CISA added this pre-auth Windows RCE to KEV today with a 2026-08-21 deadline after Unit 42 and Tenable tied it to active attack campaigns, and any internet-facing Windows VPN endpoint left unpatched this week is exposed to unauthenticated remote code execution.
A double free in the Windows IKE (Internet Key Exchange) Service Extensions lets an unauthenticated attacker execute code by sending crafted packets to any Windows machine with IKEv2 enabled — commonly VPN servers and RAS endpoints listening on UDP 500/4500. No credentials or user interaction are required.
Affected: Windows 10 1607/1809/21H2/22H2 below builds 14393.9060 / 17763.8644 / 19044.7184 / 19045.7184; Windows 11 22H3/23H2 below 22631.6936, 24H2 below 26100.8246, 25H2 below 26200.8246, 26H1 below 28000.1836; Windows Server 2016 below 14393.9060
How to Test
Run `winver` or `Get-ComputerInfo OsBuildNumber` and compare against the fixed builds above. Check whether IKE is exposed with `Get-Service IKEEXT` and `netstat -anp UDP | findstr ":500 :4500"`; any system listening on UDP 500/4500 from untrusted networks and below the fixed build is directly exploitable.
How to Patch
Install the April 2026 (or later) Windows security updates that bring systems to the fixed builds listed above. If you cannot patch immediately, block inbound UDP 500 and 4500 on systems that do not use IKE, and restrict inbound UDP 500/4500 to known VPN peer IP addresses on systems that require IKE — per Microsoft this reduces exposure but does not replace the update.
CISA added this unauthenticated vCenter code-execution flaw to KEV today with a 2026-08-21 deadline and VulnCheck records known ransomware use, so an unpatched vCenter reachable by attackers is a direct path to full-hypervisor-estate compromise this week.
A path traversal vulnerability in VMware vCenter allows an attacker with network access to vCenter to execute arbitrary code without credentials, granting control of the virtualization management plane and everything it manages. Exploitation has been linked to ransomware operations.
Affected: VMware vCenter 9.1.x < 9.1.0.0300; VMware vCenter 9.0.x < 9.0.2.0100; VMware vCenter 8.0 < 8.0 U3k; VMware Cloud Foundation 5.x, 9.0.x, 9.1.x; VMware vSphere Foundation 9.0.x, 9.1.x; VMware Telco Cloud Platform 5.1.x; VMware Telco Cloud Infrastructure 3.0
How to Test
Check the vCenter build in the vSphere Client (Help > About) or via `vpxd -v` on the appliance; anything below 8.0 U3k, 9.0.2.0100, or 9.1.0.0300 is vulnerable. Verify whether vCenter (port 443) is reachable from untrusted networks, and review vCenter access logs (/var/log/vmware/) for requests containing traversal sequences (../) or unexpected file paths.
How to Patch
Apply the VMSA-2026-0006 updates: upgrade vCenter to 8.0 U3k, 9.0.2.0100, or 9.1.0.0300 as appropriate, and apply the corresponding Cloud Foundation / vSphere Foundation / Telco Cloud async patches. Broadcom lists no workaround, so if immediate patching is impossible, restrict network access to vCenter to a dedicated management network or VPN and treat any exposed instance as potentially compromised, following CISA's forensics triage guidance.
CISA added this flaw to KEV today with a 2026-08-21 deadline after the Dutch NCSC confirmed active exploitation and public exploits appeared, so any Mac with Screen Sharing enabled and reachable over the network can be taken over without credentials this week.
An improper-authentication flaw in macOS Screen Sharing lets an attacker on the network authenticate to a Mac's Screen Sharing service without any valid credentials, resulting in full interactive control of the machine. Public exploit code is available.
Affected: macOS Sonoma < 14.8.9; macOS Sequoia < 15.7.9; macOS Tahoe < 26.6.1
How to Test
Check the macOS version via `sw_vers -productVersion`; anything below 14.8.9, 15.7.9, or 26.6.1 is vulnerable. Determine exposure with `sudo lsof -iTCP:5900 -sTCP:LISTEN` or by checking System Settings > General > Sharing for Screen Sharing/Remote Management enabled, and review screensharingd entries in the unified log (`log show --predicate 'process == "screensharingd"'`) for successful sessions from unknown source IPs.
How to Patch
Update to macOS Sonoma 14.8.9, Sequoia 15.7.9, or Tahoe 26.6.1 (released August 6, 2026) via Software Update or MDM. Until patched, disable Screen Sharing and Remote Management on all Macs where they are not essential, and block inbound TCP 5900 at network and host firewalls from untrusted networks.
