Vulnerabilities exploitable today
404,880in current view
Single score combining CVSS, KEV membership and EPSS. Every CVE with its own record — timeline from publication to active exploitation.
In KEV catalog1,760
New KEV · 24H0
Exploit Today ≥ 701,693
Distribution · last window
- Critical5,629
- High21,804
- Medium20,192
- Low2,115
Filters
Window
Severity
Flags
CVECVSSEPSSKEVRExploitTitleMod.
CVE-2026-1101834.3 MED—
———PostgreSQL Anonymizer contains a vulnerability related to static masking and statistics persistence. Statistics pertaining to the original data remain visible in pg_stats until an ANALYZE operation is performed on the table, either manually or by autovacuum. This issue is fixed in PostgreSQL Anonymizer 3.3.0 and later versions.6hCVE-2026-1186157.8 HIG—
———Ultralytics torch_safe_load Command Injection Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Ultralytics. User interaction is required to exploit this vulnerability in that the target must visit a malicious page or open a malicious file.
The specific flaw exists within the torch_safe_load function. The issue results from the lack of proper validation of a user-supplied string before using it to execute a system call. An attacker can leverage this vulnerability to execute code in the context of the current user. Was ZDI-CAN-30122.7hCVE-2026-1186147.8 HIG—
———Ultralytics check_det_dataset Code Injection Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Ultralytics. User interaction is required to exploit this vulnerability in that the target must visit a malicious page or open a malicious file.
The specific flaw exists within the check_det_dataset function. The issue results from the lack of proper validation of a user-supplied string before using it to execute Python code. An attacker can leverage this vulnerability to execute code in the context of the current user. Was ZDI-CAN-30119.7hCVE-2026-116131——
———HTTP::Body versions before 1.23_01 for Perl allow memory and temporary file exhaustion via an unlimited number of form fields and multipart parts.
The urlencoded parser splits the whole body on its separators at once, creating one scalar per separator before any field is stored. This costs about 90 bytes of memory per separator byte. The multipart parser creates a temporary file for every part that carries a filename. Neither parser limits how many fields or parts a body may hold.
A remote attacker can exhaust worker memory with a body made of separators, or create thousands of temporary files with one request of small file parts.6hCVE-2026-1186827.5 HIG—
———KaringX ClashMi through 1.0.32.1801 disables TLS certificate validation by default in the HttpUtils helper, allowing attackers to intercept HTTPS traffic by presenting any certificate. Attackers operating the user's proxy node can serve a forged update manifest omitting sha256 to deliver a malicious installer, achieving code execution when the update runs.2hCVE-2026-1163907.2 HIG—
———Deserialization of Untrusted Data vulnerability in stepasyuk Filter Everything — WordPress & WooCommerce Filters filter-everything allows Object Injection.This issue affects Filter Everything — WordPress & WooCommerce Filters: from n/a through 1.9.7.9hCVE-2026-1163898.8 HIG—
———Deserialization of Untrusted Data vulnerability in crossi72 Phoenix Media Rename phoenix-media-rename allows Object Injection.This issue affects Phoenix Media Rename: from n/a through 3.13.5.9hCVE-2026-1159689.8 CRI—
———Deserialization of Untrusted Data vulnerability in Ultimate Member Ultimate Member ultimate-member allows Object Injection.This issue affects Ultimate Member: from n/a through 2.13.1.8hCVE-2026-1186116.8 MED—
———Alpine iLX-F511 Bluetooth OBEX Packet Handling Heap-based Buffer Overflow Remote Code Execution Vulnerability. This vulnerability allows physically present attackers to execute arbitrary code on affected installations of Alpine iLX-F511 devices. Authentication is not required to exploit this vulnerability.
The specific flaw exists within the handling of Bluetooth OBEX packets. The issue results from the lack of proper validation of the length of user-supplied data prior to copying it to a heap-based buffer. An attacker can leverage this vulnerability to execute code in the context of root. Was ZDI-CAN-29012.7hCVE-2026-1186108.8 HIG—
———Alpine iLX-F511 Android Auto Service Stack-based Buffer Overflow Remote Code Execution Vulnerability. This vulnerability allows network-adjacent attackers to execute arbitrary code on affected installations of Alpine iLX-F511 devices. Authentication is not required to exploit this vulnerability.
