The internet is broken right now.
45 critical CVEs without a fix in the last 48 hours. Developers are crying quietly into their IDEs. 54 more already have a fix that someone still has to apply.
-
CVE-2026-119545
Forgejo through 15.0.11, 16.0.0 through 16.0.7, and 17.0.0 through 17.0.1 contains a server-side request forgery vulnerability that allows unauthenticated attackers to make the server issue requests by submitting arbitrary OpenID identifier URLs. Attackers can POST crafted identifiers to /user/login/openid to trigger blind discovery GET requests to loopback, cloud metadata, and internal addresses for port scanning and probing.
-
CVE-2026-61576
Lightdash is an open source business intelligence platform. In affected versions through 0.3213.0, the dbt Cloud webhook `POST /api/v1/projects/{projectUuid}/dbt-cloud/webhook` has no authentication middleware or signature validation. In `packages/backend/src/controllers/projectController.ts`, `testWebhook()` accepts `accountId` and `runId` from a `job.run.completed` request and passes them to `createPreviewWithExplores()` in `packages/backend/src/services/ProjectService/ProjectService.ts`. These unsanitized values are interpolated into the server-side dbt Cloud API URL, which is requested with the victim project's API key in the `Authorization` header. Path traversal through `accountId` changes the requested path on `cloud.getdbt.com` without changing its host, so the credential-bearing request reaches an attacker-influenced path on that fixed host. The advisory claims API-key theft, but the supplied evidence does not establish that the attacker can observe the `Authorization` header. The unauthenticated webhook also allows preview creation. The target must be a `dbt Cloud IDE` project and the attacker must know its `projectUuid`; the UUID is not enumerable but can be discovered in shared or multi-tenant contexts. Version 0.3214.0 fixes identifier path traversal and verifies webhook HMAC signatures only when a secret is configured; without a secret, unsigned webhooks continue to be processed.
-
CVE-2026-112576
A memory integrity issue was addressed with improved handling of copy-on-write pages. This issue is fixed in iOS 26.7.2 and iPadOS 26.7.2, iOS 27.2 and iPadOS 27.2, macOS Golden Gate 27.2, macOS Sequoia 15.8.2, macOS Tahoe 26.7.2, tvOS 27.2, visionOS 27.2, watchOS 27.2. An app may be able to cause unexpected changes to copy-on-write memory.
-
CVE-2026-33521
There is s denial of service vulnerability in Esri ArcGIS Server versions 12.0 and prior that may allow a remote, unauthenticated attacker to supply crafted requests that may, under exceptional circumstances, consume all available worker threads and deny service to users. A patch is provided for ArcGIS Server software in General Availability and Extended support status, which includes versions 12.0, 11.5, 11.3, and 11.1. Other versions are in mature or retired status, which are not patch eligible. Users working with unlisted versions are advised to upgrade and apply the patch.
-
CVE-2026-116213
Webasyst Framework through 4.4.0 contains an arbitrary file read vulnerability in waArchiveTar::writeItem() that creates symlinks from theme archives without validating link targets. Authenticated users with theme management rights can upload themes containing symlinks to wa-config/db.php or /etc/passwd and read them through the public theme path.
-
CVE-2026-115506
libexpat through 2.9.0 contains a denial of service vulnerability because billion laughs protection in accountingDiffTolerated() in xmlparse.c detects entity amplification too late. Attackers can supply a moderately sized XML document with a large DTD and nested entity expansions to consume tens of seconds of CPU per parse.
-
CVE-2026-95791
A vulnerability in the web management and file transfer component of AOS web could allow an authenticated remote attacker to perform arbitrary file reads. Successful exploitation could allow an attacker to access sensitive files on the underlying system, potentially exposing configuration data or credentials.
-
CVE-2026-130382
AdminNeo before 5.9.1 contains an authentication bypass vulnerability in OtpLoginPlugin that allows attackers knowing the database password to skip TOTP verification by abusing the permanent-login cookie. Attackers can submit a login with a wrong OTP and permanent=1 to obtain a neo_permanent cookie, then replay it to gain full authenticated database access.
