Deck for 2026-11-13 (UTC)12 briefs
ThreatDeck — 2026-11-13
12 cards today · 6 actively exploited or must-patch · 6 affecting AI tooling.
- 01
Act Now Critical KEV CVE-2026-111560 AI-assisted
CISA adds Barracuda multiple products OS command injection (CVE-2026-111560) to KEV
CISA added CVE-2026-111560 to its Known Exploited Vulnerabilities catalog on 2026-11-10. The OS command injection flaw affects Barracuda ESG, Load Balancer, ADC, Message Archiver, WAF Control Center, WAF, and Web Security Gateway. An authenticated administrator can run OS commands as root when a crafted timezone value is processed. The flaw can be chained with CVE-2026-116606. The federal due date is 2026-11-13.
Why it matters. Barracuda perimeter and email security appliances are exposed devices, and exploitation yields root-level command execution. Chaining with CVE-2026-116606 could let an attacker without admin credentials reach this flaw, though the source only states that the two can be chained.
Do this. Apply Barracuda's vendor mitigations and patches to all affected products now, following the Barracuda security advisory.
2026-11-10 · score 63 · CISA Known Exploited Vulnerabilities
- 02
Act Now Critical KEV CVE-2026-116606 AI-assisted
CISA adds Barracuda ESG authentication bypass CVE-2026-116606 to KEV; due Nov 13
CISA added CVE-2026-116606 to its Known Exploited Vulnerabilities catalog on 2026-11-10. The flaw in Barracuda Email Security Gateway lets an unauthenticated remote attacker invoke a configuration-import function and restore a crafted backup archive. This changes the appliance's global configuration, and the flaw can be chained with CVE-2026-111560.
Why it matters. Organizations running Barracuda ESG appliances exposed to the internet face unauthorized configuration changes by unauthenticated attackers. Mail flow and security controls on the gateway could be altered, and the chain with CVE-2026-111560 may increase impact.
Do this. Apply the vendor mitigations described in the Barracuda security advisory on pre-authentication authorization bypass and command injection immediately.
2026-11-10 · score 63 · CISA Known Exploited Vulnerabilities
- 03
Act Now Critical KEV CVE-2026-125350 AI-assisted
CISA KEV: Chromium XML use-after-free (CVE-2026-125350) enables sandboxed code execution
CISA added CVE-2026-125350, a use-after-free flaw in the XML component of Google Chromium, to its Known Exploited Vulnerabilities catalog on 2026-11-10. A remote attacker could execute arbitrary code inside the sandbox through a crafted webpage. Other Chromium-based browsers, including Chrome, Edge and Opera, may be affected. The due date for required action is 2026-11-24.
Why it matters. The flaw is listed as exploited, and any organization with Chromium-based browsers is exposed to drive-by code execution within the browser sandbox. An attacker could use it as a first step in a chain toward a full compromise.
Do this. Update Google Chrome to the latest stable release listed in the Chrome Releases blog notes for this CVE.
2026-11-10 · score 62 · CISA Known Exploited Vulnerabilities
- 04
Act Now High KEV CVE-2026-20320 AI-assisted
CISA adds Cisco BroadWorks XXE flaw CVE-2026-20320 to Known Exploited Vulnerabilities catalog
CISA added CVE-2026-20320, an XML external entity (XXE) vulnerability in Cisco BroadWorks, to its Known Exploited Vulnerabilities catalog on 2026-11-10. An unauthenticated, remote attacker could read sensitive configuration information on an affected system. The remediation due date is 2026-11-24, and use in ransomware campaigns is listed as unknown.
Why it matters. Organizations running Cisco BroadWorks are exposed to remote, unauthenticated reads of sensitive configuration data. Inclusion in the KEV catalog indicates exploitation has been observed, and attackers could use leaked configuration details to support further attacks.
Do this. Apply Cisco's mitigations and fixes as described in advisory cisco-sa-bworks-xxe-uwUd7CEt.
2026-11-10 · score 55 · CISA Known Exploited Vulnerabilities
- 05
Act Now High KEV CVE-2026-124194 AI-assisted
CISA adds Microsoft Office Outlook spoofing flaw CVE-2026-124194 to KEV catalog
CISA added CVE-2026-124194 to its Known Exploited Vulnerabilities catalog on 2026-11-10. The flaw is a missing user interaction for potentially unsafe action in Microsoft Office Outlook. It allows an unauthorized attacker to perform spoofing over a network. The due date for required action is 2026-11-24.
Why it matters. Listing in the KEV catalog indicates the flaw is known to be exploited. Organizations running Outlook are exposed to network-based spoofing that needs no authorization from the attacker.
Do this. Apply Microsoft's mitigations and updates per the MSRC guidance for CVE-2026-124194.
2026-11-10 · score 55 · CISA Known Exploited Vulnerabilities
- 06
Act Now High KEV CVE-2026-111793 AI-assisted
CISA adds Openwave Webtop Media Transcoder CVE-2026-111793 to Known Exploited Vulnerabilities catalog
CISA added CVE-2026-111793 to its Known Exploited Vulnerabilities catalog on 2026-11-10. The flaw is an incorrect resource transfer between spheres in Openwave Messaging Limited Webtop Media Transcoder. An authenticated user may be able to process and convert unintended document types if bypassing the Webmail experience. Known ransomware use is listed as Unknown.
Why it matters. Inclusion in the KEV catalog indicates exploitation has been observed. Organizations running Webtop Media Transcoder are exposed to authenticated users abusing the transcoder outside the intended Webmail path. Federal agencies face a due date of 2026-11-24.
