🚨 CVE-2026-70639
llama.cpp builds b1886 through b7445 contain a null pointer dereference vulnerability in the LLaMA-Android JNI wrapper where the bench_1model() function fails to validate the model context pointer before dereferencing it. Attackers can supply a malicious, corrupt, or truncated model file to trigger a null context condition, causing a SIGSEGV crash that terminates the Android application process and results in denial of service.

🎖@cveNotify

🚨 CVE-2026-70640
llama.cpp builds b1886 through b7445 contain a race condition use-after-free vulnerability in the LLaMA-Android JNI wrapper where bench_1model() and free_1context() lack synchronization, allowing Thread A to operate on freed memory while Thread B concurrently frees the llama_context. Attackers can exploit this by performing heap spray with attacker-controlled data containing a fake vtable to hijack the vtable pointer at offset +0x30, causing llama_batch_allocr::clear() to dereference arbitrary memory and achieve remote code execution.

🎖@cveNotify

🚨 CVE-2026-71324
Traefik is an open source HTTP reverse proxy and load balancer. Prior to 2.11.53, 3.6.24, and 3.7.9, Traefik's default HTTP reverse proxy forwards a plain HTTP/2 or HTTP/3 CONNECT request and its body to an HTTP/1.1 upstream through a shared net/http.Transport. When the upstream answers the CONNECT with a keep-alive non-2xx response and does not drain the body, Traefik returns the desynchronized backend socket to its shared pool and reuses it for other clients. An unauthenticated attacker can use this behavior to make a different client read the attacker's smuggled response, which can include authenticated or private content from another request. The ForwardAuth middleware with forwardBody true and preserveRequestMethod true can re-issue a CONNECT with the buffered body attached, exposing the auth-client pool to the same desynchronization. This issue is fixed in 2.11.53, 3.6.24, and 3.7.9.

🎖@cveNotify

🚨 CVE-2026-71325
Traefik is an open-source edge router that makes publishing services a fun and easy experience. Prior to 2.11.54, 3.6.25, and 3.7.10, cross-namespace @kubernetescrd references are not rejected for TraefikService backend references resolved by the service resolver. A tenant confined by RBAC to a single namespace can therefore bind its own router to a TraefikService owned by another namespace and expose or reroute that namespace's backend, defeating the namespace isolation allowCrossNamespace=false is meant to enforce. This issue is fixed in version 2.11.54, 3.6.25, 3.7.10.

🎖@cveNotify

🚨 CVE-2026-71326
Traefik is an open source HTTP reverse proxy and load balancer. From 3.6.11 until 3.6.25 and 3.7.10, Traefik's BasicAuth middleware in pkg/middlewares/auth/basic_auth.go deduplicates concurrent password checks with a singleflight key built from the delimiter-free concatenation of password and secret, allowing an attacker who has a valid credential and the stored hash to authenticate as an unconfigured username when headerField trusts the forwarded identity. This issue is fixed in 3.6.25 and 3.7.10.

🎖@cveNotify

🚨 CVE-2026-71327
Traefik is an open source HTTP reverse proxy and load balancer. From 3.0.0 until 3.6.25 and 3.7.10, Traefik's Kubernetes Gateway API provider in pkg/provider/kubernetes/gateway/httproute.go, grpcroute.go, tcproute.go, and tlsroute.go builds HTTPRoute, GRPCRoute, TCPRoute, and TLSRoute router and service identities by hyphen-concatenating namespace, route name, Gateway identity, entry point, and rule index, allowing colliding Routes to overwrite another namespace's backend. This issue is fixed in 3.6.25 and 3.7.10.

🎖@cveNotify

🚨 CVE-2026-71433
LangGraph Checkpoint Postgres and SQLite Checkpoint are the Postgres and SQLite implementations of LangGraph's checkpoint saver. Prior to 3.1.1, the langgraph-checkpoint-postgres and langgraph-checkpoint-sqlite packages persisted hierarchical namespaces as a dot joined string and scoped reads by matching that string as a simple prefix pattern, so a read scoped to one namespace could also match a sibling namespace whose flattened form shares the same leading characters, or a namespace label containing unescaped pattern metacharacters, allowing an authenticated caller to retrieve stored items belonging to another tenant or user through an ordinary scoped search or list namespaces call, with no crafted input required. This issue is fixed in versions 3.1.1 of langgraph-checkpoint-postgres and langgraph-checkpoint-sqlite.