The specific flaw exists within the Android Auto service. The issue results from the lack of proper validation of the length of user-supplied data prior to copying it to a fixed-length stack-based buffer. An attacker can leverage this vulnerability to execute code in the context of root. Was ZDI-CAN-29015.7hCVE-2026-1186096.8 MED—
———Alpine iLX-F511 Android Auto Connection Handler Stack-based Buffer Overflow Remote Code Execution Vulnerability. This vulnerability allows physically present attackers to execute arbitrary code on affected installations of Alpine iLX-F511 devices. Authentication is not required to exploit this vulnerability.
The specific flaw exists within the Android Auto connection handler. The issue results from the lack of proper validation of the length of user-supplied data prior to copying it to a fixed-length stack-based buffer. An attacker can leverage this vulnerability to execute code in the context of root. Was ZDI-CAN-29016.7hCVE-2026-938518.0 HIG—
———Mattermost versions 11.9.x <= 11.9.1, 11.8.x <= 11.8.5, 11.7.x <= 11.7.10 fail to validate the target post ID in the Agents plugin's mention reminder component, which allows an authenticated user who can post in a channel visible to a system administrator to cause the administrator's browser to send an unintended API request, resulting in replacement of an installed plugin and code execution on the server, via a crafted post containing a forged mention reminder that the administrator clicks. Mattermost Advisory ID: MMSA-2026-007422hCVE-2026-1156038.5 HIG—
———Improper Neutralization of Special Elements used in an SQL Command ('SQL Injection') vulnerability in Advanced Ads GmbH Advanced Ads – Tracking advanced-ads-tracking allows Blind SQL Injection.This issue affects Advanced Ads – Tracking: from n/a before 4.0.0.9hCVE-2026-111108——
———Apache Kafka's code-generated protocol readers pre-size an array's collection from the element count declared by the sender. That count is checked only against the bytes remaining in the request buffer, which assumes about one byte per element; because most elements need several bytes, a request can declare far more elements than the buffer holds and still pass. For a keyed array the reader then allocates an ImplicitLinkedHashCollection of capacity (2 * count +1), so a request of a few tens of megabytes can force a heap allocation several times larger before any element is read. The pattern is emitted into every generated reader, so it affects all Kafka API types with an array field.
Impact:
Request parsing happens before any authorization check. On a listener that does not require authentication (for example PLAINTEXT), an unauthenticated attacker can send a few concurrent requests that each declare a very large=
array length and allocate many times their wire size in broker heap, crashing the broker with an OutOfMemoryError. The reporter demonstrated three concurrent ~90 MB ProduceRequest (v3) connections exhausting a 1 GB heap (~8x amplification). On listeners that require authentication, the attacker needs valid credentials but no specific ACL, since parsing precedes authorization..
Mitigation:
Upgrade to Apache Kafka 4.2.2, 4.3.2, or 4.4.0. Until then, lowering socket.request.max.bytes and restricting network access to broker listeners reduces, but does not remove, exposure.
Resolution:
From Apache Kafka 4.2.2, 4.3.2, and 4.4.0: generated readers cap the initial array pre-allocation at MessageUtil.MAX_PREALLOCATED_ARRAY_CAPACITY (1000) and grow it on demand, and reject a declared array length above MessageUtil.MAX_ARRAY_LENGTH (1,000,000).4hCVE-2026-111115——
———Apache Kafka's code-generated readers for flexible protocol versions read a tagged-field count declared by the sender and retain a RawTaggedField object for each unknown field. That count was bounded only by the outer request frame size (socket.request.max.bytes); it was not checked against the remaining buffer bytes or capped. Because each entry can be encoded in as little as two bytes (unknown tag 0, size 0), a modest request can declare tens of millions of tagged fields and force one object allocation per declared field. Flexible request headers (header version 2, used by flexible APIs such as ApiVersions v3) are parsed this way, so the amplification applies to the request header itself.
Impact:
The flexible request header is parsed on the network thread before authorization, request dispatch, and request-time quota accounting. A client can send a request of a few tens of megabytes whose header declares about 20 million tagged fields, forcing roughly that many RawTaggedField allocations and exhausting broker heap or inducing process-wide GC stalls, denying broker service. Because the header is parsed for pre-authentication APIs such as ApiVersions (v3), the allocation can occur before authentication; no Kafka resource ACL or quota is required.
