high2 sourcesCVE-2026-95281 CVE-2026-95284 +11
Google has rolled out Chrome version 154 to the Stable channel across Linux, Windows, and macOS endpoints to address a staggering 108 security vulnerabilities. The update brings Linux builds to version `154.0.8037.57` while pushing versions `154.0.8037.57` and `154.0.8037.58` to Windows and macOS environments.
Included in this massive release are 11 critical-severity bugs primarily clustered around Chrome's graphics processing, user interface, and WebGL rendering pathways. Specifically, three critical buffer overflow flaws were mitigated within ANGLE—the internal abstraction layer responsible for translating graphics API calls—tracked as `CVE-2026-95350`, `CVE-2026-95281`, and `CVE-2026-95284`.
The update also eliminates out-of-bounds write defects in the GPU subsystem with `CVE-2026-95357` and `CVE-2026-95322`, alongside WebGL memory-safety issues tracked as `CVE-2026-95349` and `CVE-2026-95329`. In addition to the graphics pipeline, several critical use-after-free conditions were corrected in components including `ServiceWorker` (`CVE-2026-95339`), `Fullscreen` (`CVE-2026-95313`), `WindowDialog` (`CVE-2026-95356`), and `AdFilter` (`CVE-2026-95310`).
Beyond the critical category, the release patches high-severity issues in the V8 JavaScript engine, including an out-of-bounds write (`CVE-2026-95304`) and a type confusion flaw (`CVE-2026-95306`) reported by OpenAI Codex Security. Additional high-severity fixes span use-after-free bugs and authorization flaws in modules such as `PDFium`, `WebAudio`, `Bluetooth`, `HID`, `DevTools`, `Aura`, and the extensions framework.
Google indicated that public access to in-depth bug details and exploit reproduction materials remains restricted until a significant portion of the user base has successfully applied the update.
Read more → about Massive Chrome 154 Update Drops Fixes for 108 Security Vulnerabilities Across Graphics Stack and V8 Enginecritical2 sourcesCVE-2026-102255 CVE-2026-102256 +6
SonicWall has rolled out emergency hotfixes to resolve four distinct security flaws targeting its enterprise-grade Secure Mobile Access (SMA) 1000 series products.
Leading the advisory is CVE-2026-102255, a maximum-severity, pre-authentication server-side request forgery (SSRF) flaw residing directly inside the Appliance WorkPlace interface. The issue stems from an unintended alternate access-path weakness that enables unauthenticated attackers to execute low-complexity network calls on behalf of the appliance.
By leveraging this unauthorized path, external actors can force the device to query trusted internal endpoints and trigger unauthorized operations that normally require administrative credentials.
Alongside this flaw, SonicWall patched CVE-2026-102256, an authenticated OS command injection bug also discovered by researcher Benoît Sevens. The release also covers CVE-2026-102257 and CVE-2026-102258, which represent path traversal and cross-site scripting vulnerabilities in the Appliance Management Console discovered by Brian Mariani.
The impacted lineup consists of the 6210 and 7210 hardware appliances as well as the virtual 8200v platform. Neither the SMA 100 series nor the core SonicOS firewall SSL-VPN stacks share this vulnerable codebase.
While vendor telemetry has not yet identified active exploitation, internet scans by Shadowserver revealed over 400 exposed SMA1000 installations visible online. SonicWall has urged administrators to immediately deploy hotfixes 12.4.3-03670, 12.5.0-03082, or later releases.
This update follows a persistent assault on the platform throughout 2026, including zero-day chains CVE-2026-15409, CVE-2026-15410, CVE-2026-83548, and CVE-2026-83549, which delivered custom payloads such as `OrangeTail` and `RootRun`.
Read more → about SonicWall Issues Emergency Hotfixes for Maximum-Severity Pre-Auth SSRF in SMA1000 Gatewayshigh2 sourcesCVE-2021-26086 CVE-2026-21589
Atlassian has published a critical security advisory addressing CVE-2026-21589, an arbitrary file access vulnerability affecting eight self-hosted Data Center and Server products.
The vulnerability carries an internal CVSS 4.0 rating of 9.3 and affects Bitbucket Data Center, Confluence Data Center, Jira Software Data Center, Jira Service Management Data Center, Bamboo Data Center, Crowd Data Center, Crucible, and Fisheye.