🎖@cveNotify

🚨 CVE-2026-71434
Statamic is a Laravel and Git powered content management system (CMS). Prior to 5.74.3 and 6.24.2, public frontend forms did not enforce the file upload restrictions that the Control Panel enforces, so an unauthenticated visitor could upload file types an administrator had intended to disallow through a form's assets or files field, and for assets fields, files could be stored on a public, web-accessible disk, though the application's global upload allowlist still blocked executable types such as .php and .html. This issue is fixed in versions 5.74.3 and 6.24.2.

🎖@cveNotify

🚨 CVE-2026-71435
Statamic is a Laravel and Git powered content management system (CMS). Prior to 5.74.3 and 6.24.2, the default ("automagic") form notification email rendered user-submitted values without escaping, allowing an unauthenticated form submitter to inject HTML into the notification emails sent to the configured recipients. This issue is fixed in versions 5.74.3 and 6.24.2.

🎖@cveNotify

🚨 CVE-2026-71436
Mermaid is a JavaScript tool that uses Markdown-inspired text to create and modify diagrams and charts. From version 10.6.0 until 10.9.8 and 11.16.1, Mermaid XY Charts are vulnerable to an infinite loop denial of service in the setXAxisRangeData function when configuring an X-Axis with invalid parameters. Because each loop iteration appends an element to an array, this generally causes a RangeError to appear after a few seconds, but it may instead cause the page or JavaScript process to crash from memory exhaustion, depending on the environment. This issue is fixed in versions 10.9.8 and 11.16.1.

🎖@cveNotify

🚨 CVE-2026-71437
Mermaid is a JavaScript tool that uses Markdown-inspired text to create and modify diagrams and charts. From version 11.5.0 until 11.16.1, Mermaid Architecture Diagrams are vulnerable to prototype pollution when a diagram defines a group with an id of __proto__. Because the group id is used directly as an object property key without validation, an attacker who can supply diagram text can pollute Object.prototype, potentially affecting the behavior of the embedding application. This issue is fixed in version 11.16.1.

🎖@cveNotify

🚨 CVE-2026-71438
Mermaid is a JavaScript tool that uses Markdown-inspired text to create and modify diagrams and charts. Prior to 10.9.8 and 11.16.1, Mermaid's configuration setters (mermaid.initialize, mermaidAPI.setConfig, and mermaidAPI.updateSiteConfig) merge caller-supplied configuration into Mermaid's internal config using the assignWithDepth deep-merge helper, which is vulnerable to prototype pollution. This is only exploitable if an application forwards untrusted data directly into one of these configuration entry points, which is outside their documented usage; diagram-supplied configuration (e.g. %%{init: {}}%% or YAML frontmatter) is not affected. This issue is fixed in versions 10.9.8 and 11.16.1.

🎖@cveNotify

🚨 CVE-2026-71439
Mermaid is a JavaScript tool that uses Markdown-inspired text to create and modify diagrams and charts. From version 11.6.0 until 11.16.1, Mermaid Radar Diagrams allow arbitrary large values for the ticks parameter, which can cause high CPU usage and freeze the rendering webpage or JavaScript process for long periods of time, potentially until the process is killed from memory exhaustion. This issue is fixed in version 11.16.1.

🎖@cveNotify

🚨 CVE-2026-71445
AIL Framework contained a reflected cross-site scripting vulnerability in the /tag/add_tags endpoint. When an error occurred while processing a tag operation, the application returned the error value directly as an HTML response using str(res[0]).

If attacker-controlled input was included in the generated error message, a crafted request could cause arbitrary HTML or JavaScript to be reflected in the response without appropriate output encoding. An attacker could exploit the vulnerability by convincing an authenticated AIL Framework user to open a specially crafted link.

Successful exploitation could allow JavaScript to execute in the victim’s browser within the security context of the AIL Framework application. Depending on the victim’s privileges and the application’s protections, the attacker could perform actions using the victim’s session, access information available to the victim, or modify data through authenticated application requests.

The vulnerability requires user interaction because the authenticated victim must follow or open the crafted request. The attacker does not necessarily require an AIL Framework account, provided that the crafted request can be delivered to an already authenticated user.

🎖@cveNotify

🚨 CVE-2026-71446
AIL Framework contains a stored cross-site scripting vulnerability in the crawler domain view. Crawled URLs were embedded directly into the JavaScript onclick handler used to display a stored screenshot, without context-appropriate encoding.

