The internet is broken right now.
45 critical CVEs without a fix in the last 48 hours. Somewhere a staging server is public. It always is. 46 more already have a fix that someone still has to apply.
-
CVE-2026-55506
Wavelog is web-based amateur radio logging software. Prior to 2.5, the filename parameter accepted by POST /qsl/addQsoToQsl in application/controllers/Qsl.php addQsoToQsl() is stored without rejecting directory separators or relative path components, and POST /qsl/delete reaches application/models/Qsl_model.php deleteQsl(), which passes the planted qsl_images filename to unlink(). A standard authenticated user can associate the malicious filename with the user's own QSO, satisfy the ownership check, escape the per-user QSL image directory, and delete files writable by the PHP worker. Deleting configuration or operational files can cause denial of service, while deleting installer gate files can reopen the unauthenticated installer and enable application identity takeover through attacker-controlled database configuration. This issue is fixed in version 2.5.
-
CVE-2026-130765
Obot before 0.27.0 contains an open redirect vulnerability that allows unauthenticated attackers to bypass same-site validation of the rd parameter by inserting a tab character. Attackers can craft sign-in or sign-out links on the Obot domain that browsers resolve to external hosts, redirecting victims to phishing pages.
-
CVE-2026-130333
The SNORDIAN's H5PxAPIkatchu plugin for WordPress is vulnerable to Stored Cross-Site Scripting via the 'xapi' parameter in all versions up to, and including, 0.4.24 due to insufficient input sanitization and output escaping. This makes it possible for unauthenticated attackers to inject arbitrary web scripts in pages that will execute whenever a user accesses an injected page. The nonce required to reach the unauthenticated insert endpoint is publicly exposed in the listener.js script URL on every page displaying H5P content, making the injection trivially exploitable by any anonymous visitor.
-
CVE-2026-125033
The Smash Balloon Social Post Feed – Simple Social Feeds for WordPress plugin for WordPress is vulnerable to Stored Cross-Site Scripting via the 'attachments.data[].description' parameter in all versions up to, and including, 4.15.1 due to insufficient input sanitization and output escaping. This makes it possible for unauthenticated attackers to inject arbitrary web scripts in pages that will execute whenever a user accesses an injected page. This makes it possible for unauthenticated outside Facebook users — any party able to contribute an attachment description to a post ingested by the configured feed (visitor posts, shared posts, or attached content) — to plant payloads that execute against every anonymous visitor of any WordPress page carrying the shortcode, requiring no WordPress account whatsoever.
-
CVE-2026-130732
Traefik from 2.11.58 before 2.11.61 contains a resource leak vulnerability that allows unauthenticated attackers to exhaust resources by sending Authorization headers with NTLM or Negotiate schemes. Attackers can open many short-lived connections that each allocate a dedicated backend transport never released, exhausting memory, file descriptors and backend connections.
-
CVE-2026-126277
NocoDB before 2026.08.1 contains an uncontrolled resource consumption vulnerability that allows unauthenticated attackers to exhaust server resources by sending unbounded filter arrays to shared-view bulk endpoints. Attackers with a shared view UUID can POST a large body of minimal filter objects, triggering a full database query per entry, to crash the NocoDB process.
-
CVE-2026-126364
Astro is a web framework for content-driven websites. From 5.14.3 until 7.4.7, Astro's /_image endpoint compared the requested local-image path with the build output directory without preserving case on case-sensitive filesystems. A sibling directory whose path differed from the output directory only by letter case could therefore pass the containment check, allowing an unauthenticated requester to read image files accepted by the image service from outside the output directory. Exploitation required on-demand rendering with the @astrojs/node adapter, a case-sensitive filesystem, and a preexisting case-variant sibling directory; ordinary traversal to unrelated paths was already rejected. This issue is fixed in Astro version 7.4.7.
-
CVE-2026-126210
@hono/node-server allows running the Hono application on Node.js. Prior to 1.19.18 and from 2.0.0 until 2.1.6, the server passes a raw request-target containing a backslash in its path to the WHATWG URL parser, which treats the backslash as a path separator and resolves dot segments. A reverse proxy, load balancer, or web application firewall that preserves the backslash and applies path-based rules to the original request can therefore permit a request that the application normalizes to a blocked route. This lets an unauthenticated client reach route content that the upstream component was intended to deny, while Hono application access controls are unaffected because they operate on the normalized path. This issue is fixed in versions 1.19.18 and 2.1.6.
