November 4, 2026

What happened

NIST released cybersecurity guidance for small businesses with no paid employees other than their owners, including freelancers, sole proprietors, and independent contractors. [1] [2]

NIST said the resource was designed for solopreneurs because they may not have cybersecurity expertise. [5]

Why it matters

The source reports that about 82% of the 34.8 million U.S. small businesses fall into this category, citing the Small Business Administration’s Office of Advocacy. [3] [4]

Cited source evidence

These excerpts support the numbered claims above. They come from one source article and do not establish independent corroboration.

1. [1]

   The National Institute of Standards and Technology (NIST) released cybersecurity guidance for small businesses with no paid employees other than their owners.

2. [2]

   These businesses include freelancers, sole proprietors, and independent contractors.

3. [3]

   About 82% of the 34.8 million small businesses in the United States fall into this category, according to the U.S.

4. [4]

   Small Business Administration’s Office of Advocacy.

5. [5]

   “Because solopreneurs may not have cybersecurity expertise, we wanted to create a resource designed for them,” said NIST’s … More → The post NIST publishes cybersecurity guide for businesses too small to have IT staff appeared first on Help Net Security .

Read the original article →

What happened

Container images may include unused operating-system packages, development tools, debugging utilities and libraries, increasing the components that teams must investigate, patch or explain during audits. [1] [2] [3]

The article gives an NGINX image as an example containing 274 categorized vulnerabilities immediately after download. [4]

Why it matters

Unnecessary image components can expand the attack surface and add vulnerabilities requiring investigation or remediation. [2] [3]

Cited source evidence

These excerpts support the numbered claims above. They come from one source article and do not establish independent corroboration.

1. [1]

   Teams identify vulnerabilities, but turning those findings into lower risk is much harder.

2. [2]

   A standard container image may include operating-system packages, development tools, debugging utilities and libraries that an application never uses.

3. [3]

   Each unnecessary component can expand the attack surface and add vulnerabilities to investigate, patch or explain during an audit.

4. [4]

   Consider an NGINX image that contains 274 categorized vulnerabilities straight from download.

Read the original article →

What happened

Ghostunnel is an open-source proxy that provides TLS for programs unable to handle TLS themselves. [1]

Ghostunnel supports server and client modes. [2]

In server mode, Ghostunnel accepts TLS connections, checks client certificates, and forwards traffic to a backend over TCP or a UNIX socket. [3]

In client mode, Ghostunnel accepts plain local connections, wraps them in TLS, and presents a client certificate to the remote endpoint. [4]

Cited source evidence

These excerpts support the numbered claims above. They come from one source article and do not establish independent corroboration.

1. [1]

   Ghostunnel is an open-source proxy that handles TLS for programs that cannot handle it themselves.

2. [2]

   It works in two directions.

3. [3]

   In server mode it sits in front of a backend, accepts TLS connections, checks the client’s certificate, then passes traffic to the service behind it over TCP or a UNIX socket.

4. [4]

   In client mode it does the reverse, listening for plain local connections, wrapping them in TLS and presenting a client certificate to the remote … More → The post Ghostunnel: Open-source TLS proxy with mutual authentication appeared first on Help Net Security .

Read the original article →

What happened

In a Help Net Security interview, Lindsay Kaye of HUMAN Security discusses what AI-generated malware looks like when examined experimentally. [1]

Kaye identifies tidy comments, uniform logging messages, and emojis as signs that a script may have been generated by a model, while noting that decompilation removes much of this evidence. [2]

Why it matters

The interview also addresses reduced patching time for defenders and the use of AI to scale fraud by criminals. [3]

Cited source evidence

These excerpts support the numbered claims above. They come from one source article and do not establish independent corroboration.

1. [1]

   In this Help Net Security interview, Lindsay Kaye, VP of Threat Intelligence at HUMAN Security, talks about what AI-generated malware looks like on the bench.

2. [2]

   She points to tidy comments, uniform logging messages, and emojis as signs a script came from a model, and notes that decompiled code strips much of that away.

3. [3]

   Kaye also covers how AI is squeezing the time teams have to patch, where it helps criminals scale fraud, and why her … More → The post What AI-generated malware looks like under the microscope appeared first on Help Net Security .