An attacker who can cause a specially crafted URL to be recorded in the crawler history can inject JavaScript syntax into the stored URL value. The payload remains stored by AIL and is subsequently included in the domain view. When an authenticated analyst clicks the screenshot icon associated with the malicious URL, the injected JavaScript executes in the analyst’s browser within the security context of the AIL Framework application.

Successful exploitation could allow an attacker to access information available to the analyst’s session, modify displayed content, or perform application actions using the analyst’s privileges. Exploitation requires the victim to interact with the affected screenshot entry.

The vulnerability was corrected by serializing the crawled URL with Jinja’s tojson filter before inserting it into the JavaScript handler. This safely escapes characters that could otherwise terminate the JavaScript string and introduce executable code.

🎖@cveNotify

🚨 CVE-2026-71447
AIL Project contains a stored cross-site scripting vulnerability in the translation controls displayed for chat messages and forum posts.

The affected templates inserted message and post identifiers directly into inline JavaScript onclick handlers:

onclick="translateMessageToPreferredLanguage('{{ message['id'] }}', '{{ mess_id_escape }}', this)"

and:

onclick="translatePostToPreferredLanguage('{{ post['id'] }}', '{{ post_id_escape }}', this)"

These values were HTML-template escaped but were not safely encoded for use as JavaScript string literals inside an HTML attribute. A specially crafted identifier containing quotation marks, escape characters, or other JavaScript syntax could therefore terminate the expected string argument and inject arbitrary JavaScript into the event handler.

Because the affected values are associated with indexed chat messages or forum posts, a malicious value may remain stored by AIL and be rendered whenever an analyst accesses the corresponding chat or forum explorer view.

Successful exploitation requires the victim to click the affected Translate to preferred language button. The injected code would then execute in the victim’s browser under the security origin of the AIL instance.

🎖@cveNotify

🚨 CVE-2026-71476
Nx is a monorepo solution for TypeScript and polyglot codebases. From version 20.8.0 until 22.7.7 and 23.0.2, the Nx self-hosted HTTP remote cache extracts downloaded cache artifacts without constraining where files are written. A malicious or on-path (MITM) remote cache server can return a crafted tar archive whose entries escape the cache directory and write to arbitrary locations on the machine running Nx, which can be escalated to remote code execution. Nx's default local cache and Nx Cloud are not affected; only workspaces configured to use a self-hosted remote cache are affected. This issue is fixed in versions 22.7.7 and 23.0.2.

🎖@cveNotify

🚨 CVE-2026-71478
league/commonmark is a PHP library for parsing and rendering CommonMark Markdown. From 1.5.0 until 2.9.0, the AttributesExtension's href and src unsafe-link filter can be bypassed by embedding control bytes, such as a tab, carriage return, line feed, or leading C0 control character, in a javascript: URL that browsers discard before parsing the scheme, causing the browser to still execute the script even when the unsafe-link filter is enabled. This issue is fixed in 2.9.0.

🎖@cveNotify

🚨 CVE-2026-71488
league/commonmark is a PHP library for parsing and rendering CommonMark Markdown. From 0.6.0 until 2.9.0, specially crafted Markdown lines can cause the parser to have quadratic time complexity when converting, because several parsing paths repeatedly rescan growing portions of a line to translate between character positions and byte positions, and the Autolink extension can also copy and validate the remaining line at every URL-like prefix, allowing an attacker who can submit Markdown for conversion to consume disproportionate CPU time with a comparatively small request. This issue is fixed in 2.9.0.

🎖@cveNotify

🚨 CVE-2026-71497
jsoup is a Java library for working with real-world HTML. From 1.14.3 until 1.23.1, jsoup's HTML parser could incorrectly handle a malformed tag name ending in a control character, causing the tag to acquire the parsing behavior of a different element. When a custom Safelist permits certain raw-text elements, this misparsing can cause content that should remain inert text to be emitted as active markup after serialization, potentially resulting in cross-site scripting. jsoup's built-in Safelists are not affected. This issue is fixed in version 1.23.1.

🎖@cveNotify

🚨 CVE-2026-71498
node-re2 provides RE2 regular expression bindings for Node.js. Prior to version 1.26.1, passing a Buffer whose final bytes form a truncated (incomplete) multi-byte UTF-8 sequence could cause the native binding to read past the end of the allocated buffer while attempting to decode the final, incomplete code point. This could result in an out-of-bounds read and potential disclosure of adjacent memory contents. This issue is fixed in version 1.26.1.

🎖@cveNotify