-
CVE-2026-116325
SQL injection in PostgreSQL reindexdb allows an object creator to execute arbitrary SQL as the database user running reindexdb, via a crafted index name embedded in the parallel catalog query for --jobs with --index. Within major versions 17-18, minor versions before PostgreSQL 18.7 and 17.12 are affected. Versions before PostgreSQL 17 are unaffected.
-
CVE-2026-122044
Philips Hue Bridge Pro Protection Mechanism Failure Authentication Bypass Vulnerability. This vulnerability allows network-adjacent attackers to bypass authentication on affected installations of Philips Hue Bridge Pro. Authentication is not required to exploit this vulnerability. The specific flaw exists within the HTTP service, which listens on TCP port 443 by default. The issue results from the failure to apply a protection mechanism. An attacker can leverage this vulnerability to bypass authentication on the system. Was ZDI-CAN-33086.
-
CVE-2026-125122
compose-lint is a Docker Compose security linter that identifies dangerous misconfigurations before deployment. In version 0.31.0, compose-lint does not sanitize the ##[ workflow-command opener inside strings written by --format json or --format sarif to stdout. A pull request can place that opener in a service name or other Compose value, and the GitHub Actions runner recognizes it even in the middle of a serialized line when the package, pre-commit hook, or Docker image prints JSON or SARIF to the job log. Writing the output to a file and using the compose-lint GitHub Action are not affected, and the flaw does not change the lint pass or fail result. This issue is fixed in version 0.32.0.
-
CVE-2026-123563
Devalue before 5.9.3 contains a validation bypass vulnerability that allows custom ArrayBuffer revivers to skip typed-array allocation checks in the parse function. Attackers can supply very small crafted input processed by a malformed ArrayBuffer reviver to allocate massive ArrayBuffers, causing memory exhaustion and denial of service.
-
CVE-2026-93429
The VikBooking Hotel Booking Engine & PMS plugin for WordPress is vulnerable to Stored Cross-Site Scripting via 'guests[][][documents]' Parameter in all versions up to, and including, 1.8.17 due to insufficient input sanitization and output escaping. This makes it possible for unauthenticated attackers to inject arbitrary web scripts in pages that will execute whenever a user accesses an injected page. This requires the attacker to possess a valid booking sid/ts pair, which can be obtained by making a standard public booking, as the session token needed to satisfy the JSession::checkToken() gate is emitted in the publicly accessible pre-checkin form.
-
CVE-2026-109230
The WP Visitor Statistics (Real Time Traffic) plugin for WordPress is vulnerable to Stored Cross-Site Scripting via the 'browser' parameter in all versions up to, and including, 8.7.1 due to insufficient input sanitization and output escaping. This makes it possible for unauthenticated attackers to inject arbitrary web scripts in pages that will execute whenever a user accesses an injected page. The payload is stored via an unauthenticated HTTP request to the public visitor tracking endpoint (wmcAction=wmcTrack) and executes when the bundled jqPlot PieRenderer inserts the HTML-decoded series label directly as HTML in the authenticated administrator's browser.
-
CVE-2026-115548
The Quttera ThreatSign – Web Malware Scanner for WordPress plugin for WordPress is vulnerable to Stored Cross-Site Scripting via the 'HTTP_USER_AGENT' parameter in all versions up to, and including, 4.1.0.48 due to insufficient input sanitization and output escaping. This makes it possible for unauthenticated attackers to inject arbitrary web scripts in pages that will execute whenever an administrator accesses an injected page.
-
CVE-2026-103479
The Pubjet | پابجت plugin for WordPress is vulnerable to Privilege Escalation in all versions up to, and including, 5.4.8. This is due to the `checkToken()` AJAX handler — registered under `wp_ajax_pubjet-check-token` — performing no capability check and no nonce verification before unconditionally writing the attacker-supplied `token` GET parameter directly into the site-wide `pubjet_settings` option via `pubjet_update_setting()`, prior to any remote validation. This makes it possible for authenticated attackers, with Subscriber-level access and above, to overwrite the plugin's stored REST API credential with an attacker-chosen value, immediately elevating their effective privileges to those of an authorized Pubjet API client and enabling them to invoke otherwise-restricted endpoints — `createReportage`, `updateReportage`, `deleteReportage`, `toggleCopyright`, and `createBacklink` — to create, modify, and delete arbitrary reportage posts and manipulate plugin state that Subscribers cannot normally access. The privilege escalation is self-reinforcing: because `RestApi::isTokenValid()` authenticates solely by comparing the incoming `Authorization` header against the value returned by `pubjet_token()`, which reads from the same option the unprotected handler writes to, the attacker-planted token is trusted as legitimate on all subsequent requests.