-
CVE-2026-121853
The AppPresser WordPress plugin before 5.0.0 does not restrict access to two of its REST API endpoints, allowing unauthenticated visitors to retrieve the installed versions of AppPresser, its add-ons, several third-party AppPresser WordPress plugin before 5.0.0 and its .
-
CVE-2026-91879
Dell Terraform for ObjectScale and Powerstore, versions prior to 1.1.1 (ObjectScale) and versions 1.2.1 through 1.3.0 (PowerStore) contains an insertion of sensitive information into Log File vulnerability in the client initialization component. A low privileged attacker with remote access could potentially exploit this vulnerability, leading to information exposure and unauthorized access.
-
CVE-2026-96137
Exposure of sensitive system information to an unauthorized control sphere in Windows Graphics Kernel allows an authorized attacker to disclose information locally.
-
CVE-2026-96058
Out-of-bounds read in Windows Kernel allows an unauthorized attacker to disclose information with a physical attack.
-
CVE-2026-108402
Missing release of memory after effective lifetime in Windows Layer-2 Bridge Network Driver allows an unauthorized attacker to deny service over an adjacent network.
-
CVE-2026-120409
The All-in-One Video Gallery WordPress plugin before 5.1.1 does not check whether a video is password protected before serving it through its standalone player page and its download link, allowing unauthenticated visitors to obtain the media source of password-protected videos, such as unlisted YouTube or Vimeo links and external file URLs.
-
CVE-2026-17190
IBM API Connect 10.0.8.0 through 10.0.8.12 and 12.1.0.0 through 12.1.1.2 could allow a remote authenticated attacker to execute arbitrary JavaScript due to cross-site scripting. HTML content supplied by a provider is not sufficiently sanitized before it is displayed, which could allow script content to run in the user's browser.
-
CVE-2026-16709
IBM API Connect 10.0.8.0 through 10.0.8.12 and 12.1.0.0 through 12.1.1.2 could allow a remote authenticated attacker to execute arbitrary JavaScript in the victim's browser due to cross-site scripting. User-supplied input is not properly neutralized when a web page is updated, so script content could run in another user's session.
-
CVE-2019-25779
WooThemes Diarise through 1.5.9 contains a file read vulnerability that allows unauthenticated attackers to read arbitrary files through the calendar parameter of download.php. Attackers can send GET requests with file:// scheme paths to the calendar parameter to disclose sensitive files like /etc/passwd. Exploitation evidence was first observed by the Shadowserver Foundation on 2023-10-19.
-
CVE-2026-44929
Server-Side Request Forgery (SSRF) vulnerability in Apache XML Graphics Batik. This issue affects Apache XML Graphics Batik: through 1.19. Users are recommended to upgrade to version 1.20, which fixes the issue.
-
CVE-2026-90993
XXE vulnerability in Apache XML Graphics Batik. This issue affects Apache XML Graphics Batik: through 1.19. Users are recommended to upgrade to version 1.20, which fixes the issue.
-
CVE-2026-125327
pgEdge RAG Server is an API server for retrieval-augmented generation over PostgreSQL content using pgvector. Prior to 2.0.0, the unauthenticated query endpoint uses FetchDocuments in internal/database/search.go to read the entire filtered corpus without a LIMIT and rebuild an in-memory BM25 index on every request. Because HybridEnabled is set by default in internal/config/loader.go and the endpoint has no rate limit, inexpensive remote queries can impose database and CPU work proportional to corpus size and exhaust service resources. This issue is fixed in version 2.0.0.
-
CVE-2026-105580
Heap-based buffer overflow in Windows Management Instrumentation allows an unauthorized attacker to execute code locally.
-
CVE-2026-120013
The agent_memory module accepts a caller-supplied tenant parameter in create_client(tenant=...). Prior to version 0.50.0, the tenant parameter was not validated before being used to construct filesystem paths for secret resolution. A value containing path traversal sequences (for example, attacker-corp/../victim-corp) could cause the resolved path to escape the intended directory and access another tenant's secrets under:Plain Text/etc/secrets/appfnd/hana-agent-memory/<tenant>/Successful exploitation could allow unauthorized access to another tenant's mounted credentials, including OAuth client IDs and client secrets. This vulnerability could enable cross-tenant access to sensitive credentials.
