CVEs

OpenClaw versions prior to 2026.3.2 contain an archive extraction vulnerability in the tar.bz2 installer path that bypasses safety checks enforced on other archive formats. Attackers can craft malicious tar.bz2 skill archives to bypass special-entry blocking and extracted-size guardrails, causing local denial of service during skill installation.

OpenClaw versions prior to 2026.2.25 contain a time-of-check-time-of-use vulnerability in approval-bound system.run execution where the cwd parameter is validated at approval time but resolved at execution time. Attackers can retarget a symlinked cwd between approval and execution to bypass command execution restrictions and execute arbitrary commands on node hosts.

OpenClaw versions 2026.2.22 prior to 2026.2.25 contain a privilege escalation vulnerability allowing unpaired device identities to bypass operator pairing requirements and self-assign elevated operator scopes including operator.admin. Attackers with valid shared gateway authentication can present a self-signed unpaired device identity to request and obtain higher operator scopes before pairing approval is granted.

WWBN AVideo is an open source video platform. Prior to version 26.0, the `listFiles.json.php` endpoint accepts a `path` POST parameter and passes it directly to `glob()` without restricting the path to an allowed base directory. An authenticated uploader can traverse the entire server filesystem by supplying arbitrary absolute paths, enumerating `.mp4` filenames and their full absolute filesystem paths wherever they exist on the server — including locations outside the web root, such as private or premium media directories. Version 26.0 contains a patch for the issue.

WWBN AVideo is an open source video platform. Prior to version 26.0, the Scheduler plugin's `run()` function in `plugin/Scheduler/Scheduler.php` calls `url_get_contents()` with an admin-configurable `callbackURL` that is validated only by `isValidURL()` (URL format check). Unlike other AVideo endpoints that were recently patched for SSRF (GHSA-9x67-f2v7-63rw, GHSA-h39h-7cvg-q7j6), the Scheduler's callback URL is never passed through `isSSRFSafeURL()`, which blocks requests to RFC-1918 private addresses, loopback, and cloud metadata endpoints. An admin can configure a scheduled task with an internal network `callbackURL` to perform SSRF against cloud infrastructure metadata services or internal APIs not otherwise reachable from the internet. Version 26.0 contains a patch for the issue.

The Autoptimize plugin for WordPress is vulnerable to Stored Cross-Site Scripting via the lazy-loading image processing in all versions up to, and including, 3.1.14. This is due to the use of an overly permissive regular expression in the `add_lazyload` function that replaces all occurrences of `\ssrc=` in image tags without limiting to the actual attribute. This makes it possible for authenticated attackers, with Contributor-level access and above, to inject arbitrary web scripts in pages that will execute whenever a user accesses an injected page by crafting an image tag where the `src` URL contains a space followed by `src=`, causing the regex to break the HTML structure and promote text inside attribute values into executable HTML attributes.

The EmailKit – Email Customizer for WooCommerce & WP plugin for WordPress is vulnerable to arbitrary file read via path traversal in all versions up to, and including, 1.6.3. This is due to the action() function in the TemplateData class passing user-supplied input from the 'emailkit-editor-template' REST API parameter directly to file_get_contents() without any path validation, sanitization, or restriction to an allowed directory. This makes it possible for authenticated attackers, with Administrator-level access, to read arbitrary files on the server (such as /etc/passwd or wp-config.php) by supplying a traversal path. The file contents are stored as post meta and can subsequently be retrieved via the fetch-data REST API endpoint. Notably, the CheckForm class in the same plugin implements proper path validation using realpath() and directory restriction, demonstrating that the developer was aware of the risk but failed to apply the same protections to the TemplateData endpoint.

The Image Alt Text Manager plugin for WordPress is vulnerable to Stored Cross-Site Scripting via the post title in all versions up to, and including, 1.8.2. This is due to insufficient input sanitization and output escaping when dynamically generating image alt and title attributes using a DOM parser. This makes it possible for authenticated attackers, with Author-level access and above, to inject arbitrary web scripts in pages that will execute whenever a user accesses an injected page.