-
CVE-2026-111490
Insecure direct object reference (IDOR) vulnerability in the MatchPoint API, which relies on the `actingasuser` HTTP header without verifying the requester's authorization for the specified user. Manipulation of sequential identifiers allows access to personal, banking, and financial data, private messages, and QR codes for physical access, as well as the ability to modify third-party accounts, change passwords, impersonate users, and transfer virtual balances.
-
CVE-2026-107130
Joomla Extension - balbooa.com - Unauthenticated deletion of submission/field files in Gridbox < 2.20.5.0
-
CVE-2026-100228
Improper Input Validation in Order Returns in Apache OFBiz 24.09.07 on all platforms allows an authenticated user without the ORDERMGR_CREATE permission to persist an arbitrary price on a return item via the createReturnItem service, bypassing the intended server-computed price ceiling. Users are recommended to upgrade to version 24.09.08, which fixes this issue.
-
CVE-2026-12356
IBM MQ 9.2.0.0 through 9.2.0.44 LTS, 9.3.0.0 through 9.3.0.42 LTS, 9.4.0.0 through 9.4.0.26 LTS, 9.4.0.0 through 9.4.5.1 CD, and 10.0.0.0 through 10.0.0.5 LTS and CD could allow an authenticated user with command queue access to execute arbitrary code by injecting malicious arguments into Ant script commands.
-
CVE-2026-116650
Tanium addressed an unauthorized code execution vulnerability in Threat Response.
-
CVE-2026-116651
Tanium addressed an unauthorized code execution vulnerability in Threat Response.
-
CVE-2026-120596
A remotely exploitable denial-of-service vulnerability exists in the TCP options parser of the Nut/Net TCP/IP stack included in Nut/OS, developed by Ethernut/Egnite. Nut/OS releases up to version 5.1 are affected. An unauthenticated attacker can send a single TCP segment containing an unknown TCP option with a length byte of zero to any open TCP port on a vulnerable device. When processed, this malformed option triggers an infinite loop in the function NutTcpInputOptions, permanently blocking the TCP processing thread and rendering all TCP-based services unresponsive until the device is rebooted.
-
CVE-2026-96894
The ERP: Complete HR, Accounting & CRM Suite Built for WooCommerce plugin for WordPress is vulnerable to Stored Cross-Site Scripting via Inbound Email Body in all versions up to, and including, 1.17.10 due to insufficient input sanitization and output escaping. This makes it possible for unauthenticated attackers to inject arbitrary web scripts in pages that will execute whenever a user accesses an injected page. The inbound email router's only gate is a regex match on the References header for a reply token whose hash is never cryptographically validated, meaning any external party who received or can mimic a CRM outbound email can satisfy the check without a WordPress account.
-
CVE-2026-87086
Tanium addressed a denial of service vulnerability in Provision.
-
CVE-2026-116877
Semantic MediaWiki from 2.5.0 before 7.3.1 contains an uncontrolled resource consumption vulnerability that allows low-privilege editors to bypass smwgMaxNumRecurringEvents by supplying a large limit parameter. Attackers can save pages using #set_recurring_event without an end date to generate thousands or millions of date subobjects, exhausting CPU, memory, and database resources.
-
CVE-2026-112640
A use-after-free issue was addressed with improved memory management. This issue is fixed in Safari 27.2, iOS 26.7.2 and iPadOS 26.7.2, iOS 27.2 and iPadOS 27.2, macOS Golden Gate 27.2, tvOS 27.2, visionOS 27.2, watchOS 27.2. Processing maliciously crafted web content may lead to memory corruption.
-
CVE-2026-115572
pypdf is a free and open-source pure-python PDF library. Prior to 6.21.2, an attacker can craft a PDF that repeatedly references the same nonexistent indirect object, causing relevant file parts such as pages to trigger repeated full-file recovery loads and long runtimes when accessed. This issue is fixed in version 6.21.2.
