Security Daily Brief
20 events37 source observations
Lead stories
Attackers' AI Agents Found a Zero-Day in Medical-Imaging Software and Gained Root Access
Suspected China-linked hackers used autonomous AI agents to find a previously unknown flaw in MONAI Label, an open-source medical imaging server, and gain root access to a university server, Amazon Threat Intelligence said this week.
Read the analysisThe day in context
Amazon Threat Intelligence says suspected China-linked attackers used autonomous agents to find an unknown MONAI Label flaw and gain root access to one university server. The operator chose the targets; the agents chose methods. Amazon’s attribution carries moderate confidence. Versions 0.8.5 and earlier have authentication disabled by default, allowing network-reachable clients to trigger an out-of-folder file write. Root access depends on service privileges. A source fix is merged, but no fixed release is available; operators should check their exposure.
CISA added four vulnerabilities to its Known Exploited Vulnerabilities catalog, citing active exploitation. They affect SAP NetWeaver, Cisco Expressway, VMware vCenter and Apache Roller. Federal civilian agencies must prioritize qualifying catalog flaws on publicly exposed assets where exploitation grants total control, and check for prior compromise when required. CISA encourages other organizations to manage exposure by risk. Listing establishes exploitation, not compromise of every deployment.
HUMAN’s Satori researchers say BADBOX rebuilt after last year’s disruption and reached roughly 40 billion fraudulent ad bid requests daily at its peak. Infected Android-based devices can play hidden ads and join residential proxy networks without owners’ consent. Some fraudulent requests involved neither a real device nor ad playback. HUMAN says renewed disruption and traffic filtering are underway, although earlier action failed to prevent the botnet’s return.
Lead stories
Attackers' AI Agents Found a Zero-Day in Medical-Imaging Software and Gained Root Access
Published:
Amazon Threat Intelligence says suspected China-linked attackers used autonomous Cairn agents to discover and exploit a zero-day in MONAI Label, an open-source medical imaging server, gaining root access to one university server. Amazon reconstructed the activity from the operator’s unprotected control panel. It attributes the campaign to a China-linked actor with moderate confidence, without identifying a particular group or establishing government involvement. The flaw affects versions 0.8.5 and earlier: authentication is off by default, and a network-accessible client can cause files to be written outside the intended storage directory. Writes are limited by the service account’s permissions, so root access is not inevitable on every deployment. Maintainers have merged a source-code fix, but no fixed release is yet available to install. Teams using this separate server for image labeling or model training should check their version, network exposure and service permissions while watching for the release.
Attackers' AI Agents Found a Zero-Day in Medical-Imaging Software and Gained Root Access
Published:
Amazon Threat Intelligence says suspected China-linked attackers used autonomous Cairn agents to discover and exploit a zero-day in MONAI Label, an open-source medical imaging server, gaining root access to one university server. Amazon reconstructed the activity from the operator’s unprotected control panel. It attributes the campaign to a China-linked actor with moderate confidence, without identifying a particular group or establishing government involvement. The flaw affects versions 0.8.5 and earlier: authentication is off by default, and a network-accessible client can cause files to be written outside the intended storage directory. Writes are limited by the service account’s permissions, so root access is not inevitable on every deployment. Maintainers have merged a source-code fix, but no fixed release is yet available to install. Teams using this separate server for image labeling or model training should check their version, network exposure and service permissions while watching for the release.
CISA Adds Four Known Exploited Vulnerabilities to Catalog | CISA
Published:
CISA has added four vulnerabilities to its Known Exploited Vulnerabilities catalog after finding evidence of active exploitation. They affect SAP NetWeaver, Cisco Expressway, VMware vCenter and Apache Roller, and involve, respectively, improper handling of special elements in data queries, hard-coded credentials, an authentication bypass and deserialization of untrusted data. Inclusion signals evidence of exploitation, not that every organization using these products has been compromised. Under BOD 26-04, federal civilian executive branch agencies must prioritize rapid remediation of cataloged high-risk flaws on publicly exposed assets where exploitation grants total control, and check for compromise before patching when required. The directive applies only to those agencies, but CISA encourages all organizations to use risk-based vulnerability management and prioritize catalog entries. Administrators should match the listed CVEs against their inventories and consider exposure when scheduling remediation.
