Customer portal
Articles Tagged with

SOS Intelligence

"FLASH
Flash Alert

FLASH Alert – Information Disclosure vulnerability in Check Point’s Quantum Gateway

CVE-2024-24919

CVSS 7.5 HIGH (Provisional)

On 27 May 2024, Check Point disclosed a vulnerability impacting the following products:

  • CloudGuard Network
  • Quantum Maestro
  • Quantum Scalable Chassis
  • Quantum Security Gateways
  • Quantum Spark Appliances

CVE-2024-24919 is an information disclosure vulnerability which would allow an unauthenticated threat actor to read certain information on Check Point Security Gateways once connected to the internet and enabled with remote Access VPN or Mobile Access Software Blades.

The following versions are known to be affected:

  • R77.20 (EOL)
  • R77.30 (EOL)
  • R80.10 (EOL)
  • R80.20 (EOL)
  • R80.20.x
  • R80.20SP (EOL)
  • R80.30 (EOL)
  • R80.30SP (EOL)
  • R80.40 (EOL)
  • R81, R81.10
  • R81.10.x
  • R81.20

The vulnerability is exploitable on affected systems if ONE of the following conditions is met:

  • The IPsec VPN Blade is enabled, but ONLY when included in the Remote Access VPN  community.
  • The Mobile Access Software Blade is enabled.

Check Point has issued detailed instructions for applying hotfixes to affected services to mitigate this vulnerability.  Additionally, The following has also been recommended:

  • Change the password of the Security Gateway’s account in Active Directory
  • Prevent Local Accounts from connecting to VPN with Password Authentication

The announcement of this vulnerability comes after Check Point identified a small number of login attempts on older local VPN accounts that used an unrecommended password-only authentication method.  This indicates that the vulnerability is being exploited in the wild, and so the recommended hotfixes should be applied as soon as practicable.

"China
Investigation

China – A Step Ahead In Digital Espionage

In the digital age, data has emerged as one of the most valuable resources, driving economies, shaping public opinion, and determining the success of nations. Amid this reality, cybercrime has become a potent tool for state actors, with China often cited as a significant player in the realm of cyber espionage and cybercrime. This article delves into how China has allegedly used cybercrime to obtain data, the motivations behind these actions, their methods, and the implications on global geopolitics.

UPDATE – join us on the 13th June for the accompanying webinar.

The Who – Those Working In The Shadows

On the digital battlefield,  whether state-sponsored or self-motivated hacker, anonymity is key.  This makes the task of attributing the activity of threat actors to real-world identities that much harder.  More often than not, we see the evidence of digital crime, and can use available intelligence to make best estimates of a culprit, but a threat actor who wants to remain anonymous can do so with a reasonable application of effort. However, despite these efforts, identification of threat actors and attribution of criminal activity can be possible.

China’s cyber activities are primarily conducted by state-sponsored groups. These groups, often referred to as Advanced Persistent Threats (APTs), include:

APT 1

APT1, also known as the Comment Crew or Shanghai Group, is a highly active cyber espionage unit linked to the Chinese military, specifically PLA Unit 61398. Identified by cybersecurity firm Mandiant in a 2013 report, APT1 is known for targeting a wide array of industries, including information technology, aerospace, telecommunications, and scientific research.

Their primary method of infiltration involves spear-phishing emails, followed by deploying custom and publicly available malware to maintain access and exfiltrate sensitive data. The group’s activities have largely focused on U.S.-based organisations, aiming to steal intellectual property and trade secrets to benefit Chinese companies and government entities.

APT 10

APT10, also known as Stone Panda or MenuPass Group, is a cyber espionage group attributed to the Chinese government. The group has been active since at least 2009 and is known for targeting managed IT service providers (MSPs) and their clients across various industries, including healthcare, aerospace, and manufacturing. APT10’s operations typically involve sophisticated tactics such as spear-phishing, the use of custom malware, and leveraging legitimate credentials to infiltrate networks and exfiltrate data. Their focus on MSPs allows them to gain access to multiple organisations through a single breach, maximising the impact of their espionage efforts.