The RepairBuddy – Repair Shop CRM & Booking Plugin for WordPress is vulnerable to unauthorized access in all versions up to, and including, 4.1132. The plugin exposes two AJAX handlers that, when combined, allow any authenticated user to modify admin-level plugin settings. First, the wc_rb_get_fresh_nonce() function (registered via wp_ajax and wp_ajax_nopriv hooks) allows any user to generate a valid WordPress nonce for any arbitrary action name by simply providing the nonce_name parameter, with no capability checks. Second, the wc_rep_shop_settings_submission() function only verifies the nonce (wcrb_main_setting_nonce) but performs no current_user_can() capability check before updating 15+ plugin options via update_option(). This makes it possible for authenticated attackers, with subscriber-level access and above, to modify all plugin configuration settings including business name, email, logo, menu label, GDPR settings, and more by first minting a valid nonce via the wc_rb_get_fresh_nonce endpoint and then calling the settings submission handler.

The Contact List plugin for WordPress is vulnerable to Stored Cross-Site Scripting via the '_cl_map_iframe' parameter in all versions up to, and including, 3.0.18. This is due to insufficient input sanitization and output escaping when handling the Google Maps iframe custom field. The saveCustomFields() function in class-contact-list-custom-fields.php uses a regex to extract <iframe> tags from user input but does not validate or sanitize the iframe's attributes, allowing event handlers like 'onload' to be included. The extracted iframe HTML is stored via update_post_meta() and later rendered on the front-end in class-cl-public-card.php without any escaping or wp_kses filtering. This makes it possible for authenticated attackers, with Contributor-level access and above, to inject arbitrary web scripts in pages that will execute whenever a user accesses an injected page.

The Autoptimize plugin for WordPress is vulnerable to Stored Cross-Site Scripting via the 'ao_post_preload' meta value in all versions up to, and including, 3.1.14. This is due to insufficient input sanitization in the `ao_metabox_save()` function and missing output escaping when the value is rendered into a `<link>` tag in `autoptimizeImages.php`. This makes it possible for authenticated attackers, with Contributor-level access and above, to inject arbitrary web scripts in pages that will execute whenever a user accesses an injected page, granted the "Image optimization" or "Lazy-load images" setting is enabled in the plugin configuration.

The iTracker360 plugin for WordPress is vulnerable to Cross-Site Request Forgery leading to Stored Cross-Site Scripting in all versions up to and including 2.2.0. This is due to missing nonce verification on the settings form submission and insufficient input sanitization combined with missing output escaping. This makes it possible for unauthenticated attackers to inject arbitrary web scripts via a forged request granted they can trick an administrator into performing an action such as clicking on a link.

The Scoreboard for HTML5 Games Lite plugin for WordPress is vulnerable to Stored Cross-Site Scripting via the 'scoreboard' shortcode in all versions up to, and including, 1.2. The shortcode function sfhg_shortcode() allows arbitrary HTML attributes to be added to the rendered <iframe> element, with only a small blacklist of four attribute names (same_height_as, onload, onpageshow, onclick) being blocked. While the attribute names are passed through esc_html() and values through esc_attr(), this does not prevent injection of JavaScript event handler attributes like onfocus, onmouseover, onmouseenter, etc., because these attribute names and simple JavaScript payloads contain no characters that would be modified by these escaping functions. The shortcode text is stored in post_content and is only expanded to HTML at render time, after WordPress's kses filtering has already been applied to the raw post content. This makes it possible for authenticated attackers, with Contributor-level access and above, to inject arbitrary web scripts in pages that will execute whenever a user accesses an injected page.

The Keep Backup Daily plugin for WordPress is vulnerable to Stored Cross-Site Scripting via the backup title alias (`val` parameter) in the `update_kbd_bkup_alias` AJAX action in all versions up to, and including, 2.1.2. This is due to insufficient input sanitization and output escaping. While `sanitize_text_field()` strips HTML tags on save, it does not encode double quotes. The backup titles are output in HTML attribute contexts without `esc_attr()`. This makes it possible for authenticated attackers, with Administrator-level access and above, to inject arbitrary web scripts via attribute injection that will execute whenever another administrator views the backup list page.