CISA Adds Four Known Exploited Vulnerabilities to Catalog | CISA
Published:
CISA has added four vulnerabilities to its Known Exploited Vulnerabilities catalog after finding evidence of active exploitation. They affect SAP NetWeaver, Cisco Expressway, VMware vCenter and Apache Roller, and involve, respectively, improper handling of special elements in data queries, hard-coded credentials, an authentication bypass and deserialization of untrusted data. Inclusion signals evidence of exploitation, not that every organization using these products has been compromised. Under BOD 26-04, federal civilian executive branch agencies must prioritize rapid remediation of cataloged high-risk flaws on publicly exposed assets where exploitation grants total control, and check for compromise before patching when required. The directive applies only to those agencies, but CISA encourages all organizations to use risk-based vulnerability management and prioritize catalog entries. Administrators should match the listed CVEs against their inventories and consider exposure when scheduling remediation.
Sources · 8
BADBOX botnet returns with record-breaking 40 billion daily ad requests
Published:
BADBOX has rebuilt after last year’s disruption, with its latest variant generating roughly 40 billion fraudulent advertising bid requests a day at its peak, according to HUMAN’s Satori researchers. They describe BADBOX 3.0 as the largest documented ad-fraud operation by daily request volume. The malware infects inexpensive Android Open Source Project phones, tablets and streaming boxes. It can play ads in hidden windows while impersonating legitimate streaming apps, and enroll devices in residential proxy networks without their owners’ consent. Researchers also observed fraudulent requests for ads that were never played on a real device. Earlier disruption by HUMAN and Shadowserver disconnected more than 500,000 devices from control servers, but did not prevent the network’s return. Consumers should pay attention to device certification and source; advertisers should not treat request volume as evidence of genuine viewing. HUMAN says it has helped disrupt related infrastructure and is filtering associated fraud traffic.
BADBOX botnet returns with record-breaking 40 billion daily ad requests
Published:
BADBOX has rebuilt after last year’s disruption, with its latest variant generating roughly 40 billion fraudulent advertising bid requests a day at its peak, according to HUMAN’s Satori researchers. They describe BADBOX 3.0 as the largest documented ad-fraud operation by daily request volume. The malware infects inexpensive Android Open Source Project phones, tablets and streaming boxes. It can play ads in hidden windows while impersonating legitimate streaming apps, and enroll devices in residential proxy networks without their owners’ consent. Researchers also observed fraudulent requests for ads that were never played on a real device. Earlier disruption by HUMAN and Shadowserver disconnected more than 500,000 devices from control servers, but did not prevent the network’s return. Consumers should pay attention to device certification and source; advertisers should not treat request volume as evidence of genuine viewing. HUMAN says it has helped disrupt related infrastructure and is filtering associated fraud traffic.
Complete briefing
Vulnerabilities & patches
6The Restrictions for BuddyPress plugin for WordPress is vulnerable to Remote Code Execution in...
Published:
The Restrictions for BuddyPress plugin for WordPress is vulnerable to Remote Code Execution in all versions up to, and including, 1.5.4 This is due to the plugin splicing the attacker-controlled request URI into a PHP expression inside a double-quoted string and evaluating it with eval(). Because PHP interpolates double-quoted strings, a payload of the form {${...}} in the URL path is executed during interpolation, so no quote breakout is required and WordPress' slashing does not mitigate it.
The Restrictions for BuddyPress plugin for WordPress is vulnerable to Remote Code Execution in...
Published:
The Restrictions for BuddyPress plugin for WordPress is vulnerable to Remote Code Execution in all versions up to, and including, 1.5.4 This is due to the plugin splicing the attacker-controlled request URI into a PHP expression inside a double-quoted string and evaluating it with eval(). Because PHP interpolates double-quoted strings, a payload of the form {${...}} in the URL path is executed during interpolation, so no quote breakout is required and WordPress' slashing does not mitigate it.
GnuPG Warns Critical GPGME Flaw Could Expose App Data via Malicious OpenPGP Keys
Published:
A malicious OpenPGP key with an unusually long user ID can trigger a critical memory bug in GPGME, the standard library that programs use to work with GnuPG.