CISA added this SharePoint authentication bypass to KEV today with a 2026-08-21 deadline while a public Rapid7 exploit is circulating, so any unpatched on-prem SharePoint farm can be anonymously impersonated within days.
A weak-authentication flaw in SharePoint's JWT token handling lets an unauthenticated attacker make an anonymous network connection, bypass authentication, and impersonate users — allowing disclosure of files and modification of data across the farm. A public proof-of-concept exploit is available.
Affected: SharePoint Enterprise Server 2016 < 16.0.5561.1001; SharePoint Server 2019 < 16.0.10417.20175; SharePoint Server Subscription Edition < 16.0.19725.20434
How to Test
Check the farm build number in Central Administration (Servers in Farm) or via `(Get-SPFarm).BuildVersion` and compare to the fixed builds above. Review IIS logs on SharePoint front-ends for anonymous requests to authentication/token endpoints followed by authenticated activity from the same source IP, and audit ULS logs for unexpected token validation successes.
How to Patch
Install the July 2026 SharePoint security updates bringing farms to 16.0.5561.1001 (2016 — same KB covers SharePoint Server and Enterprise Server 2016), 16.0.10417.20175 (2019), or 16.0.19725.20434 (Subscription Edition), then run the SharePoint Products Configuration Wizard on every server. If external exposure is not required, restrict internet access to the farm behind VPN or an authenticating reverse proxy until patched.
VulnCheck added this CVSS 10.0 flaw to its known-exploited catalog on 2026-08-14 and Security Affairs confirmed in-the-wild attacks the next day, so unpatched Data Hub Adapter instances are being probed and compromised right now.
A default authentication client in the SAP Commerce Cloud Data Hub Adapter lets an unauthenticated remote attacker submit crafted input to functions lacking validation, achieving arbitrary code execution and full compromise of confidentiality, integrity, and availability of internal components. The CVSS 3.1 score is a maximum 10.0 with a network, no-privileges, no-interaction, scope-changed vector.
Affected: SAP Commerce Cloud (Data Hub Adapter) COM_CLOUD 2211; SAP Commerce Cloud (Data Hub Adapter) 2211-JDK21
How to Test
Inventory whether your SAP Commerce Cloud 2211 or 2211-JDK21 deployment includes the Data Hub Adapter extension and whether its endpoints are reachable from the internet; check that the default authentication client shipped with the adapter has not been left enabled with stock credentials. Review web/application access logs for unauthenticated requests hitting Data Hub Adapter endpoints, especially unexpected POSTs with crafted payloads since mid-August 2026, and look for anomalous processes or outbound connections from the Commerce application nodes.
How to Patch
Apply SAP Security Note 3771065 (released on SAP Security Patch Day) to the Data Hub Adapter on COM_CLOUD 2211 / 2211-JDK21 immediately. If patching cannot happen today, disable or remove the default authentication client used by the Data Hub Adapter, restrict network access to the adapter endpoints to trusted integration hosts only, and treat any exposed unpatched instance as potentially compromised pending forensic review.
CISA added this flaw to KEV yesterday with a 2026-08-20 remediation deadline, and exploitation reporting ties it to botnet activity, so any team running an unpatched Ray cluster or dashboard is an active target this week.
A code injection flaw in Ray allows remote code execution on machines running Ray, and can be triggered against developers through the Firefox and Safari browsers, meaning even Ray instances bound to non-public interfaces can be reached via a victim's browser. Successful exploitation gives the attacker code execution on the Ray node.
Affected: Ray (ray-project) < 2.52.0
How to Test
Run `ray --version` (or check `pip show ray`) on all nodes; any version below 2.52.0 is vulnerable. Check whether the Ray dashboard (default port 8265) or Ray client ports are listening on non-loopback interfaces with `ss -tlnp | grep 8265`, and review dashboard/API access logs for unexpected job submissions or requests with browser Origin headers.
How to Patch
Upgrade all Ray installations to 2.52.0 or later (`pip install -U 'ray>=2.52.0'`). Until upgraded, bind the Ray dashboard and job-submission API to 127.0.0.1 only, firewall port 8265 and Ray cluster ports from untrusted networks, and avoid browsing untrusted sites from machines running a local Ray instance.
The flaw entered VulnCheck KEV today with CISA's SSVC decision marked active, and teams that leave MLflow tracking servers below 3.15.0 reachable this week risk cloud IAM credential theft via the instance metadata service.