Mitigation:
Upgrade to Apache Kafka 4.2.2, 4.3.2, or 4.4.0. Until then, lowering socket.request.max.bytes and restricting network access to broker listeners reduces, but does not remove, exposure.
Resolution:
From Apache Kafka 4.2.2, 4.3.2, and 4.4.0: generated readers bounds-check the tagged-field count for flexible-versioned messages, rejecting a negative count, a count exceeding the remaining buffer bytes, or a count exceeding a fixed maximum (10,000), before any per-field allocation.4hCVE-2026-1155907.2 HIG—
———Deserialization of Untrusted Data vulnerability in weDevs Dokan dokan-lite allows Object Injection.This issue affects Dokan: from n/a through 5.3.1.9hCVE-2026-1186856.5 MED—
———ToolJet from 3.20.2-lts before 3.20.242-lts contains an authorization bypass in SignupDisableGuard, which treats the string signup setting 'false' as truthy, so instance-level signup restrictions are not enforced. Unauthenticated attackers can send POST requests to /api/signup without an organizationId to create an account, join the default workspace, and receive a valid session.5hCVE-2026-118245——
———Highcharts JS is a JavaScript charting library based on SVG. From 1.0.0 until 4.2.2, Highcharts Dashboards uses chartConstructor as an unchecked property lookup in the Highcharts namespace and invokes the selected member with the component container and chartOptions. An attacker who controls dashboard configuration can select a public utility function instead of a supported chart constructor, causing attacker-controlled options to be copied to the container as properties or attributes. Markup and event handlers can then execute in the hosting page origin when the board renders. No user interaction beyond rendering the board is required. This issue is fixed in 4.2.2.6hCVE-2026-615754.7 MED—
———AVideo is a video-sharing Platform. Prior to 8.1.0, `view/queue.php` uses `pathinfo` on the caller-controlled `fileURI` and stores its filename through `setTitle` in `encoder_queue.title` without neutralization. `fileURI` is not in the `objects/security.php` filter lists, and the `Login::canUpload()` capability needed to submit a queue entry can be self-granted using an attacker-controlled login site without legitimate encoder credentials. `view/index.php` loads `queue.json` from `view/queue.json.php` into the `#grid` jQuery Bootgrid table; its title formatter concatenates `row.title` into raw cell HTML. An attacker can therefore store an HTML title whose event handler runs in a privileged viewer's browser. `Encoder::getAll` limits nonadministrators to their own streamer's rows, so the victim must be an administrator who sees all rows or an operator on the attacker's streamer. A second `.html()` sink in the progress panel uses the same title but additionally depends on a helper script loaded from the streamer; the Bootgrid sink is reachable from encoder code alone. The script can read session-visible data and rendered streamer credentials, act with the viewer's privileges, and execute again for each qualifying viewer until the row is removed. This issue is fixed in version 8.1.0.2hCVE-2026-105420——
———Wing FTP Server does not verify user's base folder. An attacker with admin privileges can add user directories via scripts, allowing the added directories to escape the base folder. This allows the attacker to write files on the system and can lead to remote code execution with process owner privileges.Critically, since the installation guide mentions running the command with `sudo`, the files are created as “root.”
This issue was fixed in version 8.2.6.9hCVE-2026-1186076.8 MED—
———Alpine iLX-F511 diagapp Command Injection Vulnerability. This vulnerability allows physically present attackers to execute arbitrary code on affected installations of Alpine iLX-F511 devices. Authentication is not required to exploit this vulnerability.
The specific flaw exists within the diagapp component. The issue results from the lack of proper validation of a user-supplied string before using it to execute a system call. An attacker can leverage this vulnerability to execute code in the context of root. Was ZDI-CAN-28750.7hCVE-2026-1186068.8 HIG—
———Alpine iLX-F511 Framework Service Exposed Dangerous Function Remote Code Execution Vulnerability. This vulnerability allows network-adjacent attackers to execute arbitrary code on affected installations of Alpine iLX-F511 devices. Authentication is not required to exploit this vulnerability.