APT10’s activities have had significant global repercussions, prompting extensive investigations and responses from cybersecurity firms and government agencies. In December 2018, the U.S. Department of Justice indicted two Chinese nationals associated with APT10, accusing them of stealing sensitive data from dozens of companies and government agencies.

APT 31

APT31, also known as Zirconium, Judgment Panda, or Bronze Vinewood, is a Chinese state-sponsored cyber espionage group. The group is known for its advanced and persistent cyber operations targeting a wide range of sectors, including government, finance, technology, and aerospace. APT31 employs sophisticated tactics such as spear-phishing, supply chain attacks, and the deployment of custom malware to infiltrate and maintain access to targeted networks. Their primary goal is to steal sensitive information and intellectual property to support Chinese national interests and provide strategic advantages.

The activities of APT31 have significant global implications, prompting extensive countermeasures from affected organisations and governments. Notably, in 2020, APT31 was linked to cyberattacks targeting the U.S. presidential election campaign, highlighting the group’s capability and intent to influence political processes.

APT 41

APT41, also known as Winnti, Barium, or Wicked Panda, is a Chinese state-sponsored cyber threat group known for its dual role in cyber espionage and financially motivated cybercrime. Active since at least 2012, APT41 targets a wide range of sectors, including healthcare, telecommunications, finance, and video game industries. The group employs diverse tactics, techniques, and procedures (TTPs), such as spear-phishing, supply chain compromises, and the use of custom malware to infiltrate networks. APT41 is particularly notable for its ability to pivot from traditional espionage activities to financially driven attacks, including ransomware and cryptocurrency mining.

The activities of APT41 have led to significant economic and security repercussions globally. In September 2020, the U.S. Department of Justice charged five Chinese nationals associated with APT41 with hacking into over 100 companies and entities worldwide.

These groups are composed of highly skilled hackers and often operate under the direction of the Chinese government, particularly the Ministry of State Security (MSS) and the People’s Liberation Army (PLA).

The What & The Why – China’s Motivations For Stealing Data

“Know yourself and know your enemy, and you shall never be defeated.”

Chinese Advanced Persistent Threats (APTs) target a wide range of data across various sectors. The specific data targeted and stolen can vary depending on the APT group and their specific objectives, but generally includes the following types:

  1. Intellectual Property (IP) and Trade Secrets:
    • Technological innovations: This includes sensitive information from sectors where technological innovation is key, such as aerospace (e.g., designs for new aircraft or satellite technology), biotechnology (e.g., genetic research), semiconductors (e.g., chip designs), and automotive (e.g., electric vehicle technology). The aim is often to reduce the time and cost associated with research and development by acquiring innovations from other nations.
    • Manufacturing processes: This encompasses proprietary methods, production techniques, and formulas used in manufacturing. For example, a pharmaceutical company’s proprietary process for producing a drug or an electronics company’s methods for fabricating microchips.
  1. Corporate Data:
  1. Strategic plans: Corporate strategies can include market expansion plans, new product launches, or competitive tactics. Accessing this information gives competitors an unfair advantage.
  2. Client and partner information: Information about key clients, partners, and their contracts or negotiations can be exploited to undercut or sabotage business deals.
  3. Employee data: Personal information about employees, such as social security numbers, addresses, and employment history, can be used for targeted attacks or to compromise individuals who hold critical positions within an organisation.
  1. Government and Military Information:
  1. Defence and military secrets: This includes detailed information about defence systems, weapons designs, military operational plans, and intelligence reports. Such data is critical for national security and military advantage.
  2. Diplomatic communications: Sensitive communications between diplomats, government officials, and international bodies. This can provide insights into negotiation tactics, foreign policy strategies, and international relations.
  1. Healthcare Data:
  1. Patient records: Patient data includes medical histories, diagnoses, treatments, and personal identification information. This data is valuable not only for identity theft but also for crafting highly targeted social engineering attacks.
  2. Medical research: Data from clinical trials and research into new treatments and drugs is invaluable for both economic and public health reasons. Stealing this data can provide a competitive edge in the pharmaceutical industry.
  1. Financial Data:
  1. Banking information: Includes account numbers, transaction histories, credit card information, and other financial records. This data can be used for financial fraud or to gain insights into the financial health of organisations.
  2. Payment systems: Information related to the security and operation of payment processing systems, such as those used in banking and retail. Compromising these systems can lead to large-scale financial theft or disruption.
  1. Energy and Infrastructure Data:
  1. Operational data: Details about the daily operations of critical infrastructure such as power grids, water supply systems, and telecommunications networks. This information can be used to disrupt services or to understand and replicate operational efficiencies.
  2. Designs and security details: Blueprints and security protocols for infrastructure facilities, which can be used to plan attacks or unauthorised access.
  1. Academic and Research Data:
  1. Scientific research: Data from academic research projects, particularly those in cutting-edge fields like artificial intelligence, quantum computing, and nanotechnology. This can accelerate a nation’s technological progress by acquiring the latest scientific breakthroughs.
  2. Educational resources: Curricula, exam results, and other educational materials can be used to understand and influence the educational standards and outputs of other countries.