Read the original article →

What happened

The rsync project released version 3.5.2, which fixes a flaw allowing a connecting client to consume server CPU with ineffective checksum calculations. [1]

rsync is an open-source utility used to copy files locally and across a network. [2]

Version 3.5.2 adds an option to place a transfer at the back of the CPU and disk queue, reducing competition from large synchronizations. [3]

Cited source evidence

These excerpts support the numbered claims above. They come from one source article and do not establish independent corroboration.

1. [1]

   The rsync project released version 3.5.2, and it closes a flaw that let a connecting client tie up a server’s CPU with checksum math that leads nowhere.

2. [2]

   rsync is an open-source utility that copies files locally and across a network, and it is widely used … More → The post Open-source file sync tool rsync 3.5.2 fixes a CPU exhaustion flaw appeared first on Help Net Security .

3. [3]

   The new version also lets you push a transfer to the back of the queue for CPU and disk, so a big sync stops elbowing everything else on the machine.

Read the original article →

What happened

Google said a critical Android flaw could let a remote attacker run code on a phone without user action; the flaw is CVE-2026-58868 and affects Android 14 through 17. [1] [2]

The fix is included in security patch level 2026-11-01 or later, and phone makers and carriers distribute the patch on their own schedules. [3]

Known limitations

The available source is truncated and does not state what Google’s bulletin reported as of November 4. [4]

Cited source evidence

These excerpts support the numbered claims above. They come from one source article and do not establish independent corroboration.

1. [1]

   A critical Android flaw could allow an attacker to run code on a phone without any user action remotely, Google said in its November security bulletin.

2. [2]

   The flaw, CVE-2026-58868, affects Android 14 through 17.

3. [3]

   Phones get the fix with security patch level 2026-11-01 or later, which phone makers and carriers send on their own schedules.

4. [4]

   As of November 4, Google's bulletin did not report

Read the original article →

What happened

CVE-2026-58868 is a critical Android heap buffer overflow in FrameDecoder.cpp’s onOutputReceived function, affecting Android versions 14 through 17. [1] [2] [3]

The flaw can cause an out-of-bounds write and potentially enable remote arbitrary-code execution without user interaction or additional execution privileges. [4] [5]

Google has fixed CVE-2026-58868; devices receive the fix through security patch level 2026-11-01 or later, depending on manufacturers and carriers. [2] [6]

Separately, CVE-2026-58882 could reportedly let an attacker remotely disable a phone by sending a specially crafted message; exploitation reportedly requires no user interaction and only the victim’s phone number. [8] [9] [10]

A technical analysis by security researcher Mark B reportedly reproduced crashes on several Android models; the crash can terminate ongoing calls, including emergency calls, and may recur until the offending message is removed. [11] [12] [13] [14] [15]

Why it matters

Google’s November 2026 bulletin says it is aware of active exploitation of the flaws covered in that bulletin. [7]

Known limitations

Google’s bulletin and the CVE record identify decodeWapPdu, but do not explicitly confirm that CVE-2026-58882 is the same issue described in the researcher’s SMS of Death report. [16] [17]

The source says the technical overlap warrants investigation, but the reports should not be treated as definitively describing the same vulnerability without further confirmation. [18]

Cited source evidence

These excerpts support the numbered claims above. They come from one source article and do not establish independent corroboration.

1. [1]

   Critical Android flaw CVE-2026-58868 could allow remote code execution without user interaction on devices running Android 14 through 17.

2. [2]

   Google has fixed CVE-2026-58868, a critical Android vulnerability affecting versions 14 through 17.

3. [3]

   The flaw is a critical Android vulnerability caused by a heap buffer overflow in the onOutputReceived function of FrameDecoder.cpp .

4. [4]

   The flaw can trigger an out-of-bounds write, potentially allowing a remote attacker to execute arbitrary code on a vulnerable device.

5. [5]

   Exploitation requires neither user interaction nor additional execution privileges, making it a potential zero-click attack.

6. [6]

   User interaction is not needed for exploitation.” Devices receive the fix through security patch level 2026-11-01 or later, depending on manufacturers and carriers.

7. [7]

   Google is now aware of active exploitation of the flaws in its November 2026 bulletin.

8. [8]

   One of the critical issues fixed by Google is CVE-2026-58882 (aka SMS of Death ), which could allow attackers to remotely disable a phone by sending a specially crafted SMS message.