The specific flaw exists within the framework service. The issue results from an exposed dangerous function. An attacker can leverage this vulnerability to execute code in the context of root. Was ZDI-CAN-29006.7hCVE-2026-935165.5 MED—
———A use-after-free caused by lack of lifecycle management when adding / removing XInput master devices in X.org X Server before 21.1.25 and XWayland before 24.1.14 could be used by authenticated X Clients to crash the X server or potentially execute code in X server context.10hCVE-2026-118256——
———Vue I18n is the internationalization plugin for Vue.js. From 9.0.0-beta.5 until 11.4.14 and 12.0.0-alpha.5, enabling escapeParameter, called escapeParameterHtml in the Legacy API, does not escape named interpolation values when the same translation call supplies both list and named values because escapeParams() processes only the list branch. Applications that render translations as HTML and pass untrusted named values with a list can insert attacker-controlled markup, including phishing links, forms, and layout changes, and can permit script execution when combined with the srcdoc sanitizer bypass or on releases without the sanitizer. This issue is fixed in versions 11.4.14 and 12.0.0-alpha.5.5hCVE-2026-111496——
———Kafka Streams derives its stream-time clock from record timestamps. PunctuationSchedule.next() computed the next punctuation time as timestamp + (intervalsMissed + 1) * interval with no overflow check. A record with a timestamp at or near Long.MAX_VALUE advances stream time to that value, so the next stream-time punctuation overflows to a negative time. PunctuationQueue then re-queues that time and loops on it indefinitely while holding a lock, stalling the StreamThread.
Impact:
A single record is enough. The affected StreamThread stops processing, and because the record stays in the topic the hang returns on restart; an application whose partitions are all affected becomes unresponsive. Only the Streams application is affected, not the brokers or the cluster. Triggering it needs write access to a source topic whose topology uses stream-time punctuation, and a broker that stores the record: brokers from Apache Kafka 4.0.0 reject such a timestamp by default (log.message.timestamp.after.max.ms, one hour), so it is reachable only on older brokers or where that bound has been relaxed.
Mitigation:
Upgrade Kafka Streams to 4.2.2, 4.3.2, or 4.4.0. As defense in depth, keep log.message.timestamp.after.max.ms bounded on source topics and restrict write access to them.
Resolution:
From Apache Kafka 4.2.2, 4.3.2, and 4.4.0: the next punctuation time is computed with overflow-checked arithmetic (clamped to Long.MAX_VALUE), and a rescheduled punctuation is re-queued only when its time moves forward.4hCVE-2026-544537.2 HIG—
———OpenCATS is an open-source applicant tracking system and recruitment CRM for staffing agencies and hiring teams. Prior to version 0.10.0, the Professional License configuration feature in `SettingsUI` (`manageProfessional`) accepts administrator-controlled license key input that is reconstructed into a PHP `define()` statement in `config.php` without safe escaping. Because the application includes `config.php`, crafted input can alter the file's PHP syntax and cause persistent arbitrary PHP execution with the web server process privileges. An authenticated administrator must have Settings module access and permission to update Professional License configuration, and `config.php` must be writable for this write-based path. Successful exploitation can read, modify or delete application data, expose candidate, contact, company, job order and user account information, alter the installation or configuration, and provide a foothold for further attacks. Unauthenticated exploitation and exploitation without Settings access are not established. This issue is fixed in version 0.10.0.2hCVE-2026-118246——
———Highcharts JS is a JavaScript charting library based on SVG. From 10.0.0 until 11.4.9, 12.6.1, and 13.1.0, Highcharts applies breadcrumbs.style as SVG attributes rather than CSS and does not filter the attribute names. An attacker who controls chart configuration can provide event-handler keys through drilldown.breadcrumbs, navigation.breadcrumbs, or hierarchical series breadcrumb options. The keys become active handlers on the breadcrumb button and execute in the hosting page origin after the user navigates to a level and hovers over or activates the button. Styled-mode charts do not apply this option and are not affected. This issue is fixed in 11.4.9, 12.6.1 and 13.1.0.6hCVE-2026-541624.7 MED—
———Ember is a real-time terminal dashboard for Caddy servers with FrankenPHP support. Prior to 1.4.2, interactive TUI row formatters pass Caddy access-log fields, particularly the request URI, through the `fitCellLeft` render helper without neutralizing terminal escape or control sequences. An unauthenticated remote client can place ANSI, OSC, or CSI control bytes in an ordinary HTTP request target. Caddy JSON-escapes those bytes, `ParseLogLine` decodes them back into raw controls, and the formatters concatenate them into `View()` output reaching `os.Stdout`. When the operator views the default `Logs` tab, the terminal interprets these sequences, allowing monitoring-dashboard spoofing, clipboard hijacking through `OSC 52`, and window-title spoofing through `OSC 0/2`. This works even with a loopback-only log listener because the input arrives through normal Caddy access logs. `Certificates` and `Routes` share secondary unneutralized render sinks. The `--json`, `--once`, and daemon `--expose` modes are unaffected because they do not render an interactive terminal. Consequences depend on the terminal emulator and usually require further operator action to escalate; this is not general-purpose remote code execution. This issue is fixed in version 1.4.2.4hCVE-2026-35833.1 LOW—
———When the `allowed_scopes` configuration is enabled, the OAuth2 token introspection mechanism does not immediately invalidate a token in the cache after a revocation request. This results in a scenario where an access token that was introspected immediately before a revocation request may be returned again to the client after revocation, until the relevant cache entry is cleared.