Under specific conditions, an unauthenticated remote attacker can read sensitive files stored within the target web application root directory without requiring any user interaction or valid credentials.
Exploitation requires the attacker to possess prior knowledge of the target file's precise name and path, as the flaw does not facilitate directory enumeration or listing.
While Atlassian patched its managed cloud offerings with no reported signs of compromise, administrators managing self-hosted installations are strongly advised to upgrade immediately.
For environments where instant patching is unfeasible, Atlassian recommends isolating internet-facing instances or applying specific URL filtering mitigations at the reverse proxy or WAF layer.
These temporary mitigations focus on blocking requests containing dot-dot path traversal sequences next to slashes or colons, alongside Tomcat `RewriteValve` rules or `urlrewrite.xml` configurations depending on the product.
Discrepancies emerged across CVE documentation and internal tickets regarding specific fixed versions for products like Crowd and Bamboo, as well as legacy Server editions.
Security teams are urged to inspect historic web access logs by searching for decoded and double-decoded traversal patterns to detect potential reconnaissance or exploitation attempts.
Read more → about Critical Path Traversal Bug Exposes Sensitive Files Across Eight Atlassian Data Center Productscritical6 sourcesCVE-2025-7775 CVE-2026-19490 +12
Citrix has released emergency security updates addressing eight vulnerabilities across customer-managed NetScaler ADC and NetScaler Gateway deployments, including two critical zero-day flaws tracked as CVE-2026-88771 and CVE-2026-88772. The patches arrived shortly after national cyber authorities, including the Dutch NCSC-NL and Australia's ACSC, began privately notifying administrators and advising them to shut down internet-facing NetScaler appliances immediately.
Both zero-day vulnerabilities enable unauthenticated threat actors to achieve remote code execution on target devices, sporting CVSS 4.0 scores of 9.5. Specifically, CVE-2026-88771 stems from improper input validation and affects all default configurations across NetScaler ADC and Gateway appliances without requiring special feature flags. The companion flaw, CVE-2026-88772, is a memory overflow vulnerability affecting systems where Datagram Transport Layer Security (DTLS) is enabled, which is configured on by default across VPN virtual servers.
The broader advisory also mitigates six additional flaws, ranging from HTTP request smuggling in CVE-2026-88773 (CVSS 9.3) to memory overflows and TCP initial sequence number prediction issues. In response to confirmed active exploitation, CISA updated its Known Exploited Vulnerabilities catalog and mandated that US federal agencies patch all vulnerable appliances under Binding Operational Directive 26-04 by September 30.
Remediation builds include NetScaler 14.1-73.37, 13.1-64.23, and corresponding updated FIPS and NDcPP packages, while unsupported releases such as versions 12.1 and 13.0 require immediate platform migration. Threat monitoring from Shadowserver currently detects over 23,000 reachable NetScaler IP addresses across the internet, underlining a massive attack surface for edge infrastructure. Citrix has published basic indicators of compromise through NetScaler Console, while simultaneously cautioning operators that these telemetry artifacts might be insufficient to identify sophisticated breaches.
Read more → about Citrix Patches Dual NetScaler Zero-Days Under Active Exploitation Following Emergency Shutdown Advisorieshigh1 sourceCVE-2026-71963 CVE-2026-72718
Researchers at Manifold Security disclosed a vulnerability class dubbed GitSpawn that affects multiple AI-powered developer tools, including Claude Code, Cursor, OpenAI Codex, Goose, Hermes Agent, Grok Build, and Qwen Code.
The issue stems from how coding agents automatically gather repository context on launch by executing background commands like `git status` or `git diff` before receiving user input or establishing workspace trust boundaries.
When a developer opens a project directory pre-populated with a `.git` folder, the local repository configuration dictates how Git executes certain features.
The primary attack vector leverages Git's legitimate `core.fsmonitor` configuration, which is designed to call external helper utilities to optimize file-change monitoring in massive codebases.
Because the automated agents run background Git commands without first sanitizing local repository settings, Git parses the attacker-supplied `core.fsmonitor` value and executes the specified executable on the host system.
The resulting payload runs with the developer's logged-in permissions, completely bypassing standard AI sandboxing and prompt-based confirmation dialogues.
Standard operations like `git clone` or `git pull` do not carry local repository configuration files, meaning attacks rely on direct directory transfers like ZIP archives, shared storage drives, or sync folders.