-
CVE-2026-93639
A missing authorization vulnerability in Zimbra Collaboration Suite (ZCS) allows a delegated administrator with no explicit administrative grants to overwrite global Zimlet configuration through ConfigureZimletRequest. The uploaded configuration can also add an attacker-controlled domain to the proxy allowlist of the default class of service. If a signed-in user in that class opens a crafted /service/proxy URL, Zimbra serves attacker-controlled HTML under its own origin. The resulting JavaScript can read mailbox data and perform authenticated actions as the victim. Exploitation requires a delegated-administrator account and access to the admin SOAP service; the browser compromise additionally requires the victim to open the crafted URL. ZCS 10.1.16 and 10.1.20 are affected.
-
CVE-2026-56840
PraisonAI is a multi-agent teams system. From version 4.5.2 to before version 4.6.59, praisonai.bots.HTTPApproval renders pending tool approval arguments directly into the approval dashboard HTML. An attacker-controlled tool argument can inject JavaScript into that page. When a human opens the approval URL to inspect the risky tool request, the script runs in the dashboard origin and can POST to the same request's /approve/{request_id}/decide endpoint, causing HTTPApproval to return approved=True. This issue has been patched in version 4.6.59.
-
CVE-2026-92450
The authentication framework used by affected Kiloview E3 NDI encoder firmware signs session JSON Web Tokens using an HMAC-SHA256 secret hardcoded as a static string within the firmware image, identical across all deployed devices. An attacker with access to the publicly distributed firmware can recover this secret and forge arbitrary session tokens, including administrative tokens, without valid credentials, enabling unauthenticated access to privileged device functions.
-
CVE-2026-92709
Rsyslog is a rocket-fast system for log processing. From versions 1.15.0 to 8.2610.0, rsyslog's omfile dynaFile facility intentionally allows templates to generate flexible output paths. This supports valid deployments that route logs into different directories or use dynamically generated file names. A deployment becomes unsafe when a dynaFile template uses a value influenced by untrusted input as a path component without explicit secure-path handling. In that configuration, the rendered path can escape the operator's intended log directory. This is not limited to the RFC5424 HOSTNAME field. Any untrusted value that reaches a dynaFile path component can create the condition. At time of publication, there are no publicly available patches.
-
CVE-2026-133254
In JetBrains IntelliJ IDEA before 2026.3, 2026.2.4 code execution was possible in untrusted projects via the Code Provenance plugin
-
CVE-2026-132144
MISP contains an authorization flaw in the user login-email change logic. When an external authentication backend is configured (LDAP, Shibboleth, ApacheSecureAuth, or CustomAuth), the login email address serves as the key by which the directory service resolves a user's identity, organization, and role on each login. The __canChangeLogin() method in UsersController only checked the MISP.disable_user_login_change setting and site-admin status; it did not account for whether an external authentication backend was active. As a result, any authenticated non-admin user (including org admins) could change their own login email to an address that matches a higher-privilege account in the external directory. On the user's next login, the external backend would resolve that email to the target directory identity, granting the attacker the organization and role associated with that identity. This constitutes an account-identity capture and privilege escalation: the attacker effectively assumes the identity of any directory account whose email they can guess or discover. The vulnerability requires an authenticated session with a valid API key or session token and a MISP instance configured with at least one external authentication backend. The impact is limited to the scope of the captured directory identity's privileges within MISP.
-
CVE-2026-71425
Outstatic is a static site CMS for Markdown and MDX content stored in GitHub repositories. Prior to 2.2.1, the authenticated GitHub REST proxy accepts an unvalidated endpoint value, constructs an upstream URL by string concatenation, and attaches the current dashboard user's GitHub OAuth token without ensuring the destination remains the exact GitHub API origin. An authenticated dashboard user can cause the server to send that user's token to a non-GitHub host, potentially exposing private repositories and other resources allowed by the token's scopes. The issue does not let an unauthenticated attacker access private data and does not directly expose another user's token without an additional same-origin execution or request-forging primitive. This issue is fixed in version 2.2.1.
-
CVE-2026-119635
The Groundhogg plugin for WordPress is vulnerable to Stored Cross-Site Scripting via the 'post_title' merge-tag of the [ghr] shortcode in versions up to, and including, 4.10 The post_title replacement is registered with 'nested' => true, so the nested branch of Replacements::do_replacement() returns the merge value through tackle_replacements() without passing it through escape_merge_tags()/escape_shortcodes(); the callback single_post() calls html_entity_decode(get_the_title(), ENT_QUOTES), converting the entity-encoded text that wp_filter_kses leaves intact on a Contributor's post title back into raw HTML, which is then emitted into the rendered post. 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.