The Where – Understanding Which Nations Are Targeted

Chinese Advanced Persistent Threat (APT) groups, which are often associated with state-sponsored cyber activities, have targeted a wide range of countries over the years. Some of their primary targets include:

  1. United States:
    • Chinese APT groups have consistently targeted U.S. government agencies, including defence, diplomatic, and intelligence entities, to gather political and military intelligence.
    • Additionally, they have sought to steal intellectual property from U.S. corporations, particularly in the technology, aerospace, healthcare, and energy sectors.
    • Some notable incidents include the hacking of the Office of Personnel Management (OPM) in 2015, which compromised the sensitive personal data of millions of federal employees, and the targeting of defence contractors involved in sensitive military projects.
  1. European Countries:
  1. European nations have been targeted for intellectual property theft, economic espionage, and political influence operations.
  2. Chinese APT groups have focused on stealing cutting-edge technology, research, and development data from industries such as aerospace, automotive, telecommunications, and pharmaceuticals.
  3. European governments and diplomatic institutions have also been targeted for intelligence gathering and monitoring political developments.
  1. Asian Countries:
  1. China’s regional rivals, such as Japan and South Korea, have been targeted for political and military intelligence gathering, as well as stealing advanced technology.
  2. Countries like India have experienced cyber intrusions aimed at accessing sensitive government information, military strategies, and technological advancements.
  3. Southeast Asian nations have been targeted for economic espionage, particularly related to infrastructure projects, natural resources, and geopolitical influence.
  1. Taiwan:
  1. Due to the ongoing political tensions between China and Taiwan, Taiwanese government agencies, defence contractors, and organisations have been frequent targets of Chinese cyber espionage.
  2. The aim is to gather intelligence on Taiwan’s defence capabilities, political developments, and cross-strait relations.
  1. Australia:
  1. Australian government institutions, defence contractors, and companies across various sectors have been targeted for intellectual property theft, economic espionage, and monitoring of political developments.
  2. Notable incidents include cyber intrusions targeting universities and research institutions to steal sensitive research data and technology.
  1. Canada:
  1. Canadian government agencies, particularly those involved in defence, foreign affairs, and natural resources, have been targeted for intelligence gathering.
  2. Chinese APT groups have also targeted Canadian companies in sectors such as aerospace, telecommunications, and mining for economic espionage purposes.
  1. Africa and Latin America:
  1. While less extensively reported, there have been instances of Chinese cyber espionage targeting countries in Africa and Latin America.
  2. These activities often revolve around gaining access to natural resources, monitoring infrastructure projects, and influencing political developments in alignment with China’s strategic interests.

Overall, Chinese APT groups demonstrate a global reach in their cyber operations, driven by motivations such as geopolitical competition, economic advantage, and technological advancement. They employ sophisticated techniques to infiltrate networks, exfiltrate data, and maintain persistent access for intelligence gathering and other strategic objectives.

The When – A Timeline of Chinese Threat Actor Activity

“If you know the enemy and know yourself, you need not fear the result of a hundred battles.”

For 2500 years, the teachings of Sun Tzu have been the backbone of Chinese military doctrine.  As his text, The Art of War, has become more popularised, its teachings have been applied across other walks of life, particularly business and governance.  Chinese offensive cyber policy has also followed these principles, and what we have previously discussed shows how China is utilising cyber crime activity to gather information and intelligence to satisfy a broad range of objectives.