Researchers identified eight distinct vulnerabilities across the affected agents, observing that the spawned processes run entirely in the background without UI alerts or user interaction.
Remediation has begun across the industry, with Goose patching CVE-2026-72718 in version 1.44.0 and Claude Code addressing its `core.fsmonitor` vector in version 2.1.196 alongside fixes for Cursor and Codex.
However, several products remained vulnerable at disclosure time, including Hermes Agent (CVE-2026-71963) in version 0.21.0, Grok Build version 1.0.13, Qwen Code version 0.22.3, and an additional unpatched `ultrareview` configuration execution sink in Claude Code.
Read more → about GitSpawn Flaw in AI Coding Agents Executes Malicious Commands via Git Configurationshigh2 sourcesCVE-2026-84323 CVE-2026-84324 +24
Google has rolled out an update for Chrome's desktop Stable Channel addressing 26 security vulnerabilities across the browser ecosystem.
The update bumps Chrome to version 152.0.7977.75/.76 for Windows and macOS, alongside versions 152.0.7977.75 or 152.0.7977.76 on Linux.
Leading the release are two critical memory safety defects, tracked as `CVE-2026-84353` and `CVE-2026-84352`.
`CVE-2026-84353` represents a critical use-after-free bug in the Shared Tab Groups component, while `CVE-2026-84352` is a use-after-free flaw inside Chrome’s WebGL graphics rendering pipeline.
In addition to the critical items, the release resolves nine high-severity vulnerabilities affecting subsystems like `V8`, `Skia`, `Omnibox`, `FileSystem`, and the GPU process.
These high-severity flaws include an uninitialized resource condition in `V8` (`CVE-2026-84326`), a GPU buffer overflow (`CVE-2026-84351`), and an incorrect authorization bug in `FileSystem` (`CVE-2026-84354`).
The update also mitigates fifteen medium and low-severity defects spanning components like WebRTC, TabStrip, Navigation, and FullScreen UI handling.
Google indicated that public bug details remain restricted until a majority of active desktop clients apply the fixes.
Furthermore, neither Google nor tracking sources noted any evidence of active exploitation in the wild for these specific CVEs prior to patch deployment.
Read more → about Google ships patches for 26 Chrome security flaws including critical memory bugs in WebGL and Shared Tab Groupscritical3 sourcesCVE-2025-3248 CVE-2026-0768 +6
Threat actors have started actively exploiting a critical unauthenticated remote code execution vulnerability, tracked as CVE-2026-0768, within the open-source AI platform Langflow.
The flaw carries a CVSS score of 9.8 and resides in the code validator of the framework's custom component editor, affecting all versions up to 1.4.2.
Because user-supplied strings sent to the `validate` endpoint are processed without adequate sanitization, the application ultimately passed malicious input to an underlying `exec()` call, allowing arbitrary Python execution as root.
Cybersecurity intelligence firm VulnCheck initially captured attacks on its honeypot infrastructure in the United Kingdom, observing attempts surge from 50 over the weekend to roughly 360 hits by Monday.
Telemetry indicates that the attack traffic primarily originates from Russian IP space and focuses heavily on early reconnaissance and key exfiltration.
Adversaries are actively querying environment variables such as `LANGFLOW_SUPERUSER`, `OPENAI_API*`, `AWS_ACCESS*`, and `AWS_SECRET*` to harvest operational secrets.
Beyond environment dumps, the incoming exploit payloads inspect the size of `.bash_history`, examine `.ssh` folder configurations, and read the application secret key file located at `/root/.cache/langflow/secret_key`.
This incident follows an aggressive series of exploits targeting the platform throughout the year, including CVE-2026-33017, CVE-2026-5027, CVE-2026-55255, and CVE-2026-0770.
While the original zero-day was reported via the Zero Day Initiative back in July 2025 and publicly disclosed in January 2026, researchers noted that the problematic `exec()` call was eliminated in version 1.10.1.
Security teams running self-hosted Langflow deployments are urged to update instances immediately to release 1.11.6 or later to defend against active exploitation.