-
CVE-2026-131090
Use after free in Geometry in Google Chrome prior to 158.0.8127.36 allowed a remote attacker to execute arbitrary code inside the sandbox via a crafted HTML page. (Chromium security severity: High)
-
CVE-2026-130951
Use after free in Plugins in Google Chrome prior to 158.0.8127.36 allowed a remote attacker to execute arbitrary code inside the sandbox via a crafted HTML page. (Chromium security severity: High)
-
CVE-2026-120835
Deserialization of Untrusted Data vulnerability in VillaTheme AFFI – Affiliate Marketing for WooCommerce affi-affiliate-marketing-for-woo allows Object Injection.This issue affects AFFI – Affiliate Marketing for WooCommerce: from n/a through 1.0.11.
-
CVE-2026-119512
Deserialization of Untrusted Data vulnerability in Tribulant Software Newsletters newsletters-lite allows Object Injection.This issue affects Newsletters: from n/a through 4.18.2.
-
CVE-2026-127085
Ryot before 10.5.6 contains an authorization bypass vulnerability that allows authenticated users to remove entries from other users' collections by supplying arbitrary collectionName and creatorUserId values. Attackers can obtain user ids from the usersList query and target fixed collection names like Watchlist to delete collection_to_entity rows belonging to other users.
-
CVE-2026-126253
A authentication bypass by primary weakness vulnerability in Fortinet FortiOS 8.0.0 through 8.0.1, FortiOS 7.6.0 through 7.6.7, FortiOS 7.4.0 through 7.4.12, FortiOS 7.2.0 through 7.2.13 may allow an unauthenticated attacker to broadcast LSA via specially crafted request
-
CVE-2026-92441
Kong Gateway Enterprise and Kong AI Gateway include decoded CR/LF characters from the request path in the X-Forwarded-Prefix header on regex Routes, allowing HTTP header injection and request smuggling to the upstream.
-
CVE-2026-125271
Incorrect normalization of operation names in the custom security policy enforcement in awslabs aws-api-mcp-server before 1.5.1 might allow an actor to execute AWS API operations across AWS services that the operator's security policy was configured to deny or require elicitation for, via calls to operations whose names contain consecutive capital letters. To remediate this issue, users should upgrade to version 1.5.1.
-
CVE-2026-119670
IBM Storage Copy Data Management 2.2.0.0 through 2.3.1.0 could allow an authenticated user to execute arbitrary commands with normal user privileges on the system due to improper validation of user supplied input.
-
CVE-2026-19121
IBM API Connect 12.1.0.0 through 12.1.1.2 could allow an authenticated attacker to gain unauthorized access due to improper authorization. Tenant context is not sufficiently verified against the authenticated user, which could allow access to another tenant's resources.
-
CVE-2026-123859
When identifier escaping is disabled through a non-default configuration, the write_pandas() data-loading functionality in the Snowflake Connector for Python and the Snowflake Connector for Python (Universal Core) did not escape DataFrame column names before including them in generated SQL, which could allow SQL injection. In both products, a crafted column name could have been used to chain additional SQL statements onto the generated statement, potentially allowing an attacker to read, modify, or delete other data accessible to the application's Snowflake role. Exploitation required an application to pass attacker controlled column names into this functionality. The fix is available in Snowflake Connector for Python version 4.8.1 and Snowflake Connector for Python (Universal Core) version 5.2.0. Users must upgrade manually.
-
CVE-2026-16725
IBM Watson Speech Services Cartridge 4.0.0 through 5.4.0 could allow a remote authenticated attacker to read arbitrary files from the local filesystem due to path traversal.
-
CVE-2026-10859
IBM MQ 9.2.0.0 through 9.2.0.44 LTS, 9.3.0.0 through 9.3.0.42 LTS, 9.4.0.0 through 9.4.0.26 LTS, 9.4.0.0 through 9.4.5.1 CD, and 10.0.0.0 through 10.0.0.5 LTS and CD could allow an authenticated attacker to execute arbitrary code on consumer applications due to insecure deserialization of ObjectMessage payloads.
-
CVE-2026-78527
Contributor PHP Object Injection in GS Testimonial Slider <= 3.3.10 versions.