document with an application that relies on the library. The required user interaction (UI:P) prevents a fully unattended trigger, and the assessed impact is limited to low confidentiality, integrity and availability effects rather than confirmed code execution. No public exploit identified at time of analysis; an upstream fix commit has been published, and the issue fits a routine monthly patch cycle rather than an emergency response.
Reflected cross-site scripting (CWE-79) in the video library web application of the Castile-La Mancha Regional Parliament allows remote, unauthenticated attackers to execute arbitrary JavaScript in a victim's browser when the victim interacts with attacker-crafted content or a crafted URL (CVSS 4.0 base 5.1, Medium). The application embeds user-supplied content into the page without sanitising it before rendering, so the payload executes in the security context of the site; because the flaw is reflected there is no stored payload and each victim must be individually lured, which is the main limiting factor. Impact is bounded to Low confidentiality and integrity within the browser session of users of this single, narrowly deployed application. No public exploit has been identified at time of analysis and there is no CISA KEV entry; the priority label assigned by vuln.today is this_month.
Reflected cross-site scripting in the Zuinq Studio IsMyGym gym management software reservation endpoint lets remote attackers inject arbitrary JavaScript through the 'ret' parameter, which is echoed back in the HTTP response without sanitization. Rated 5.1 under CVSS 4.0 (assessed CVSS 3.1: AV:N/AC:L/PR:N/UI:R/S:C/C:L/I:L/A:N), the flaw requires the victim to actively open a crafted link, after which the script executes in that user's browser session with the privileges of that session. No public exploit code and no confirmed active exploitation were identified at time of analysis, and the affected version range is not enumerated - the CPE is a wildcard covering all IsMyGym releases; INCIBE published the originating advisory.
Stored cross-site scripting in the web interface of WolfVision Cynap collaboration devices allows an attacker who can influence a stored, externally supplied value to execute script in the browser of another user who later views the page rendering that value. The flaw affects the Cynap Videobar (before 1.72e), the Cynap Pro, Cynap Core Pro, Cynap Receiver, Cynap Pure, Cynap Pure Mini and Cynap Pure Pro family (before 1.78a), and Cynap Pure (Next Gen) and Cynap Bridge (before 2.00d), with Cynap, Cynap Core and Cynap Pure SDM also listed by the EUVD record. Per the assessed vector (CVSS:3.1/AV:N/AC:L/PR:L/UI:R/S:C/C:L/I:L/A:N) the attacker needs some level of authenticated access to write the value and the victim must be induced to load the affected page, so this is stored XSS rather than unauthenticated remote code execution; no public exploit has been identified at time of analysis and EPSS is low at 0.25% (17th percentile).
Cross-site scripting in DOMPurify versions up to and including 3.4.16 lets an attacker slip a javascript: URL past the sanitizer's SMIL guard when the integrating application parses content with PARSER_MEDIA_TYPE set to 'application/xhtml+xml' and allowlists the SVG animate (or set) element together with the camelCase attributeName attribute. In that mode the guard that strips attributeName="href" never fires, so a sanitized SVG keeps an <animate> that rewrites its parent link's href, and a victim who clicks that link executes the injected script. Exploitation requires this non-default configuration plus user interaction, which is why the vendor classifies it as a hardening bug rather than a default-config bypass; EPSS is low at 0.22% (12th percentile), there is no CISA KEV listing, and a public proof-of-concept shipped with the advisory demonstrates the behavior in Chrome, with the vendor releasing 3.4.17 as the fix.
DOMPurify 3.4.4 through 3.4.16 can be made to preserve attacker-chosen template expressions in the text content of an attached shadow root when SAFE_FOR_TEMPLATES is enabled together with a DOM-returning mode (RETURN_DOM, RETURN_DOM_FRAGMENT or IN_PLACE); because a disallowed separator element can sit between two individually harmless text fragments, the fragments become adjacent once that element is removed and reconstitute a complete `${...}`, `{{...}}` or `<%...%>` expression. The exposure is not a default-configuration flaw: the application must sanitize a DOM node that already carries a shadow root and must subsequently feed that shadow root's textContent or innerHTML into a template evaluator, so real impact depends entirely on the downstream engine's evaluation context. Severity is mid-tier and conditional - EPSS is 0.18% (7th percentile), there is no CISA KEV entry, and the vendor has released a fix in 3.4.17; a public PoC reference (a demonstration snippet in the GitHub advisory) exists but is not verified as working exploit code.
DOM-based XSS in DOMPurify (all versions before 3.4.17) survives sanitization when an application runs IN_PLACE mode with a hook that removes a parent, ancestor, or not-yet-visited sibling: the detached subtree is never walked or neutralized, so an already-loading <img onerror> fires after sanitize() returns. Exploitation is conditional on that specific non-default configuration - remote, unauthenticated, but requiring user interaction and attacker markup containing a deferred handler - and the vendor has released 3.4.17 with a fix (PR 1672 / commit a77a3e7). Risk signals are modest and this is not an emergency: EPSS is 0.16% (6th percentile), there is no CISA KEV entry, and priority is 'this_month'; the vendor advisory itself includes a public proof-of-concept that demonstrates the handler firing in Chromium.
ECDSA JWS signature malleability in erlang-jose (potatosalad:jose) 1.0.0 through 1.11.12 lets anyone holding a single legitimately issued ES256/ES256K/ES384/ES512 token produce multiple distinct signature byte strings that all verify against the same key, because verify/4 never checks the RFC 7518 §3.4 fixed signature length. It cannot forge new payloads or impersonate a signer - the underlying (r, s) pair is unchanged - but it defeats controls that assume a token or signature is unique, such as token-revocation-by-hash, replay/idempotency caches and deduplication keyed on raw signature bytes; applications that enforce uniqueness via a claim such as jti are not meaningfully affected. Exploitation requires an already-valid token as input (assessed vector AV:N/PR:L), a demonstration PoC reference ships with the vendor advisory but it is not verified as working exploit code against third-party deployments, EPSS is 0.09% (1st percentile), there is no CISA KEV listing, and the vendor released a fix in 1.11.13.