The Injection Guard plugin for WordPress is vulnerable to Stored Cross-Site Scripting via malicious query parameter names in all versions up to and including 1.2.9. This is due to insufficient input sanitization in the sanitize_ig_data() function which only sanitizes array values but not array keys, combined with missing output escaping in the ig_settings.php template where stored parameter keys are echoed directly into HTML. When a request is made to the site, the plugin captures the query string via $_SERVER['QUERY_STRING'], applies esc_url_raw() (which preserves URL-encoded special characters like %22, %3E, %3C), then passes it to parse_str() which URL-decodes the string, resulting in decoded HTML/JavaScript in the array keys. These keys are stored via update_option('ig_requests_log') and later rendered without esc_html() or esc_attr() on the admin log page. This makes it possible for unauthenticated attackers to inject arbitrary web scripts in the admin log page that execute whenever an administrator views the Injection Guard log interface.

The Keep Backup Daily plugin for WordPress is vulnerable to Limited Path Traversal in all versions up to, and including, 2.1.1 via the `kbd_open_upload_dir` AJAX action. This is due to insufficient validation of the `kbd_path` parameter, which is only sanitized with `sanitize_text_field()` - a function that does not strip path traversal sequences. This makes it possible for authenticated attackers, with Administrator-level access and above, to list the contents of arbitrary directories on the server outside of the intended uploads directory.

Discourse is an open-source discussion platform. Prior to versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2, a non-staff user with elevated group membership could access deleted posts belonging to any user due to an overly broad authorization check on the deleted posts index endpoint. Versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2 contain a patch. No known workarounds are available.

Discourse is an open-source discussion platform. Prior to versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2, an unauthenticated attacker can cause a legitimate Discourse authorization page to display an attacker-controlled domain, facilitating social engineering attacks against users. Versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2 contain a patch. No known workarounds are available.

Service information is not encrypted when transmitted as BACnet packets over the wire, and can be sniffed, intercepted, and modified by an attacker. Valuable information such as the File Start Position and File Data can be sniffed from network traffic using Wireshark's BACnet dissector filter. The proprietary format used by WebCTRL to receive updates from the PLC can also be sniffed and reverse engineered.

WebCTRL systems that communicate over BACnet inherit the protocol's lack of network layer authentication. WebCTRL does not implement additional validation of BACnet traffic so an attacker with network access could spoof BACnet packets directed at either the WebCTRL server or associated AutomatedLogic controllers. Spoofed packets may be processed as legitimate.

Discourse is an open-source discussion platform. Prior to versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2, users with tag-editing permissions could edit and create synonyms for tags hidden in restricted tag groups, even if they lacked visibility into those tags. Versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2 contain a patch. No known workarounds are available.

Under certain conditions, an attacker could bind to the same port used by WebCTRL. This could allow the attacker to craft and send malicious packets and impersonate the WebCTRL service without requiring code injection into the WebCTRL software.

Discourse is an open-source discussion platform. Prior to versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2, unauthenticated users can determine whether a specific user is a member of a private group by observing changes in directory results when using the `exclude_groups` parameter. Versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2 contain a patch. As a workaround, disable public access to the user directory via Admin → Settings → hide user profiles from public.

Discourse is an open-source discussion platform. Prior to versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2, an attacker can grant access to a private message topic through invites even after they lose access to that PM. Versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2 contain a patch. No known workarounds are available.

flatted is a circular JSON parser. Prior to version 3.4.2, the parse() function in flatted can use attacker-controlled string values from the parsed JSON as direct array index keys, without validating that they are numeric. Since the internal input buffer is a JavaScript Array, accessing it with the key "__proto__" returns Array.prototype via the inherited getter. This object is then treated as a legitimate parsed value and assigned as a property of the output object, effectively leaking a live reference to Array.prototype to the consumer. Any code that subsequently writes to that property will pollute the global prototype. This issue has been patched in version 3.4.2.

Discourse is an open-source discussion platform. Prior to versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2, staff can modify any user's group notification level. Versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2 contain a patch. No known workarounds are available.

Charging station authentication identifiers are publicly accessible via web-based mapping platforms.

Discourse is an open-source discussion platform. Prior to versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2, the `ip_address` of a flagged user is exposed to any user who can access the review queue, including users who should not be able to see IP addresses. Versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2 contain a patch. No known workarounds are available.