9. [9]

   The vulnerability is particularly concerning because exploitation reportedly requires no user interaction.

10. [10]

    An attacker only needs the victim’s phone number to send a malicious message.

11. [11]

    According to a technical analysis published on October 29 by security researcher Mark B, a malformed WAP Push message can crash com.android.phone , the Android system component responsible for handling telephony functions.

12. [12]

    The crash can terminate ongoing calls, including emergency calls.

13. [13]

    Tests on several Android phone models reportedly reproduced the crash consistently.

14. [14]

    The phone can also become stuck in a crash loop because Android retries processing the malicious message whenever the device starts.

15. [15]

    As a result, the device may continue crashing until the offending message is removed.

16. [16]

    There is an important distinction, however.

17. [17]

    Google’s security bulletin and the CVE record identify the vulnerable decodeWapPdu function, but neither explicitly confirms that CVE-2026-58882 is the same issue described in the researcher’s SMS of Death report.

18. [18]

    The technical overlap is worth investigating, but the two reports shouldn’t be treated as definitively describing the same vulnerability without further confirmation.

Read the original article →

What happened

A tampered software installer and a fake job interview reportedly provided attackers access used to steal about $6.34 billion in digital assets from Binance. [1]

Binance characterized the incident as more than a single isolated breach or simple phishing incident. [2]

The incident report attributed the attack to combined techniques including fake-recruitment social engineering, malicious software downloads, compromised internal AI tooling, endpoint compromise, and an advanced transaction-authorisation method. [3]

Cited source evidence

These excerpts support the numbered claims above. They come from one source article and do not establish independent corroboration.

1. [1]

   A tampered software installer and a fake job interview gave attackers the access they needed to steal about $6.34 billion in digital assets from Binance.

2. [2]

   “It was not a single isolated breach or a simple phishing incident,” Binance wrote in its incident report.

3. [3]

   “The attackers combined a range of techniques, including social engineering by fake recruitment interviews, malicious software downloads, compromised internal AI tooling, endpoint compromise, and an advanced transaction-authorisation method.” The money left Binance … More → The post Fake job interview opened the door to $6.34 billion Binance heist appeared first on Help Net Security .

Read the original article →

What happened

An attacker compromised cPGuard’s update infrastructure and distributed a malicious update to nearly 3,000 customer Linux web-hosting servers. [1]

OPSShield confirmed that customer contact information and hashed passwords were stolen from its billing and licensing server. [2]

The attacker uploaded fake cPGuard updates identified as versions 5.88.12 through 5.88.19. [3]

Cited source evidence

These excerpts support the numbered claims above. They come from one source article and do not establish independent corroboration.

1. [1]

   An attacker has compromised the update infrastructure of cPGuard, a security suite for Linux web hosting servers, and pushed out a malicious update to nearly 3,000 customer servers.

2. [2]

   OPSShield, the company behind the security suite, also confirmed that the attacker made off with customer contact information and hashed passwords from the firm’s billing and licensing server.

3. [3]

   How the attack unfolded The attacker uploaded fake cPGuard updates – versions 5.88.12 through 5.88.19 – to OPSShield’s update … More → The post Poisoned cPGuard update hits thousands of web hosting servers appeared first on Help Net Security .

Read the original article →

What happened

Two Xen hypervisor flaws allow a malicious guest VM to read and write the memory of other VMs on the same host, including dom0. [1] [2]

All Xen versions from 4.6 onward are vulnerable. [3]

Patches are available, and the Xen Project advisory reported no known mitigation. [4]

Cited source evidence

These excerpts support the numbered claims above. They come from one source article and do not establish independent corroboration.

1. [1]

   Two flaws in the Xen hypervisor allow a malicious guest virtual machine (VM) to read and write the memory of other VMs on the same host, including dom0.

2. [2]

   Dom0 is the VM that, by default, has full permissions and manages the others.

3. [3]

   All Xen versions from 4.6 onward are vulnerable,

4. [4]

   Patches are out, and there is no known mitigation, the Xen Project said in an advisory on November 3.

Read the original article →

What happened

A critical flaw in Ruby on Rails Active Storage could let an unauthenticated attacker write arbitrary files to an application server by uploading a crafted file when the application processes uploads with ImageMagick. [1] [2]

The Rails project disclosed the flaw and fixed it in Rails versions 8.1.4.1 and 8.0.5.2. [3]

Cited source evidence

These excerpts support the numbered claims above. They come from one source article and do not establish independent corroboration.