Over the years, the finger of blame has been levelled at China for some of the biggest data breaches and incidents of corporate espionage.  We look at some of these below:

Chinese Data Breaches

The How – Common TTPs Utilised By Chinese Threat Actors

Chinese Advanced Persistent Threat (APT) groups employ various sophisticated techniques to steal data from targeted organisations. Their methods often involve multiple stages, including reconnaissance, initial compromise, establishing a foothold, escalating privileges, internal reconnaissance, data exfiltration, and covering their tracks. Here are some common techniques and tactics used by Chinese APTs:

  1. Reconnaissance

Chinese APTs conduct thorough reconnaissance to tailor their attacks effectively:

  • Open Source Intelligence (OSINT): Gathering information from social media platforms, corporate websites, and public records to identify key personnel and network architecture.
  • Phishing Campaigns: Utilising spear-phishing emails targeting specific individuals within an organisation to collect credentials or deliver malware. For example, APT41 has been known to send emails mimicking trusted contacts or business partners.
  1. Initial Compromise

Common methods for initial network penetration by Chinese APTs include:

  • Spear-Phishing Emails: Highly targeted emails containing malicious attachments or links. APT10 frequently used this method to deliver malware like PlugX or Poison Ivy.
  • Exploiting Zero-Day Vulnerabilities: Identifying and exploiting vulnerabilities before they are publicly known. APT3, for instance, has leveraged zero-days in widely used software such as Adobe Flash and Internet Explorer.
  • Supply Chain Attacks: Compromising software updates or hardware components. APT41 has been implicated in attacks on software supply chains, embedding malware in legitimate software updates.
  1. Establishing a Foothold

Once access is gained, Chinese APTs work to maintain a persistent presence:

  • Malware Deployment: Installing Remote Access Trojans (RATs) like Sakula, used by APT10, or variants of the Cobalt Strike framework employed by APT41.
  • Setting Up Command and Control (C2) Channels: Creating secure channels to communicate with infected systems. APT41 often uses DNS tunnelling and HTTP/S protocols to evade detection.
  1. Privilege Escalation

To gain higher privileges, Chinese APTs use various techniques:

  • Credential Dumping: Tools like Mimikatz are frequently used by groups such as APT41 to extract credentials from Windows systems.
  • Exploiting Privilege Escalation Vulnerabilities: Utilising known vulnerabilities in operating systems and applications. APT3 has exploited vulnerabilities in Windows to escalate privileges and move laterally within networks.
  1. Internal Reconnaissance

Mapping the internal network to locate valuable data involves:

  • Network Scanning: Using tools like Nmap to identify live hosts and services. APT10 often employs custom network scanning tools.
  • Lateral Movement: Utilising credentials and tools like PsExec or WMI to move across the network. APT41 is known for its proficiency in lateral movement, using legitimate administrative tools to avoid detection.
  1. Data Exfiltration

Stealing data while avoiding detection is critical:

  • Data Compression and Encryption: Compressing and encrypting data to expedite transfer and evade detection. APT10 has been known to use tools like WinRAR for compression and encryption.
  • Steganography: Embedding data within other files or images. APT groups may use steganography to hide data within innocuous files.
  • Covert Channels: Employing techniques like DNS tunnelling or HTTPS to transfer data. APT41, for example, has used custom protocols to exfiltrate data over HTTPS.
  1. Covering Tracks

Chinese APTs employ various methods to avoid detection and analysis:

  • Log Deletion and Manipulation: Removing or altering logs to erase evidence of their activities. APT10 has been observed cleaning up after themselves by deleting logs and temporary files.
  • Use of Proxy Chains: Routing traffic through multiple compromised systems to obscure the origin of their actions. APT41 often uses a series of compromised machines to route their traffic, making it difficult to trace.
  • Anti Forensic Techniques: Using tools to thwart forensic investigations, such as wiping tools or encrypting malware payloads. APT3 has been known to employ these techniques to hinder analysis.