Read more → about Zero-Day Exploit Wave Targets Langflow RCE Flaw to Scrape Cloud and AI API Keyshigh1 sourceCVE-2026-75604
On August 25, 2026, Vercel released security updates for Next.js, publishing versions 15.5.24 and 16.3.3 to resolve two critical unauthenticated remote code execution vulnerabilities. The first flaw carries a CVSS score of 9.5 and stems from an underlying dependency in the `sharp` image optimization package. Specifically, `sharp` delegates AVIF image decoding to the C library `libheif`, which suffers from a memory corruption bug. When processing a crafted AVIF file containing two transparency layers of differing depths, `libheif` allocates memory based on the first layer but writes the larger second layer data, resulting in an attacker-controlled heap overflow of approximately 16 KB. This issue affects any application invoking `heif_decode_image()` within vulnerable `libheif` builds. Within Next.js, this affects versions 10.0.0 through 15.5.23 and 16.x up to 16.3.2, though AVIF optimization is disabled by default in `next.config.js`. The second vulnerability, tracked as CVE-2026-75604 with a CVSS score of 9.0, is a directory traversal flaw leading to remote code execution on Windows environments. Exploitation requires three specific conditions: the deployment must run on a Windows file system, rely simultaneously on both the Pages Router and the App Router, and run without Cache Components enabled. Versions 13.4 through 15.5.23 and 16.0 through 16.3.2 are affected by the traversal issue. Vercel confirmed that applications deployed directly on their managed cloud platform are not exposed due to backend Linux environments and proactive disabling of AVIF optimization.
Read more → about Vercel Patches Pair of Unauthenticated RCE Flaws in Next.js Image Processing and Windows Routingcritical1 sourceCVE-2026-18431
A critical vulnerability tracked as CVE-2026-18431 with a 9.8 CVSS score enables unauthenticated remote code execution on WordPress instances running the Avada theme. The exploit ties together six distinct flaws spanning authorization, input validation, trust-boundary, and file-handling weaknesses into a zero-click attack path. Successful exploitation allows an attacker to execute arbitrary PHP code directly on the target host. This level of access enables complete site takeover, database exfiltration, malware deployment, malicious redirects, and the creation of rogue administrator accounts. The vulnerability impacts Avada versions up to 7.16 alongside Fusion Builder plugin versions up to 3.16. While both components are strictly required to trigger the full exploit chain, Fusion Builder is a mandatory plugin bundled with the Avada theme. Consequently, all sites using the Avada theme ship with the necessary dependencies and remain vulnerable. Defiant's Wordfence team discovered and reproduced the issue using an automated agentic framework named Argus, which also developed functional proof-of-concept exploit code. ThemeFusion resolved the flaws by issuing patches in Avada version 7.16.1 and Fusion Builder version 3.16.1. Full technical details are currently withheld by researchers to allow system administrators time to apply the updates.
Read more → about Critical Exploit Chain in Avada Theme and Fusion Builder Unlocks Unauthenticated RCEcritical1 sourceCVE-2026-77537 CVE-2026-77545 +3
Ubiquiti released a major security advisory addressing 22 vulnerabilities across nearly its entire product stack. The impacted software includes UniFi Protect, UniFi Network, UniFi Access, UniFi Talk, UniFi Connect, and the core UniFi OS platform. Out of the 22 addressed flaws, 21 carry a critical severity rating, with three vulnerabilities scoring a maximum CVSS rating of 10.0. Specifically, CVE-2026-77537 enables unauthenticated network attackers to execute arbitrary commands on the UniFi Protect host due to poor input validation. A related unauthenticated command injection bug, tracked as CVE-2026-77554, affects the UniFi Talk application. For the underlying operating platform, CVE-2026-77550 allows attackers to bypass authentication on UniFi OS instances via improper neutralization of CRLF sequences. Another notable vulnerability, CVE-2026-77545, carries a 9.0 CVSS score and stems from leftover active debugging code in UniFi OS. Additionally, CVE-2026-77549 presents another authentication bypass condition linked to CRLF handling. Ubiquiti indicated that these bugs are easily exploitable across local or remote networks without user interaction, though active in-the-wild exploitation was not confirmed prior to patch publication. Fixed software releases have been deployed across all affected product lines, requiring administrators to update to versions such as UniFi Network Application 10.5.67 and UniFi OS Server 5.1.37.