1. [1]

   A critical flaw in Ruby on Rails could allow an unauthenticated attacker to write arbitrary files on an app's server by uploading a crafted file if the app processes uploads using ImageMagick.

2. [2]

   The flaw is in Active Storage, the part of Rails that handles file uploads.

3. [3]

   The Rails project disclosed the flaw in an advisory and fixed it in versions 8.1.4.1 and 8.0.5.2.

Read the original article →

What happened

Impart Security launched the Runtime Defense Harness, a controlled environment where AI agents write and test virtual patches for new vulnerabilities without placing AI in the traffic path. [1]

Within the harness, agents build each virtual patch against a customer’s own traffic; an engineer approves it, and Impart’s engine enforces it as deterministic code. [2]

The source says security teams can receive a tested virtual patch as soon as 60 minutes after an unspecified event or condition; the supplied text is truncated and does not identify what starts that interval. [4]

Why it matters

The launch is presented in the context of exploits increasingly outpacing fixes. [3]

Cited source evidence

These excerpts support the numbered claims above. They come from one source article and do not establish independent corroboration.

1. [1]

   Impart Security has launched the Runtime Defense Harness, a controlled environment where AI agents write and test virtual patches for new vulnerabilities without putting AI in the traffic path.

2. [2]

   In the harness, agents build each virtual patch against the customer’s own traffic, an engineer approves it, and Impart’s engine enforces it as deterministic code.

3. [3]

   The launch comes as exploits increasingly outpace fixes.

4. [4]

   Security teams get a tested virtual patch as soon as 60 minutes after … More → The post Impart uses AI to build virtual patches without putting models in the traffic path appeared first on Help Net Security .

Read the original article →

What happened

CVE-2000-0411 affected a CGI-bin script used to send e-mail; adding URL parameters could leak environment variables. [1] [2] [3]

Related vulnerabilities could let attackers set the e-mail “To” address, and such flaws were heavily abused by spammers. [4]

Why it matters

Passing parameters directly to the sendmail binary could, in similar vulnerabilities, allow arbitrary code execution. [5]

Cited source evidence

These excerpts support the numbered claims above. They come from one source article and do not establish independent corroboration.

1. [1]

   One of the early web application vulnerabilities I remember was CVE-2000-0411 [CVE].

2. [2]

   This was a simple vulnerability in a CGI-bin script used to send e-mail.

3. [3]

   The vulnerability was pretty straightforward: Adding additional parameters to the URL leaked environment variables.

4. [4]

   Other vulnerabilities allowed setting the "To" address and were heavily abused by spammers.

5. [5]

   And if parameters were passed directly to the sendmail binary, similar vulnerabilities could allow arbitrary code execution.

Read the original article →

What happened

Three blood banks detected and reversed a potentially destructive temperature spike in cold rooms after a cyberattack targeted an EFS refrigeration-equipment provider. [1]

The French Blood Establishment (EFS) is described as France’s only civilian blood collection and transfusion organization, but it was not directly targeted. [1] [2]

Why it matters

Monitoring systems detected anomalies that resulted in changed operating parameters for equipped cold rooms; the source states the temperature spike was reversed in time. [1] [2]

Cited source evidence

These excerpts support the numbered claims above. They come from one source article and do not establish independent corroboration.

1. [1]

   Three blood banks have detected and reversed a potentially destructive temperature spike in their cold rooms, following a cyberattack that targeted a refrigeration equipment provider of the French Blood Establishment (EFS), the only civilian blood collection and transfusion organization in France.

2. [2]

   Monitoring systems caught the anomalies in time “Although the EFS was not directly targeted by this cybersecurity incident, anomalies were reported, resulting in changes to the operating parameters of cold rooms equipped with the … More → The post Cyberattack hits French blood banks appeared first on Help Net Security .

Read the original article →

What happened

ISC2 will host a live Q&A about the CISSP exam on November 6, 2026, at 7:00pm CET. [1]

A panel of experts will answer questions, explain the certification requirements, and review the exam domains. [2]

The event will cover self-study resources and training options for exam preparation. [3]

The listed learning topics include CISSP’s end-to-end foundational cybersecurity understanding and the demand for vendor-neutral certification. [4]

Cited source evidence

These excerpts support the numbered claims above. They come from one source article and do not establish independent corroboration.