The security consequence is that a revoked access token may be re-issued to the client during the cache lifetime. While the re-issued token is not treated as valid for API authorization, affected users may be unable to obtain a new valid access token until the cache is cleared, which can disrupt authentication flows.4hCVE-2026-116269——
———MIME::Field::ParamVal versions before 5.521 for Perl turn an RFC 2231 filename with a hyphenated language tag into an empty recommended filename.
A parameter such as filename*=utf-8'en-US'payload.exe is converted to an encoded word with the language tag appended to the charset. The decoder strips the language suffix only when it is a single word, so it cannot resolve the charset UTF-8*EN-US and returns an empty string. The recommended_filename() method returns that empty string rather than falling back to the name parameter.
Other mail readers return payload.exe. A remote sender can get an attachment past a policy that checks file extensions on the recommended filename.6hCVE-2026-1186058.0 HIG—
———Alpine iLX-F511 prh_sdp_service_search_decode Stack-based Buffer Overflow Remote Code Execution Vulnerability. This vulnerability allows network-adjacent attackers to execute arbitrary code on affected installations of Alpine iLX-F511 devices. An attacker must first obtain the ability to pair a malicious Bluetooth device with the target system in order to exploit this vulnerability.
The specific flaw exists within the prh_sdp_service_search_decode function. The issue results from the lack of proper validation of the length of user-supplied data prior to copying it to a stack-based buffer. An attacker can leverage this vulnerability to execute code in the context of root. Was ZDI-CAN-29069.7hCVE-2026-93515——
———A use-after-free in present_clear_window_notifies() in the X.org X Server before 21.1.25 and X Wayland before 24.1.14 could be used by malicious X clients to crash or potentially execute code in the X server context.10hCVE-2026-615736.1 MED—
———AVideo is a video-sharing Platform. Prior to 8.1.0, the anonymous login interface reflects request values into HTML and JavaScript without output encoding. In `view/index.php`, `webSiteRootURL` is returned verbatim by `Streamer::getStreamerURL()` in `objects/Streamer.php` and reaches navigation and form attributes; it also enters a single-quoted JavaScript string in `view/index.header.php`. It is not filtered by `objects/security.php`, while the `user` and `pass` fields are filtered only for single quotes, leaving double quotes and angle brackets capable of breaking out of their HTML attributes. In addition, `view/index.header.php` reflects POST fields using `json_encode` without `JSON_HEX_TAG`, allowing a closing script tag to break out, and `locale/function.php` stores the POST `lang` value in the session for unescaped reflection in `view/index.php`. An unauthenticated remote attacker who persuades a victim to open a crafted login link can execute arbitrary JavaScript in the encoder's origin. This can read the operator session identifier exposed in inline script, capture login credentials and perform actions as the operator against the encoder and its upstream platform. This issue is fixed in version 8.1.0.2hCVE-2026-1186083.5 LOW—
———Alpine iLX-F511 libanwbtstack.so prh_rfcomm_cc_process_test Out-Of-Bounds Read Information Disclosure Vulnerability. This vulnerability allows network-adjacent attackers to disclose sensitive information on affected installations of Alpine iLX-F511 devices. An attacker must first obtain the ability to pair a malicious Bluetooth device with the target system in order to exploit this vulnerability.