Read more → about Ubiquiti Patches 22 Vulnerabilities Across UniFi Ecosystem, Including Three Max-Severity Flawscritical2 sourcesCVE-2024-4879 CVE-2024-5178 +6
ServiceNow released security updates addressing four vulnerabilities in its AI Platform (formerly the Now Platform), including three maximum-severity flaws carrying a perfect 10.0 CVSS score. The trio of critical flaws comprises CVE-2026-18885, CVE-2026-18886, and CVE-2026-74820. Specifically, CVE-2026-18885 is a code injection flaw in the GraphQL Composite Data API enabling unauthenticated remote code execution and data tampering. CVE-2026-18886 involves improper access control in the system configuration image upload processor that allows unauthenticated actors to modify data and escalate privileges. Meanwhile, CVE-2026-74820 allows unauthenticated attackers to execute arbitrary SQL queries against the underlying database via a dynamic schema ORDER BY clause. Each of these three vulnerabilities is network-reachable with low attack complexity, requiring no prior privileges or user interaction. Alongside these, ServiceNow addressed CVE-2026-6876, a high-severity sandbox escape rated at 8.7 that allows arbitrary code execution. Affected major releases include Xanadu, Yokohama, Zurich, and Australia, spanning various patch and hotfix levels. ServiceNow has deployed automatic patches across its hosted cloud instances, but operators managing self-hosted infrastructure must deploy the fixes manually. While ServiceNow reported no known active exploitation for these specific flaws at publication time, prior platform flaws have frequently been targeted in the wild.
Read more → about ServiceNow drops urgent updates for trio of critical 10.0 flaws and sandbox escapecritical3 sourcesCVE-2015-3246 CVE-2015-5287 +5
On August 26, 2026, CISA expanded its Known Exploited Vulnerabilities catalog with six additions across enterprise technologies. A prominent entry is CVE-2026-8452, a memory overflow issue affecting NetScaler ADC and Gateway appliances configured for AAA virtual servers or VPN gateways. While Citrix initially described the vulnerability as causing unpredictable behavior or denial-of-service, security researchers at WatchTowr published a proof-of-concept demonstrating unauthenticated remote code execution. Threat telemetry firms Previdian and Defused subsequently reported in-the-wild exploitation involving web shell deployments and discovery commands like `id` and `echo`. Fixed firmware releases include versions `14.1-72.61`, `13.1-63.18`, and `13.1-37.272`. Alongside the NetScaler flaw, CISA flagged CVE-2019-1068, a remote code execution bug in Microsoft SQL Server triggered via crafted queries. This flaw enables attackers to execute arbitrary code within the database engine's service account context. The catalog additions also encompass legacy vulnerabilities including Red Hat `libuser` race conditions (CVE-2015-3246), Red Hat ABRT privilege escalation (CVE-2015-5287), Ajax.NET Professional deserialization (CVE-2021-23758), and a Linux kernel out-of-bounds write flaw (CVE-2022-0995). CISA established an August 29 remediation deadline for the Citrix and Microsoft SQL vulnerabilities for affected federal entities, with remaining flaws slated for mitigation by September 9. Historical context notes this follows rapid exploitation of another NetScaler flaw, CVE-2026-8451, earlier in the year.
Read more → about CISA Updates KEV Catalog as Threat Actors Target Citrix NetScaler and Legacy Enterprise Softwarecritical3 sourcesCVE-2026-15981 CVE-2026-61979
Two critical authentication bypass vulnerabilities in the miniOrange SAML 2.0 Single Sign-On WordPress plugin, tracked as CVE-2026-61979 and CVE-2026-15981, are being actively targeted in the wild. The flaws stem from signature algorithm confusion and improper handling of OpenSSL verification errors, allowing unauthenticated attackers to forge SAML assertions and acquire administrator session cookies. Because the vendor distributes multiple commercial editions under a shared plugin slug, many paid installations fail to show update alerts within the WordPress dashboard and require manual upgrades.
Read more → about Active Exploitation Targets Critical miniOrange SAML Flaws in WordPresscritical2 sourcesCVE-2017-10271 CVE-2020-14882 +2
CISA has added a critical remote code execution vulnerability in Oracle HTTP Server and the WebLogic Server Proxy plugin, tracked as CVE-2026-21962, to its Known Exploited Vulnerabilities catalog. The max-severity flaw allows unauthenticated network access and has been actively exploited in the wild since January. Federal civilian agencies have been given a tight deadline to apply the patches released by Oracle earlier this year.
Read more → about CISA Orders Urgent Patching for Actively Exploited Oracle WebLogic Flaw