In Conclusion

China’s use of cybercrime to obtain data is a testament to the strategic importance of information in the modern world. As China continues to leverage cyber capabilities to advance its national interests, the global community faces the challenge of balancing technological advancement with security and ethical considerations.

The ongoing cyber skirmishes highlight the need for robust international norms and cooperation to address the complexities of cyber espionage and cybercrime, ensuring a secure and stable digital future for all.

By understanding the scope, motivations, and methods behind China’s cyber activities, the international community can better prepare and respond to the evolving landscape of cyber warfare. As data becomes increasingly integral to national security and economic prosperity, safeguarding it against state-sponsored cybercrime will be crucial in maintaining global stability and trust in the digital age.

The future of cybersecurity will depend on collective efforts to strengthen defences, establish clear policies, and foster international collaboration to mitigate the risks posed by cyber espionage and cybercrime.

UPDATE – join us on the 13th June for the accompanying webinar.

Further Reading

UK Electoral Commission Breach

https://www.bbc.co.uk/news/uk-politics-68652374

MOD Payroll Breach

https://www.bbc.co.uk/news/uk-68967805

How does China use it’s data

https://www.nzz.ch/english/how-does-china-use-the-personal-data-it-steals-ld.1828192

https://www.forbes.com/sites/heatherwishartsmith/2023/11/04/trafficking-data-chinas-digital-sovereignty-and-its-control-of-your-data/?sh=2b78939543a4

F22/F35 Program Breaches

https://www.sandboxx.us/news/the-man-who-stole-americas-stealth-fighters-for-china

Photo by Li Yang on Unsplash.

"SOS
CVE Top 10

The SOS Intelligence CVE Chatter Weekly Top Ten – 27 May 2024

 

This weekly blog post is from via our unique intelligence collection pipelines. We are your eyes and ears online, including the Dark Web.

There are thousands of vulnerability discussions each week. SOS Intelligence gathers a list of the most discussed Common Vulnerabilities and Exposures (CVE) online for the previous week.

We make every effort to ensure the accuracy of the data presented. As this is an automated process some errors may creep in.

If you are feeling generous please do make us aware of anything you spot, feel free to follow us on Twitter @sosintel and DM us. Thank you!

 


 

1.  CVE-2024-4761

Out of bounds write in V8 in Google Chrome prior to 124.0.6367.207 allowed a remote attacker to perform an out of bounds memory write via a crafted HTML page. (Chromium security severity: High)

https://nvd.nist.gov/vuln/detail/CVE-2024-4761

 


 

2. CVE-2024-20356

A vulnerability in the web-based management interface of Cisco Integrated Management Controller (IMC) could allow an authenticated, remote attacker with Administrator-level privileges to perform command injection attacks on an affected system and elevate their privileges to root. This vulnerability is due to insufficient user input validation. An attacker could exploit this vulnerability by sending crafted commands to the web-based management interface of the affected software. A successful exploit could allow the attacker to elevate their privileges to root.

https://nvd.nist.gov/vuln/detail/CVE-2024-20356

 


 

3. CVE-2024-4671

Use after free in Visuals in Google Chrome prior to 124.0.6367.201 allowed a remote attacker who had compromised the renderer process to potentially perform a sandbox escape via a crafted HTML page. (Chromium security severity: High)

https://nvd.nist.gov/vuln/detail/CVE-2024-4671

 


 

4. CVE-2024-4947

Type Confusion in V8 in Google Chrome prior to 125.0.6422.60 allowed a remote attacker to execute arbitrary code inside a sandbox via a crafted HTML page. (Chromium security severity: High)

https://nvd.nist.gov/vuln/detail/CVE-2024-4947

 


 

5. CVE-2023-2551

PHP Remote File Inclusion in GitHub repository unilogies/bumsys prior to 2.1.1.

https://nvd.nist.gov/vuln/detail/CVE-2023-2551

 


 

6. CVE-2023-43770

Roundcube before 1.4.14, 1.5.x before 1.5.4, and 1.6.x before 1.6.3 allows XSS via text/plain e-mail messages with crafted links because of program/lib/Roundcube/rcube_string_replacer.php behavior.