1. [1]

   Join ISC2 on November 6, 2026 at 7:00pm CET for a live Q&A about the CISSP exam.

2. [2]

   A panel of experts will answer your questions, explain what it takes to earn the certification, and walk through the exam domains.

3. [3]

   You’ll also learn about self-study resources and training options to help you prepare for exam day.

4. [4]

   You’ll learn: How CISSP builds your end-to-end foundational understanding of cybersecurity Why vendor-neutral certification is in demand What training tools … More → The post Live Q&A: CISSP exam prep with ISC2 experts appeared first on Help Net Security .

Read the original article →

What happened

BlueFlag Security says its expanded AI Agent Governance maps access, behavior, and oversight for approved and shadow AI agents, with options for automated risk remediation. [1]

The capabilities govern AI agents and their relationships with human and machine identities, tools, pipelines, and production applications across the software development lifecycle. [2]

Why it matters

The controls are presented as addressing AI agents operating under developer identities and spanning software-development environments. [1] [2]

Known limitations

The supplied spans do not specify deployment requirements, supported products or integrations, remediation conditions, or independent validation of the capabilities. [1] [2]

Cited source evidence

These excerpts support the numbered claims above. They come from one source article and do not establish independent corroboration.

1. [1]

   BlueFlag Security has expanded AI Agent Governance that fully maps the access, behavior and oversight of both approved and shadow AI agents, including options for automated risk remediation.

2. [2]

   Activated within BlueFlag’s pioneering platform, customers can now govern all aspects of AI agents and their relationships with human/machine identities, tools, pipelines and production applications across the software development lifecycle (SDLC).

Read the original article →

What happened

Descope announced Envoy, capabilities intended to give personal AI agents a delegated sign-in path alongside a consumer app’s existing authentication flows. [1]

The announcement says consumer companies can accept agents as first-class users with distinct identities, permissions, and approval flows without rebuilding their identity stack. [2]

Why it matters

GoFundMe, You.com, GoodRx, and Composio are identified as launch and design partners that will work with Descope to improve Envoy in the coming months. [3]

Cited source evidence

These excerpts support the numbered claims above. They come from one source article and do not establish independent corroboration.

1. [1]

   Descope has announced Envoy, a set of capabilities that gives personal AI agents a delegated sign-in path alongside a consumer app’s existing authentication flows.

2. [2]

   Consumer companies can now accept agents as first-class users—with their own identities, permissions, and approval flows—without rebuilding their identity stack.

3. [3]

   GoFundMe, You.com, GoodRx, and Composio are among several launch and design partners for Envoy who will work alongside Descope to improve Envoy in the months to come.

Read the original article →

What happened

Sekoia reported that Russia’s APT29 used compromised websites, lookalike domains, and a hijacked web analytics service to target iOS and Android users with exploits for patched flaws, stealing cookies, messages, and Wi‑Fi passwords. [1]

Sekoia’s Threat Detection & Research team spotted the watering-hole operation in mid-October 2026; it targeted mobile devices with n-day exploits. [2]

Known limitations

The available excerpt says Sekoia assessed the watering-hole campaign with moderate to high confidence, but the assessment sentence is truncated and provides no further operational detail. [3]

Cited source evidence

These excerpts support the numbered claims above. They come from one source article and do not establish independent corroboration.

1. [1]

   Russia’s APT29 has been using compromised websites, lookalike domains and a hijacked web analytics service to target iOS and Android users with exploits for patched flaws and steal their cookies, messages and Wi-Fi passwords, according to Sekoia.

2. [2]

   APT29 watering holes target mobile devices with n-day exploits (Source: Sekoia) The company’s Threat Detection & Research (TDR) team spotted the operation in mid-October 2026.

3. [3]

   “This watering hole campaign, which we assess with moderate to high confidence to … More → The post Russian hackers target iOS and Android users with n-day exploits appeared first on Help Net Security .