Budibase is a low code platform for creating internal tools, workflows, and admin panels. In versions from 3.30.6 and prior, the REST datasource query preview endpoint (POST /api/queries/preview) makes server-side HTTP requests to any URL supplied by the user in fields.path with no validation. An authenticated admin can reach internal services that are not exposed to the internet — including cloud metadata endpoints (AWS/GCP/Azure), internal databases, Kubernetes APIs, and other pods on the internal network. On GCP this leads to OAuth2 token theft with cloud-platform scope (full GCP access). On any deployment it enables full internal network enumeration. At time of publication, there are no publicly available patches.

Nhost is an open source Firebase alternative with GraphQL. Prior to version 0.12.0, the storage service's file upload handler trusts the client-provided Content-Type header without performing server-side MIME type detection. This allows an attacker to upload files with an arbitrary MIME type, bypassing any MIME-type-based restrictions configured on storage buckets. This issue has been patched in version 0.12.0.

The WebSocket backend uses charging station identifiers to uniquely associate sessions but allows multiple endpoints to connect using the same session identifier. This implementation results in predictable session identifiers and enables session hijacking or shadowing, where the most recent connection displaces the legitimate charging station and receives backend commands intended for that station. This vulnerability may allow unauthorized users to authenticate as other users or enable a malicious actor to cause a denial-of-service condition by overwhelming the backend with valid session requests.

Discourse is an open-source discussion platform. Versions prior to 2026.3.0-latest.1, 2026.2.1, and 2026.1.2 have a potential stored XSS in topic titles for the solved posts stream. Versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2 contain a patch. As a workaround, ensure that the Content Security Policy is enabled, and has not been modified in a way which would make it more vulnerable to XSS attacks.

Ruby JSON is a JSON implementation for Ruby. From version 2.14.0 to before versions 2.15.2.1, 2.17.1.2, and 2.19.2, a format string injection vulnerability can lead to denial of service attacks or information disclosure, when the allow_duplicate_key: false parsing option is used to parse user supplied documents. This issue has been patched in versions 2.15.2.1, 2.17.1.2, and 2.19.2.

The WebSocket Application Programming Interface lacks restrictions on the number of authentication requests. This absence of rate limiting may allow an attacker to conduct denial-of-service attacks by suppressing or mis-routing legitimate charger telemetry, or conduct brute-force attacks to gain unauthorized access.

Discourse is an open-source discussion platform. Prior to versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2, moderators can create Zendesk tickets for topics they do not have access to view. This affects all forums that use the Zendesk plugin. Versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2 contain a patch. No known workarounds are available.

WebSocket endpoints lack proper authentication mechanisms, enabling attackers to perform unauthorized station impersonation and manipulate data sent to the backend. An unauthenticated attacker can connect to the OCPP WebSocket endpoint using a known or discovered charging station identifier, then issue or receive OCPP commands as a legitimate charger. Given that no authentication is required, this can lead to privilege escalation, unauthorized control of charging infrastructure, and corruption of charging network data reported to the backend.

Requires malware code to misuse the DDK kernel module IOCTL interface. Such code can use the interface in an unsupported way that allows subversion of the GPU to perform writes to arbitrary physical memory pages. The product utilises a shared resource in a concurrent manner but does not attempt to synchronise access to the resource.

Discourse is an open-source discussion platform. Prior to versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2, an authorization bypass vulnerability in hidden Solved topics may allow unauthorized users to accept or unaccept solutions. Versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2 contain a patch. As a workaround, ensure only trusted users are part of the Site Setting for accept_all_solutions_allowed_groups.

barebox is a bootloader. In barebox from version 2016.03.0 to before version 2026.03.1 (and the corresponding backport to 2025.09.3), an attacker could exploit a FIT signature verification vulnerability to trick the bootloader into booting different images than those that were verified as part of a signed configuration. mkimage(1) sets the hashed-nodes property of the FIT signature node to list which nodes of the FIT were hashed as part of the signing process as these will need to be verified later on by the bootloader. However, hashed-nodes itself is not part of the hash and could therefore be modified to allow booting different images than those that have been verified. This issue has been patched in barebox versions 2026.03.1 and backported to 2025.09.3.

A web page that contains unusual GPU shader code is loaded into the GPU compiler process and can trigger a write out-of-bounds write crash in the GPU shader compiler library. On certain platforms, when the compiler process has system privileges this could enable further exploits on the device. An edge case using a very large value in switch statements in GPU shader code can cause a segmentation fault in the GPU shader compiler due to an out-of-bounds write access.