The specific flaw exists within the prh_rfcomm_cc_process_test function. The issue results from the lack of proper validation of user-supplied data, which can result in a read past the end of an allocated buffer. An attacker can leverage this in conjunction with other vulnerabilities to execute arbitrary code in the context of root. Was ZDI-CAN-29018.7hCVE-2026-615728.6 HIG—
———AVideo is a video-sharing Platform. Before the initial remediation attempt in 8.1.0, `url_get_contents` in `objects/functions.php` accepts caller-controlled destinations after only `FILTER_VALIDATE_URL` syntax validation, without rejecting private, loopback, or link-local addresses, and disables TLS peer verification. `Login::isLogged` reauthenticates using the attacker-settable `encoder_aVideoURL` cookie together with arbitrary nonempty `encoder_user` and `encoder_pass` cookies, and `Login::run` stores the raw fetched body in `result`. The unauthenticated `objects/login.json.php` endpoint returns that session object, enabling full-response exfiltration through a supplied `siteURL`; its `Streamer::isURLAllowed` gate allows all sites when the operator allowlist is empty, the default. An unauthenticated remote attacker can read internal HTTP and HTTPS responses, including cloud instance-metadata credentials, and reach internal TCP services. A further pre-authentication FTP variant in `view/youtubeDl.json.php` routes an ftp `videoURL` into `FTPDownloader`; `parseFtpUrl` in `objects/FTPDownloader.php` accepts its host and port for `ftp_connect`. This also enables requests to internal FTP services. Untrusted self-signed HTTPS certificates are accepted, and no account or operator interaction is required. Version 8.1.0 contains an initial remediation attempt that removes raw-response reflection and adds destination checks, but private-network exemptions and login TLS-verification defaults leave complete remediation of the described internal-network and TLS variants unresolved.2hCVE-2026-441799.9 CRI—
———XWiki Remote Macros provides XWiki rendering macros that are useful when migrating content from Confluence. From 1.13 until 1.14.5, the `excerpt-include` macro does not properly escape the included page's title and executes `excerpt` content with the macro's rights. The title handling in `xwiki-pro-macros-ui/src/main/resources/Confluence/Macros/ExcerptInclude.xml` and the rendering of `excerpt` content into XWiki syntax both allow syntax injection. Any user able to edit a page, even without script or programming rights, can supply executable macro syntax in the title or `excerpt` and cause remote code execution when the page is included. This affects confidentiality, integrity and availability of the whole XWiki installation. This issue is fixed in version 1.14.5.8hCVE-2026-615717.3 HIG—
———AVideo is a video-sharing Platform. Prior to 8.1.0, the encoder lacks a global authentication layer and exposes control and status PHP endpoints to anonymous network callers. `view/run.php` has a commented-out guard and invokes `Encoder::run()`. `view/queue.json.php` calls `Encoder::getAll(true)`, reveals streamer URLs through `getSiteURL()`, and triggers processing with `execRun()`. `view/status.php` returns `Access-Control-Allow-Origin` with a wildcard, reads the queue, triggers processing, and exposes host memory statistics through `ServerMonitor::getMemory()` when `serverStatus` is set. `view/listFiles.json.php` checks only `Login::canBulkEncode()`, which admits anonymous callers whenever `disableBulkEncode` is empty, the default. Its `path` parameter is resolved with `realpath()` and listed with `scandir()` without a base-directory restriction. An unauthenticated attacker can drive the privileged queue processor, read queue and streamer-site information, obtain memory statistics and enumerate media file names and paths in any directory readable by the web-server user. Enumeration is limited to configured media extensions and does not itself disclose arbitrary file contents. This issue is fixed in version 8.1.0.2hCVE-2026-1033905.3 MED—
———An Out-of-Bounds Read vulnerability in the HttpParser component in Juniper Networks Session Smart Router and Session Smart Conductor allows an attacker to cause a denial of service via a specially crafted HTTP request.7hCVE-2026-1186137.8 HIG—
———Ultralytics torch_load Deserialization of Untrusted Data Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Ultralytics. User interaction is required to exploit this vulnerability in that the target must visit a malicious page or open a malicious file.
The specific flaw exists within the torch_load function. The issue results from the lack of proper validation of user-supplied data, which can result in deserialization of untrusted data. An attacker can leverage this vulnerability to execute code in the context of the current user. Was ZDI-CAN-29652.7hCVE-2026-1033915.3 MED—
———An Out-of-bounds Write vulnerability in the SSR TCP proxy/optimization path in Juniper Networks Session Smart Router allows an attacker to corrupt forwarded packet data and potentially cause a data-plane crash via a transit TCP packet with a crafted TCP option length.7h