-
CVE-2026-123371
compose-lint is a Docker Compose security linter that identifies dangerous misconfigurations before deployment. From 0.2.0 until 0.32.0, compose-lint automatically follows a .compose-lint.yml policy-file symbolic link outside the project and includes part of the target line when that external file fails YAML parsing. A pull request can point the policy file at a runner credential file, causing about 30 characters around an attacker-uncontrolled parse position to appear in standard error, JSON errors, and SARIF tool execution notifications uploaded by the GitHub Action. The run exits with an error and the attacker cannot select the exact disclosed line, but public logs or Code Scanning output can reveal part of the target. This issue is fixed in version 0.32.0.
-
CVE-2026-76185
Apache Airflow's Task SDK secret cache built its keys from the team name and secret id with unescaped `_` delimiters, so different (team, secret id) pairs could collide. In multi-team deployments a Dag author could read another team's Variable or Connection value by crafting a colliding secret id. Only deployments with both `core.multi_team = True` and `secrets.use_cache = True` are affected. Users are advised to upgrade to apache-airflow 3.4.0 or later. Deployments that cannot upgrade can set `secrets.use_cache = False`.
-
CVE-2026-124326
A flaw has been found in DataLinkDC Dinky up to 1.2.5. Impacted is the function file.getOriginalFilename of the file CatalogueController.java of the component Catalogue Upload. Executing a manipulation of the argument filename can lead to path traversal. The attack can be executed remotely. The exploit has been published and may be used. The researcher claims to have submitted a GHSA on the repository of the project.
-
CVE-2026-104738
The Advanced Form Integration — Connect Forms to 300+ Apps plugin for WordPress is vulnerable to PHP Object Injection in all versions up to, and including, 2.10.2 via deserialization of untrusted input . This makes it possible for unauthenticated attackers to inject a PHP Object. No known POP chain is present in the vulnerable software, which means this vulnerability has no impact unless another plugin or theme containing a POP chain is installed on the site. If a POP chain is present via an additional plugin or theme installed on the target system, it may allow the attacker to perform actions like delete arbitrary files, retrieve sensitive data, or execute code depending on the POP chain present. Exploitation requires that an administrator has previously configured at least one active NexForms integration (status=1) in the plugin's database table; in the free tier, only POST parameter keys containing the substrings 'name', 'email', 'first', or 'last' are routed to the vulnerable normalizer.
-
CVE-2026-116784
FreeSWITCH is a Software Defined Telecom Stack enabling the digital transformation from proprietary telecom switches to a software implementation that runs on any commodity hardware. Prior to 1.11.2, process_rtcp_report() used an RTPFB NACK packet's declared RTCP header length as its feedback-entry loop bound instead of the datagram length already validated by the caller. A short packet with an oversized declaration could drive reads far beyond the fixed 16 KiB receive buffer. Any peer able to send RTCP to an active video call can reach the flaw, including calls where remote SDP enables video RTCP even though the SIP profile has no explicit RTCP interval configured. The out-of-bounds read can crash FreeSWITCH and terminate every active call, but supplies no write primitive and does not echo the read values to the sender. This issue is fixed in version 1.11.2.
-
CVE-2026-123200
VRPN through 07.36 contains an out-of-bounds stack write vulnerability in vrpn_Endpoint::handle_sender_message() and handle_type_message() that allows unauthenticated remote peers to write outside a stack buffer via an unchecked length index. Attackers can connect, complete the version-cookie exchange, and send a SENDER_DESCRIPTION or TYPE_DESCRIPTION message with an out-of-range or negative index to crash the process.
-
CVE-2026-122054
Converting crafted BIG-HKSCS input to UCS-4 or the internal wide character encoding, for example with iconv, in the GNU C Library version 2.18 through 2.31 may result in the converter making no progress, causing the calling application to hang.