Charging station authentication identifiers are publicly accessible via web-based mapping platforms.

NLTK (Natural Language Toolkit) is a suite of open source Python modules, data sets, and tutorials supporting research and development in Natural Language Processing. In versions 3.9.3 and prior, the NLTK downloader does not validate the `subdir` and `id` attributes when processing remote XML index files. Attackers can control a remote XML index server to provide malicious values containing path traversal sequences (such as `../`), which can lead to arbitrary directory creation, arbitrary file creation, and arbitrary file overwrite. Commit 89fe2ec2c6bae6e2e7a46dad65cc34231976ed8a patches the issue.

The WebSocket backend uses charging station identifiers to uniquely associate sessions but allows multiple endpoints to connect using the same session identifier. This implementation results in predictable session identifiers and enables session hijacking or shadowing, where the most recent connection displaces the legitimate charging station and receives backend commands intended for that station. This vulnerability may allow unauthorized users to authenticate as other users or enable a malicious actor to cause a denial-of-service condition by overwhelming the backend with valid session requests.

NLTK (Natural Language Toolkit) is a suite of open source Python modules, data sets, and tutorials supporting research and development in Natural Language Processing. In versions 3.9.3 and prior, `nltk.app.wordnet_app` allows unauthenticated remote shutdown of the local WordNet Browser HTTP server when it is started in its default mode. A simple `GET /SHUTDOWN%20THE%20SERVER` request causes the process to terminate immediately via `os._exit(0)`, resulting in a denial of service. Commit bbaae83db86a0f49e00f5b0db44a7254c268de9b patches the issue.

The WebSocket Application Programming Interface lacks restrictions on the number of authentication requests. This absence of rate limiting may allow an attacker to conduct denial-of-service attacks by suppressing or mis-routing legitimate charger telemetry, or conduct brute-force attacks to gain unauthorized access.

NLTK (Natural Language Toolkit) is a suite of open source Python modules, data sets, and tutorials supporting research and development in Natural Language Processing. In versions 3.9.3 and prior, `nltk.app.wordnet_app` contains a reflected cross-site scripting issue in the `lookup_...` route. A crafted `lookup_<payload>` URL can inject arbitrary HTML/JavaScript into the response page because attacker-controlled `word` data is reflected into HTML without escaping. This impacts users running the local WordNet Browser server and can lead to script execution in the browser origin of that application. Commit 1c3f799607eeb088cab2491dcf806ae83c29ad8f fixes the issue.

WebSocket endpoints lack proper authentication mechanisms, enabling attackers to perform unauthorized station impersonation and manipulate data sent to the backend. An unauthenticated attacker can connect to the OCPP WebSocket endpoint using a known or discovered charging station identifier, then issue or receive OCPP commands as a legitimate charger. Given that no authentication is required, this can lead to privilege escalation, unauthorized control of charging infrastructure, and corruption of charging network data reported to the backend.

Halloy is an IRC application written in Rust. In versions on \*nix and macOS prior to commit f180e41061db393acf65bc99f5c5e7397586d9cb, halloy creates its config directory and files using default umask permissions, which typically results in `0644` on files and `0755` on directories. This allows any local user on the system to read plaintext credentials stored in `config.toml` or referenced `password_file` paths. Commit f180e41061db393acf65bc99f5c5e7397586d9cb patches the issue.

Avo is a framework to create admin panels for Ruby on Rails apps. Prior to version 3.30.3, a reflected cross-site scripting (XSS) vulnerability exists in the return_to query parameter used in the avo interface. An attacker can craft a malicious URL that injects arbitrary JavaScript, which is executed when he clicks a dynamically generated navigation button. This issue has been patched in version 3.30.3.

Halloy is an IRC application written in Rust. Prior to commit 0f77b2cfc5f822517a256ea5a4b94bad8bfe38b6, the DCC receive flow did not sanitize filenames from incoming `DCC SEND` requests. A remote IRC user could send a filename with path traversal sequences like `../../.ssh/authorized_keys` and the file would be written outside the user's configured `save_directory`. With auto-accept enabled this required zero interaction from the victim. Starting with commit 0f77b2cfc5f822517a256ea5a4b94bad8bfe38b6, all identified code paths sanitize filenames through a shared `sanitize_filename` function.