Read the original article →

What happened

Cycurion launched Habit, an AI-driven security operations center intended to identify threats before they disrupt operations. [1]

The launch accompanies Cycurion’s repositioning as a predictive cybersecurity company and the assembly of a team including former U.S. Intelligence Community and private-sector cyber experts. [2] [3]

Habit is described as learning normal behavior for each identity, device, and workload in a client environment, then investigating activity that breaks the pattern. [4] [5]

Known limitations

The supplied excerpt is truncated and does not provide enough detail to assess Habit’s verification process, fix validation, coverage, deployment conditions, or observed effectiveness. [6]

Cited source evidence

These excerpts support the numbered claims above. They come from one source article and do not establish independent corroboration.

1. [1]

   Cycurion has launched Habit, an AI-driven security operations center (SOC) built to identify threats before they disrupt operations.

2. [2]

   The launch also marks Cycurion’s repositioning as a predictive cybersecurity company, built around a newly assembled team of former U.S.

3. [3]

   Intelligence Community and private-sector cyber experts.

4. [4]

   Habit doesn’t chase alerts.

5. [5]

   It learns what normal looks like for every identity, device, and workload in a client’s environment, then goes after whatever breaks the pattern.

6. [6]

   It connects to the … More → The post Cycurion Habit verifies threats and validates fixes before closing cases appeared first on Help Net Security .

Read the original article →

What happened

Abraham Teklu Lemma, a U.S. naturalized citizen and IT specialist with Top Secret clearance, pleaded guilty to transmitting sensitive documents to Ethiopia and agreed to a 120–262 month prison term. [1] [2]

Lemma worked for multiple U.S. government organizations from 2019 to 2023, including the National Reconnaissance Office, State Department, and Justice Department. [3] [4]

He was authorized to move classified files between classified systems and move unclassified files from classified to unclassified systems, but access to classified files on non-State Department systems required written approval tied to official duties. [5] [6]

Court documents stated that Lemma accessed intelligence reports without need-to-know authorization, copied information from at least 133 reports, downloaded more than 100 documents, removed classification markings, and carried discs or handwritten notes out of a SCIF. [7] [8] [9] [10] [11] [12]

Prosecutors claimed Lemma transmitted information to Ethiopian officials through an encrypted messaging app beginning around August or September 2022; the material included documents, photographs, notes, maps, and information about military and logistics sites. [13] [14]

Why it matters

Prosecutors claimed Ethiopian officials remotely accessed copies of classified files stored on Lemma’s personal drives several times between April and June 2023. [15]

The Justice Department said many of Lemma’s contacts were senior Ethiopian officials in political, military, and intelligence branches, and an FBI official said the information was used in military operations during an ongoing conflict. [16] [17] [18]

Known limitations

The complaint initially placed the alleged offending’s start in July 2022, while the Justice Department later said it began in 2021. [19]

Cited source evidence

These excerpts support the numbered claims above. They come from one source article and do not establish independent corroboration.

1. [1]

   An IT specialist who worked across US government departments and held Top Secret clearance has pleaded guilty to transmitting sensitive documents to Ethiopia.

2. [2]

   Abraham Teklu Lemma, 53, admitted to the espionage offenses and agreed to a prison term between 120 and 262 months, the Justice Department confirmed on Tuesday.

3. [3]

   A naturalized citizen of the US, Lemma, of Maryland, had worked for the US government in different capacities and across different teams, including for the National Reconnaissance Office, State Department (DoS), and Justice Department (DoJ), between 2019 and 2023.

4. [4]

   These roles included working evening hours as an IT help desk technician at the DoS as of 2021, and daytime work at the DoJ as a contract management analyst, according to court documents.

5. [5]

   Lemma had what's known as "downdraft" authority, which is to say that he was permitted to move classified files between classified systems, and unclassified files from classified to unclassified systems.

6. [6]

   Crucially, he was only allowed to access classified files on non-DoS systems with written approval that it was part of his official duties, and worked in a Sensitive Compartmented Information Facility (SCIF) – a secure workplace built for handling classified files.

7. [7]

   It claims Lemma was caught trawling "several" non-DoS classified portals and accessed intelligence reports without the proper authorization to do so.

8. [8]

   On various occasions, Lemma, who had held Top Secret security clearance since 2020, copied information from at least 133 intelligence reports without need-to-know authorization, according to court documents.

9. [9]

   Prosecutors claimed that the man downloaded more than 100 documents, burned them to CD/DVDs, and deleted their classification markings.

10. [10]

    The material Lemma was copying was at the Secret and Top Secret levels, but burned onto discs with green labels, which are approved for employees and are supposed to signify the contents of the discs relating to unclassified material only.