GnuPG Warns Critical GPGME Flaw Could Expose App Data via Malicious OpenPGP Keys
Published:
A malicious OpenPGP key with an unusually long user ID can trigger a critical memory bug in GPGME, the standard library that programs use to work with GnuPG.
Sources · 4
F5 releases emergency patches for critical BIG-IP RCE flaws
Published:
US cybersecurity company F5 warned customers to patch five BIG-IP vulnerabilities that attackers can exploit to trigger denial-of-service (DoS) or gain remote code execution.
F5 releases emergency patches for critical BIG-IP RCE flaws
Published:
US cybersecurity company F5 warned customers to patch five BIG-IP vulnerabilities that attackers can exploit to trigger denial-of-service (DoS) or gain remote code execution.
A new Windows zero-day reportedly gets past BitLocker using only the screen reader and a USB stick - TechSpot
Published:
Cutting corners: BitLocker can be beaten in minutes on a fully patched Windows 11 machine, and the weapon is the operating system's own screen reader. A zero-day exploit published this week by a security researcher abuses Windows Narrator to run code from the lock screen, before any user signs in.
A new Windows zero-day reportedly gets past BitLocker using only the screen reader and a USB stick - TechSpot
Published:
Cutting corners: BitLocker can be beaten in minutes on a fully patched Windows 11 machine, and the weapon is the operating system's own screen reader. A zero-day exploit published this week by a security researcher abuses Windows Narrator to run code from the lock screen, before any user signs in.
Amazon: Suspected China-Linked Actor Exploited Medical Software Zero-Day
Published:
A threat actor used autonomous Cairn AI agents to discover and exploit a critical MONAI Label vulnerability.
Amazon: Suspected China-Linked Actor Exploited Medical Software Zero-Day
Published:
A threat actor used autonomous Cairn AI agents to discover and exploit a critical MONAI Label vulnerability.
GPGME Double Free Bug Hid in GnuPG's Glue Library for 24 Years, and Most Mail Clients Ship It - Hardware Busters
Published:
GnuPG has shipped GPGME 2.3.1 to fix a critical GPGME double free that a crafted OpenPGP key can trigger, and the list of affected software runs from mail and XMPP clients to Flatpak and Samba. If something on your machine signs or verifies mail, it probably needs this update.
GPGME Double Free Bug Hid in GnuPG's Glue Library for 24 Years, and Most Mail Clients Ship It - Hardware Busters
Published:
GnuPG has shipped GPGME 2.3.1 to fix a critical GPGME double free that a crafted OpenPGP key can trigger, and the list of affected software runs from mail and XMPP clients to Flatpak and Samba. If something on your machine signs or verifies mail, it probably needs this update.
Exploitation & attacks
10Metasploit Wrap Up: Breaking Zimbra's Identity Boundaries
Published:
This release we have three pull requests from jburgess-r7 building out coverage for the identity-boundary issues detailed in Rapid7's Breaking Zimbra Identity Boundaries, Part One research.
Metasploit Wrap Up: Breaking Zimbra's Identity Boundaries
Published:
This release we have three pull requests from jburgess-r7 building out coverage for the identity-boundary issues detailed in Rapid7's Breaking Zimbra Identity Boundaries, Part One research.
A security flaw has been discovered in D-Link DIR-600L 2.17b02. This affects the function...
Published:
A security flaw has been discovered in D-Link DIR-600L 2.17b02. This affects the function FUN_0040efb0 of the file gena.cgi of the component GENA Subscription Handler.
A security flaw has been discovered in D-Link DIR-600L 2.17b02. This affects the function...
Published:
A security flaw has been discovered in D-Link DIR-600L 2.17b02. This affects the function FUN_0040efb0 of the file gena.cgi of the component GENA Subscription Handler.
Microsoft Entra Connect Health flaw exploited to steal tenant data
Published:
Microsoft is warning organizations that it suffered a data breach after attackers exploited a vulnerability in its Entra Connect Health service to gain unauthorized access to administrator email addresses and tenant configuration information.