-
CVE-2026-118516
The Symfony Routing components CompiledUrlMatcher and UrlMatcher interpret the host and path information differently, allowing path-based authorization rules to be bypassed in Symfony versions prior to 5.4.56, 6.x prior to 6.4.50, 7.x prior to 7.4.23, and 8.x prior to 8.1.11. For routes that define a host, CompiledUrlMatcher matches the string $host.'.'.$pathinfo against the host regular expression followed by the path regular expression, while UrlMatcher matches the host and the path separately. A host placeholder can therefore match /, and consume the beginning of the path, so the request is routed to a controller, while UrlMatcher would find no route. Rules that look at the path, such as access_control entries and firewall patterns, are evaluated against the actual request path, so they do not apply while the controller of the route runs, potentially allowing an unauthenticated request to reach a protected controller. For example, with the route /admin/report/{id}, the host app.{domain} and the rule access_control: [{ path: ^/admin, roles: ROLE_ADMIN }], an anonymous request for /admin/report/1 is sent to the login page. A request with the Host header app and path /x./admin/report/1 instead reaches the controller, with domain="/x". With the required domain: .+, the real hostname works: app.example.com and /x./admin/report/1 match with domain="example.com./x". Applications are affected if they have routes with host placeholders protected by path-based rules. With the default requirement, the Host header must be the host of the route without its last labels, such as app: framework.trusted_hosts blocks that, while a web server that only serves known names may not, for instance, Nginx passing the raw Host header to FastCGI for a request line that carries an absolute URI. If the requirement accepts /, such as .+, a real hostname works, and trusted_hosts does not help.
-
CVE-2026-120422
Sogou Input Method for Windows contains a vulnerability in the configuration retrieval mechanism of its SGTool.exe component. The application retrieves configuration data over unencrypted HTTP connections, allowing an attacker capable of manipulating network traffic to replace legitimate server responses with malicious configurations. An attacker can exploit this vulnerability to instruct SGTool.exe to download attacker-controlled files to arbitrary locations, including paths used by legitimate application components. This can result in arbitrary code execution when the downloaded files are subsequently executed or loaded by the application.
-
CVE-2026-95738
Authentication bypass vulnerabilities in the affected interface of HPE Networking Fabric Composer could allow an unauthenticated remote attacker to circumvent existing authentication controls if certain preconditions outside of the attacker's control are met. Successful exploitation could allow an attacker to gain administrative access to the affected system.
-
CVE-2026-121436
Froxlor from 2.3.8 before 2.3.15 contains an information disclosure vulnerability that allows administrators with customers_see_all to view every account's API key and secret through an unscoped WHERE 1 query in api_keys.php. Attackers can read the super administrator's apikey and secret from the detail modal and replay them to the API to gain super administrator control.
-
CVE-2026-118090
Subscriber PHP Object Injection in MasterStudy LMS <= 3.7.55 versions.
-
CVE-2026-116868
Subscriber PHP Object Injection in BuddyPress Xprofile Custom Field Types <= 1.3.1 versions.
-
CVE-2026-115482
In lsquic before 4.9.0, the HTTP/3 implementation can use an excessive amount of memory because the number of buffered header sets is not limited, as demonstrated by a QPACK amplification attack. This is related to header_out_dynamic_entry and es_max_header_list_size.
-
CVE-2026-78846
Multiple AiDot devices are vulnerable to a heap-based buffer overflow via SSDP location header parsing.
-
CVE-2026-129281
N-able N-central before 2026.4.1 HF3 checks only a view permission for ServerII2 operations that write device execution configuration. An authenticated user with view-only access to a device can set configuration that the Windows agent on that device acts on.
-
CVE-2026-113449
visualTicket is vulnerable to information exposure through JSONP on session-bound cart service endpoints. Session-bound data (including the session id, cart id, item ids, ticket/event metadata, quantities and prices) is returned inside a script wrapper, and since the symfony session token can also be passed in the URL, a remote attacker who obtains the token can read the cart session token and cart data. This issue was mitigated in version 20260817: the session cookie is set with SameSite=Lax, which prevents other websites from reading cart data of the victim’s session.