https://nvd.nist.gov/vuln/detail/CVE-2023-43770

 


 

7. CVE-2024-1630

Path traversal vulnerability in “getAllFolderContents” function of Common Service Desktop, a GE HealthCare ultrasound device component

https://nvd.nist.gov/vuln/detail/CVE-2024-1630

 


 

8. CVE-2022-23940

SuiteCRM through 7.12.1 and 8.x through 8.0.1 allows Remote Code Execution. Authenticated users with access to the Scheduled Reports module can achieve this by leveraging PHP deserialization in the email_recipients property. By using a crafted request, they can create a malicious report, containing a PHP-deserialization payload in the email_recipients field. Once someone accesses this report, the backend will deserialize the content of the email_recipients field and the payload gets executed. Project dependencies include a number of interesting PHP deserialization gadgets (e.g., Monolog/RCE1 from phpggc) that can be used for Code Execution.

https://nvd.nist.gov/vuln/detail/CVE-2022-23940

 


 

9. CVE-2024-1628

OS command injection vulnerabilities in GE HealthCare ultrasound devices

https://nvd.nist.gov/vuln/detail/CVE-2024-1628

 


 

10. CVE-2024-1629

Path traversal vulnerability in “deleteFiles” function of Common Service Desktop, a GE HealthCare ultrasound device component

https://nvd.nist.gov/vuln/detail/CVE-2024-1629

 


"SOS
Ransomware

Ransomware – State of Play April 2024

SOS Intelligence is currently tracking 192 distinct ransomware groups, with data collection covering 382 relays and mirrors.

In the reporting period, SOS Intelligence has identified 365 instances of publicised ransomware attacks.  These have been identified through the publication of victim details and data on ransomware blog sites accessible via Tor. Our analysis is presented below:

Group Activity and Trends

Ransomware activity showed a 13% decrease in April when compared to the previous month, and a 7 % decrease in activity when compared to the previous year. However, the number of active groups has increased to 36 from 33 the previous month.

The overall drop in victim numbers for April is likely an ongoing effect of the dissolution of AlphV/BlackCat and the significant decrease in activity from Lockbit as a result of law enforcement activity in February.

Since February, we have closely monitored group activity for signs of where AlphV and Lockbit affiliates would take their business. The top six groups for the year-to-date are represented above and as yet, no one group has emerged above the others. Hunters International, Play and Ransomhub established themselves as the most active across April, but over the three months, we have also seen significant activity from Blackbasta and 8base. This could suggest that displaced affiliates are not settled on a final product, and have been utilising different ransomware services in the wake of the downfall of AlphV and Lockbit.

Analysis of Geographic Targeting

The volume of targeting against US-based victims has remained steady at around 50% of all reported ransomware attacks.  Targeting continues to follow financial lines, with the majority of remaining attacks targeted at G7 and BRICS bloc countries.

Compared to March, 11% fewer countries were targeted in April.  Our data is also showing interesting geographic targeting data.  We have observed emerging or developing strains targeting developing countries, whereas more established variants focus more on North America, Western Europe and Australia.

Top Strains per Country

United States
Canada
United Kingdom
Germany
Italy
– play
– play
– snatch
– ragroup
– ransomhub
– hunters
– blacksuit
– dragonforce
– 8base
– rhysida
– blacksuit
– akira
– lockbit3
– lockbit3
– ciphbit

Industry Targeting

Despite a reduction in victim volume, Manufacturing and IT & Technology remain at the forefront of threat actor targeting. Health & Social Care and Retail & Wholesale continue to see an emergence as a target of choice amongst multiple different variants, likely due to many groups removing targeting restrictions in the wake of law enforcement activity and continued western support for Ukraine.

Top Strains per Industry

Manufacturing
IT & Technology
Health & Social Care
Construction & Engineering
Retail & Wholesale
– play
– ransomhub
– incransom
– play
– hunters
– hunters
– darkvault
– qiulong
– cactus
– ransomhub
– blackbasta
– cactus
– ransomhub
– lockbit3
– lockbit3

Significant Events

8base targets the United Nations

The United Nations Development Programme (UNDP) was subject to an 8base ransomware attack, resulting in the exfiltration of human resources and procurement information. Despite significant demands being made, the UN has stood fast in its decision to not make payment.