Do this. Identify all deployments of Openwave Webtop Media Transcoder and assess their internet exposure.
2026-11-10 · score 54 · CISA Known Exploited Vulnerabilities
- 07
AI Stack Alert Critical CVE-2026-112136 AI-assisted
CVE-2026-112136: Mendix Studio Pro MCP Server HTTP listener exposed beyond loopback
Siemens Mendix Studio Pro 11 (>= V11.15.0 < V11.16.0) and Mendix Studio Pro 11.12 (>= V11.12.6 < V11.12.7) contain a flaw in the MCP Server. Its HTTP listener is not properly restricted to the loopback interface and may be reachable from remote systems. The CVSS base score is 9.2.
Why it matters. An attacker with network access to a developer workstation running an affected version could interact with the MCP Server remotely. This could expose the development environment and any tools or data the MCP Server provides.
Do this. Upgrade Mendix Studio Pro to a fixed version per the vendor advisory (V11.16.0 or later for 11.x, V11.12.7 or later for 11.12).
2026-11-10 · score 43 · NVD (AI keyword search)
- 08
AI Stack Alert High CVE-2026-126197 AI-assisted
CVE-2026-126197: pgEdge Postgres MCP Server pl-do custom tools allow SQL injection before 1.1.0
pgEdge Postgres MCP Server versions prior to 1.1.0 interpolated JSON arguments from custom pl-do tools into a PostgreSQL DO block with fixed dollar-quote tags. An argument containing the tag text could terminate the block and inject arbitrary SQL, bypassing read_only and reaching data or resources available to the connecting role. The issue is fixed in 1.1.0 and requires at least one pl-do custom tool.
Why it matters. Organizations exposing this MCP server to AI clients may have a database role that can read or modify data beyond what read_only was meant to allow. CVSS base score is 8.8.
Do this. Upgrade pgEdge Postgres MCP Server to version 1.1.0 or later.
2026-11-12 · score 36 · NVD (AI keyword search)
- 09
AI Stack Alert High CVE-2026-126196 AI-assisted
CVE-2026-126196: pgEdge Postgres MCP Server read_only bypass via simple query protocol before 1.1.0
pgEdge Postgres MCP Server enables SQL queries against PostgreSQL through MCP-compatible clients. Prior to 1.1.0, query_database on a read_only connection could send unrecognized statements without bind parameters, allowing multiple semicolon-separated statements in one request. This could let a caller make database writes when the connecting role had write privileges, and the issue is fixed in 1.1.0.
Why it matters. Organizations that rely on read_only mode as a safety boundary for MCP-driven database access may have writable sessions they believe are read-only. The flaw affects integrity and availability of the database.
Do this. Upgrade pgEdge Postgres MCP Server to version 1.1.0 or later.
2026-11-12 · score 36 · NVD (AI keyword search)
- 10
AI Stack Alert High TD-2026-0062 Unconfirmed AI-assisted
Attackers Used Open-Source Cairn AI Agent with Chinese Models to Autonomously Breach 19 Organizations
From August to September 2026, attackers used the open-source Cairn AI penetration testing framework to conduct 25 autonomous operations against at least 19 organizations. The operations were paired with Zhipu GLM and DeepSeek models. Unit 42 confirmed unauthorized access or data retrieval at 10 of the targeted organizations.
Why it matters. Semiconductor, defense and civil society organizations face faster, higher-volume intrusion attempts that need little operator effort. The confirmed outcomes included persistent administrator accounts and mass personal data exposure.
Do this. Enforce phishing-resistant MFA and unique passwords on internet-facing applications, and audit for weak or default credentials.
2026-11-12 · score 36 · Palo Alto Unit 42
- 11
AI Stack Alert High TD-2026-0066 Unconfirmed AI-assisted
AntiFox Agentic Pentest Platform Abused in Intrusions on IIS Servers and AWS Accounts
Arctic Wolf Labs reported attackers using AntiFox, an open-source Chinese-language agentic penetration testing platform, in intrusions against internet-facing web services and AWS environments. Some intrusions began with a public IIS application compromise, and others started with a valid AWS credential of unknown origin. Alibaba Cloud systems hosting AntiFox interfaces were associated with the confirmed malicious activity.
Why it matters. Organizations running internet-facing IIS applications or holding AWS credentials are exposed. Attackers gained access to database credentials, production secrets, and S3 data in observed cases.
Do this. Identify all internet-facing IIS applications and prioritize patching and protecting file upload and document-processing services.
2026-11-11 · score 36 · Security Affairs
- 12
AI Stack Alert High TD-2026-0068 Unconfirmed AI-assisted
Attackers use self-hosted AntiFox AI pentest agents on Alibaba Cloud for AWS credential theft
Attackers are using Alibaba Cloud to host self-managed instances of AntiFox, an open-source agentic penetration testing platform, to target web services and AWS environments. Arctic Wolf observed malicious AntiFox-linked infrastructure using valid AWS identities, including IAM users and assumed EC2 roles, to make authenticated requests to AWS APIs. In some intrusions, the attacker obtained production secrets.
Why it matters. Organizations running internet-facing IIS applications with credentials in configuration files or long-term AWS keys are exposed. A stolen identity can give an attacker access to application secrets, S3 objects, and cloud resources.
Do this. Revoke affected AWS sessions and keys and rotate credentials exposed through application configuration and log files.
2026-11-11 · score 36 · BleepingComputer