MLflow's outbound webhook SSRF protection resolves and validates a hostname but lets the actual request re-resolve it independently, so an attacker-controlled DNS server can pass validation with a public IP and then rebind to 169.254.169.254 or 127.0.0.1. The attacker receives the response body, enabling theft of cloud IAM credentials and secrets from instance metadata endpoints.
Affected: MLflow < 3.15.0
How to Test
Run 'pip show mlflow' (or check the server's /version endpoint) to confirm whether you are below 3.15.0; audit configured webhooks for attacker-controlled or recently changed destination URLs; review cloud audit logs (e.g., AWS CloudTrail) for IAM credential use originating from your MLflow tracking server's role.
How to Patch
Upgrade to MLflow 3.15.0, which adds an SSRF-protected HTTP adapter that validates the peer IP at connect time and disables environment proxy handling; as interim mitigation, block egress from MLflow hosts to 169.254.169.254 and internal ranges, and enforce IMDSv2 with a hop limit of 1 on AWS instances.
CERT Polska reports active exploitation and CISA's SSVC decision is 'active'; Zimbra servers not upgraded to 10.1.20 this week are exposed to unauthenticated remote command execution over SMTP.
A command injection flaw in Zimbra's SNMP monitoring component (triggered when SNMP notifications are enabled) lets an unauthenticated attacker send crafted SMTP requests that execute arbitrary OS commands as the zimbra user. Successful exploitation gives full control of the mail server and everything it stores.
Affected: Zimbra Collaboration Suite (ZCS) < 10.1.20
How to Test
Run `zmcontrol -v` to confirm the installed version; any release below 10.1.20 is vulnerable. Check whether SNMP notifications are enabled in the zimbra-snmp/monitoring configuration — the injection path requires them. Review MTA and system logs for unexpected processes or commands spawned by the zimbra user following inbound SMTP connections.
How to Patch
Upgrade to ZCS 10.1.20 immediately via `yum update` or `apt update` per your install method. If you cannot upgrade today, disable SNMP notifications in the SNMP monitoring component as an interim mitigation and restrict inbound SMTP exposure where feasible.
With Wordfence and Patchstack both reporting in-the-wild exploitation, two public exploits, and a VulnCheck KEV listing from the past three days, unpatched Pods sites face unauthenticated admin takeover this week.
The pods_admin AJAX router funnels every access check — method allowlist, nonce verification, login enforcement, and capability gates — through pods_error(), which under the JSON meta-box-loader compatibility path merely logs failures and returns false instead of terminating the request. Unauthenticated attackers can therefore invoke admin-only methods and escalate to full site compromise.
Affected: Pods 2.8 – 2.8.23.3; Pods 2.9 – 2.9.19.3; Pods 3.0 – 3.0.10.3; Pods 3.1 – 3.1.4.1; Pods 3.2 – 3.2.8.2; Pods 3.3 – 3.3.9
How to Test
Run 'wp plugin list' or check Plugins in wp-admin to confirm the Pods version; grep web server access logs for unauthenticated POSTs to admin-ajax.php with action=pods_admin, and check the PHP error log for pods_error entries indicating swallowed authorization failures; audit wp_users for unexpected administrator accounts.
How to Patch
Update Pods to 3.3.9.1 (or the corresponding patched release for your branch) immediately; if you cannot update today, deactivate the plugin or block admin-ajax.php requests with action=pods_admin from unauthenticated sources at the WAF.
VulnCheck added this CVSS 9.8 unauthenticated auth bypass to its KEV catalog with Wordfence reporting exploitation, and sites still on 3.16.4 or earlier can be taken over as administrator this week.
The plugin's wppb_log_in_user() coerces a WP_Error from wp_insert_user() to integer 1 via absint() before the error check, so a registration submitted with a 61–70 character username causes the plugin to issue an autologin nonce bound to user ID 1 — typically the site administrator. An unauthenticated attacker can use this to log in as admin and fully take over the site.
Affected: cozmoslabs User Profile Builder (profile-builder) <= 3.16.4
How to Test
Check the installed plugin version in wp-admin → Plugins or via `wp plugin get profile-builder`; anything at or below 3.16.4 is vulnerable. Review access logs for registration form POSTs containing 61–70 character usernames, and audit wp_users for unexpected recent logins or sessions on user ID 1 and any newly created administrator accounts.
How to Patch
Update the profile-builder plugin to the first release after 3.16.4 (via wp-admin or `wp plugin update profile-builder`). If you cannot update immediately, disable public registration through the plugin's forms or deactivate the plugin, and rotate administrator credentials and sessions if you find suspicious autologin activity.