Microsoft Entra Connect Health flaw exploited to steal tenant data
Published:
Microsoft is warning organizations that it suffered a data breach after attackers exploited a vulnerability in its Entra Connect Health service to gain unauthorized access to administrator email addresses and tenant configuration information.
Cyberattack on Austrian Foreign Ministry Exposes Diplomatic Staff Data
Published:
Hackers accessed Austrian Foreign Ministry records containing staff contact details, including employees at diplomatic missions abroad.
Cyberattack on Austrian Foreign Ministry Exposes Diplomatic Staff Data
Published:
Hackers accessed Austrian Foreign Ministry records containing staff contact details, including employees at diplomatic missions abroad.
Cairn AI Framework Used in Autonomous Zero-Day Exploitation
Published:
Amazon uncovered AI-powered attacks on universities, where autonomous agents found a zero-day and gained root access without human guidance.
Cairn AI Framework Used in Autonomous Zero-Day Exploitation
Published:
Amazon uncovered AI-powered attacks on universities, where autonomous agents found a zero-day and gained root access without human guidance.
The Outages Are Just Getting Started
Published:
WIRED says recent outages had individual, routine explanations, not a coordinated AI hacking spree. Researchers nevertheless argue that AI is accelerating vulnerability discovery and exploitation, increasing patching pressure and risks to system stability.
The Outages Are Just Getting Started
Published:
WIRED says recent outages had individual, routine explanations, not a coordinated AI hacking spree. Researchers nevertheless argue that AI is accelerating vulnerability discovery and exploitation, increasing patching pressure and risks to system stability.
AI-powered cyberattacks are here, and the EU just admitted to being a target for weeks - Digital Trends
Published:
The European Union has confirmed that a cyberattack using artificial intelligence hit a key institution in October, and some of the bloc’s digital tools were still offline when an official spoke on Thursday. According to the Associated Press, it is the first time the 27-member bloc has reported a cyberattack that used an AI tool.
AI-powered cyberattacks are here, and the EU just admitted to being a target for weeks - Digital Trends
Published:
The European Union has confirmed that a cyberattack using artificial intelligence hit a key institution in October, and some of the bloc’s digital tools were still offline when an official spoke on Thursday. According to the Associated Press, it is the first time the 27-member bloc has reported a cyberattack that used an AI tool.
Hackers Using Email Bombing Before Posing as IT Support on Microsoft Teams
Published:
Fortra warns of an email bombing campaign where hackers flood work inboxes, then pose as IT support on Microsoft Teams to get access to work accounts.
Hackers Using Email Bombing Before Posing as IT Support on Microsoft Teams
Published:
Fortra warns of an email bombing campaign where hackers flood work inboxes, then pose as IT support on Microsoft Teams to get access to work accounts.
Wrapped Monero Token Plunges 99.8% After Suspected Exploit
Published:
Wrapped Monero token XMR1 fell 99.8%, from about $500 to $1, on Hyperliquid in 30 minutes after a wallet sold about $13.5 million worth of tokens, according to Bitquery.
Wrapped Monero Token Plunges 99.8% After Suspected Exploit
Published:
Wrapped Monero token XMR1 fell 99.8%, from about $500 to $1, on Hyperliquid in 30 minutes after a wallet sold about $13.5 million worth of tokens, according to Bitquery.
‘So what if there’s a hole? It hasn’t collapsed’: Kyiv residents unbothered by Russian attacks on city’s bridges | Ukraine | The Guardian
Published:
It hasn’t collapsed’: Kyiv residents unbothered by Russian attacks on city’s bridges
‘So what if there’s a hole? It hasn’t collapsed’: Kyiv residents unbothered by Russian attacks on city’s bridges | Ukraine | The Guardian
Published:
It hasn’t collapsed’: Kyiv residents unbothered by Russian attacks on city’s bridges
Malware
1Fake GTA VI Streaming Client Installs Data-Stealing Malware on Windows PCs
Published:
Now, scammers are using the lack of an official PC version to deliver a fake streaming client that infects Windows computers with malware.
Fake GTA VI Streaming Client Installs Data-Stealing Malware on Windows PCs
Published:
Now, scammers are using the lack of an official PC version to deliver a fake streaming client that infects Windows computers with malware.