-
CVE-2026-126888
Buffer overflow in Turbolev in Google Chrome prior to 157.0.8103.78 allowed a remote attacker to execute arbitrary code inside the sandbox via a crafted HTML page. (Chromium security severity: High)
-
CVE-2026-76079
Probo is a self-hostable governance, risk, and compliance (GRC) platform built for engineering and security teams. Prior to version 0.224.0, an organization ADMIN can mint an OWNER membership via createUser, bypassing the owner-only set-owner authorization gate. This issue has been patched in version 0.224.0.
-
CVE-2026-92443
The caching plugins in Kong Gateway Enterprise and Kong AI Gateway may serve a response cached for a request carrying an Authorization or Proxy-Authorization to different callers.
-
CVE-2026-92442
The graphql-proxy-cache-advanced plugin in Kong Gateway Enterprise omits the caller's identity from the cache key, allowing a response cached for one caller to be served to another.
-
CVE-2026-123041
The request-callout plugin in Kong Gateway Enterprise and Kong AI Gateway omits callout headers not sent by the client from its cache key, allowing a cached callout response for a caller to be served to another.
-
CVE-2026-119084
Unsafe reflection in the connection parameter and credentials profile handling of Amazon Redshift JDBC Driver before 2.3.0 might allow an actor who can influence the JDBC connection URL or the AWS credentials profile to execute arbitrary code in the application context, provided a suitable class is available on the application classpath. To remediate this issue, users should upgrade to version 2.3.0 or later.
-
CVE-2026-82159
Dell Wyse Management Suite, versions prior to 2609.0.0.155, contain an Improper Neutralization of Input During Web Page Generation ('Cross-site Scripting') vulnerability. A low privileged attacker with remote access could potentially exploit this vulnerability, leading to Script injection.
-
CVE-2026-118994
CallBackery versions before 0.62.0 for Perl allow log injection via newlines in the login name. login() writes the login name to the application log without escaping control characters. The login RPC method needs no authentication, so a name containing a newline followed by text in the log record format adds a forged record. An unauthenticated attacker can forge log records, such as a successful administrator login from an address of their choosing.
-
CVE-2026-10757
There is a SQL injection vulnerability in Esri ArcGIS Server versions 12.1 and earlier on Windows and Linux that may allow a remote attacker with privileges to access the service (which could, depending on the service being configured in a non-default access configuration “public”, include an anonymous attacker) to issue database specific functions that should not be allowed. A patch is provided for ArcGIS Server software in General Availability and Extended support status, which includes versions 12.1, 12.0, 11.5, 11.3, and 11.1. Other versions are in mature or retired status, which are not patch eligible. Users working with unlisted versions are advised to upgrade and apply the patch.
-
CVE-2026-99083
In the Linux kernel, the following vulnerability has been resolved: ksmbd: fix heap out-of-bounds write in krb5_authenticate() In krb5_authenticate(), out_len is calculated to determine available headroom in response_buf for the incoming Kerberos AP-REP token. However, when SMB2_SESSION_SETUP is processed as a non-first element of a compounded SMB2 request, the calculation omits work->next_smb2_rsp_hdr_off. This causes out_len to overstate remaining buffer headroom by the cumulative size of prior responses in the compound chain. Consequently, the length check in ksmbd_krb5_authenticate() (*out_len <= resp->spnego_blob_len) passes erroneously, allowing memcpy() to write the AP-REP blob past the end of response_buf into adjacent kernel heap memory. Fix this by subtracting work->next_smb2_rsp_hdr_off when computing out_len, ensuring it accurately reflects physical remaining buffer space.
-
CVE-2026-125283
When NGINX Plus or NGINX Open Source is configured to use the try_files and alias directives with a map variable, certain combinations of regular expression captures in map and try_files can cause a segmentation fault in an NGINX worker process. Impact: An attacker could leverage this vulnerability to cause an NGINX worker process to terminate unexpectedly. There is no control plane exposure; this is a data plane issue only. Note: Software versions which have reached End of Technical Support (EoTS) are not evaluated.