Akira collects ransoms worth USD 42 million

An advisory provided cyber security centre’s in the USA, Netherlands and Europe has revealed that, since March 2023, the Akira ransomware strain has been responsible for attacks against 250 victims, with an estimated total ransom value of USD 42 million.

Lockbit not disappearing without a fight

The District of Columbia Department of Insurance, Securities & Banking,a local government department in the US capital, was added to the long list of Lockbit victims.  An estimated 800GB of sensitive data was obtained in the breach, which has not been made available to the public amid reports of it being sold privately.

New Groups

APT73

  • Suspected to be a LockBit spin-off – several pages on their leaksite resemble those used by LockBit
  • Listed 4 victims since appearing in late April

DarkVault

  • Suspected to be a LockBit spin-off – several pages on their leaksite resemble those used by LockBit
  • Also involved in other illicit activities, such as bomb threats, doxing, and fraud.
  • Listed 22 victims since appearing in April

Quilong

  • Currently exclusively targeting victims in Brazil
  • Listed 6 victims since appearing in April

SEXi

  • Emerged in April 2024, targeting a hosting company in Chile. 
  • Encrypts VMware ESXi servers and backups, appending the .SEXi extension to encrypted files and dropping ransom notes named SEXi.txt. The name ‘SEXi’ is believed to be a play on ‘ESXi,’ as the attacks exclusively target VMWare ESXi servers.

Space Bears

  • Sports a unique front end with corporate stock images but also maintains a classic “wall of shame” for their victims.
  • Alongside instructions for affected companies, they operate both a .onion site and a clearnet website.

Vulnerability Exploitation

Threat actors are maintaining techniques focusing on the exploitation of vulnerabilities in public-facing corporate infrastructure.  

In recent weeks, Linux variants of the Cerber ransomware have been seen to be deployed utilising exploitation of Atlassian Confluence Data Center and Server, specifically CVE-2023-22518.  CVE-2023-22518 is a critical severity (CVSS 9.1) Improper Authorisation Vulnerability which allows an unauthenticated attacker to reset Confluence and create an administrator account for persistent access.

"SOS
CVE Top 10

The SOS Intelligence CVE Chatter Weekly Top Ten – 20 May 2024

 

This weekly blog post is from via our unique intelligence collection pipelines. We are your eyes and ears online, including the Dark Web.

There are thousands of vulnerability discussions each week. SOS Intelligence gathers a list of the most discussed Common Vulnerabilities and Exposures (CVE) online for the previous week.

We make every effort to ensure the accuracy of the data presented. As this is an automated process some errors may creep in.

If you are feeling generous please do make us aware of anything you spot, feel free to follow us on Twitter @sosintel and DM us. Thank you!

 


 

1.  CVE-2024-30055

Microsoft Edge (Chromium-based) Spoofing Vulnerability

https://nvd.nist.gov/vuln/detail/CVE-2024-30055

 


 

2. CVE-2024-20356

A vulnerability in the web-based management interface of Cisco Integrated Management Controller (IMC) could allow an authenticated, remote attacker with Administrator-level privileges to perform command injection attacks on an affected system and elevate their privileges to root. This vulnerability is due to insufficient user input validation. An attacker could exploit this vulnerability by sending crafted commands to the web-based management interface of the affected software. A successful exploit could allow the attacker to elevate their privileges to root.

https://nvd.nist.gov/vuln/detail/CVE-2024-20356

 


 

3. CVE-2023-41266

A path traversal vulnerability found in Qlik Sense Enterprise for Windows for versions May 2023 Patch 3 and earlier, February 2023 Patch 7 and earlier, November 2022 Patch 10 and earlier, and August 2022 Patch 12 and earlier allows an unauthenticated remote attacker to generate an anonymous session. This allows them to transmit HTTP requests to unauthorized endpoints. This is fixed in August 2023 IR, May 2023 Patch 4, February 2023 Patch 8, November 2022 Patch 11, and August 2022 Patch 13.

https://nvd.nist.gov/vuln/detail/CVE-2023-41266

 