11. [11]

    Lemma would allegedly take these discs at the end of his shift and exit the SCIF with them tucked into his jacket pocket.

12. [12]

    On other occasions, he was observed looking at classified material on his PC, then taking handwritten notes before folding the paper and again storing it in his pocket.

13. [13]

    Prosecutors claimed that Lemma began transmitting information to Ethiopian officials via an encrypted messaging app in or around August and September 2022, weeks after the alleged offending began.

14. [14]

    The files included documents, photographs, notes, maps, and information relating to Ethiopia, and officials would ask Lemma to provide details about command centers, logistics centers, military compounds, and other matters.

15. [15]

    Lemma was also accused of storing copies of classified files on his own storage drives, which prosecutors claim were accessed remotely by Ethiopian officials numerous times between April and June 2023.

16. [16]

    Announcing the guilty plea on Tuesday, the Justice Department said many of Lemma’s contacts were senior Ethiopian officials across political, military, and intelligence branches of the government.

17. [17]

    "Despite swearing an oath to defend the Constitution of the United States, Abraham Lemma betrayed our country by committing espionage on behalf of the Ethiopian government," said Roman Rozhavsky, assistant director at the FBI's Counterintelligence and Espionage Division.

18. [18]

    "By his own admission, Lemma provided classified national defense information to Ethiopian officials, who then used it in military operations during an ongoing conflict.

19. [19]

    The criminal complaint [PDF] initially stated that the alleged offending began in July 2022, although the DoJ more recently claimed that it began in 2021.

Read the original article →

What happened

Cybersecurity researchers disclosed an OpenAI Codex jailbreak bundle designed to modify an AI agent’s trusted instructions to insert advertisements for a third-party AI gateway. [1]

Check Point Research gave the phenomenon the codename “prompt adware.” [2]

The jailbreak is called Haiou, meaning “Seagull” in Chinese, and modifies OpenAI Codex’s configuration. [3]

Cited source evidence

These excerpts support the numbered claims above. They come from one source article and do not establish independent corroboration.

1. [1]

   Cybersecurity researchers have disclosed details of an OpenAI Codex jailbreak bundle that's designed to modify an artificial intelligence (AI) agent's trusted instructions to insert advertisements for a third-party AI gateway.

2. [2]

   The phenomenon has been codenamed prompt adware by Check Point Research.

3. [3]

   Called Haiou (meaning Seagull in Chinese), the jailbreak modifies OpenAI Codex's configuration

Read the original article →

What happened

SonicWall warned of two vulnerabilities affecting end-of-life Global Management System (GMS) systems and said they will not be patched. [1]

CVE-2026-109593 affects SonicWall GMS 9.5.2 and earlier versions and has a CVSS score of 9.1. [2]

CVE-2026-109593 is an unauthenticated file-read vulnerability that could allow an attacker with network access to read internal GMS files. [2]

Cited source evidence

These excerpts support the numbered claims above. They come from one source article and do not establish independent corroboration.

1. [1]

   SonicWall has warned of two vulnerabilities impacting end-of-life Global Management System (GMS) systems that it said will not be patched.

2. [2]

   Details of the security flaws, which affect SonicWall GMS 9.5.2 and earlier versions, are below - CVE-2026-109593 (CVSS score: 9.1) - An unauthenticated file read vulnerability that could allow an attacker with network access to read internal GMS files,

Read the original article →

What happened

APT, the package manager for Debian and Ubuntu, has a flaw allowing an attacker who can modify package downloads to replace a checked package with attacker-controlled content. [1]

The attacker’s code runs as root, according to advisories published by Debian, Ubuntu, and APT developers on November 4. [2]

The attack requires APT to retry failed downloads; the supplied text states this behavior is enabled by default from Debian 12, but the sentence is truncated. [3]

Cited source evidence

These excerpts support the numbered claims above. They come from one source article and do not establish independent corroboration.

1. [1]

   A flaw in APT, the package manager for Debian and Ubuntu, allows an attacker who can modify package downloads to swap a checked package for their own.

2. [2]

   The attacker's code then runs as root, according to advisories that Debian, Ubuntu, and APT's developers published on November 4.

3. [3]

   The attack needs APT to retry failed downloads, which it does by default from Debian 12 and

Read the original article →