 

4. CVE-2023-2551

PHP Remote File Inclusion in GitHub repository unilogies/bumsys prior to 2.1.1.

https://nvd.nist.gov/vuln/detail/CVE-2023-2551

 


 

5. CVE-2023-7101

Spreadsheet::ParseExcel version 0.65 is a Perl module used for parsing Excel files. Spreadsheet::ParseExcel is vulnerable to an arbitrary code execution (ACE) vulnerability due to passing unvalidated input from a file into a string-type “eval”. Specifically, the issue stems from the evaluation of Number format strings (not to be confused with printf-style format strings) within the Excel parsing logic.

https://nvd.nist.gov/vuln/detail/CVE-2023-7101

 


 

6. CVE-2024-0519

Out of bounds memory access in V8 in Google Chrome prior to 120.0.6099.224 allowed a remote attacker to potentially exploit heap corruption via a crafted HTML page. (Chromium security severity: High)

https://nvd.nist.gov/vuln/detail/CVE-2024-0519

 


 

7. CVE-2024-1628

OS command injection vulnerabilities in GE HealthCare ultrasound devices

https://nvd.nist.gov/vuln/detail/CVE-2024-1628

 


 

8. CVE-2024-21887

A command injection vulnerability in web components of Ivanti Connect Secure (9.x, 22.x) and Ivanti Policy Secure (9.x, 22.x) allows an authenticated administrator to send specially crafted requests and execute arbitrary commands on the appliance.

https://nvd.nist.gov/vuln/detail/CVE-2024-21887

 


 

9. CVE-2021-45046

It was found that the fix to address CVE-2021-44228 in Apache Log4j 2.15.0 was incomplete in certain non-default configurations. This could allows attackers with control over Thread Context Map (MDC) input data when the logging configuration uses a non-default Pattern Layout with either a Context Lookup (for example, $${ctx:loginId}) or a Thread Context Map pattern (%X, %mdc, or %MDC) to craft malicious input data using a JNDI Lookup pattern resulting in an information leak and remote code execution in some environments and local code execution in all environments. Log4j 2.16.0 (Java 8) and 2.12.2 (Java 7) fix this issue by removing support for message lookup patterns and disabling JNDI functionality by default.

https://nvd.nist.gov/vuln/detail/CVE-2021-45046

 


 

10. CVE-2021-44228

Apache Log4j2 2.0-beta9 through 2.15.0 (excluding security releases 2.12.2, 2.12.3, and 2.3.1) JNDI features used in configuration, log messages, and parameters do not protect against attacker controlled LDAP and other JNDI related endpoints. An attacker who can control log messages or log message parameters can execute arbitrary code loaded from LDAP servers when message lookup substitution is enabled. From log4j 2.15.0, this behavior has been disabled by default. From version 2.16.0 (along with 2.12.2, 2.12.3, and 2.3.1), this functionality has been completely removed. Note that this vulnerability is specific to log4j-core and does not affect log4net, log4cxx, or other Apache Logging Services projects.

https://nvd.nist.gov/vuln/detail/CVE-2021-44228

 


"SOS
CVE Top 10

The SOS Intelligence CVE Chatter Weekly Top Ten – 13 May 2024

 

This weekly blog post is from via our unique intelligence collection pipelines. We are your eyes and ears online, including the Dark Web.

There are thousands of vulnerability discussions each week. SOS Intelligence gathers a list of the most discussed Common Vulnerabilities and Exposures (CVE) online for the previous week.

We make every effort to ensure the accuracy of the data presented. As this is an automated process some errors may creep in.

If you are feeling generous please do make us aware of anything you spot, feel free to follow us on Twitter @sosintel and DM us. Thank you!

 


 

1.  CVE-2024-20356

A vulnerability in the web-based management interface of Cisco Integrated Management Controller (IMC) could allow an authenticated, remote attacker with Administrator-level privileges to perform command injection attacks on an affected system and elevate their privileges to root. This vulnerability is due to insufficient user input validation. An attacker could exploit this vulnerability by sending crafted commands to the web-based management interface of the affected software. A successful exploit could allow the attacker to elevate their privileges to root.

https://nvd.nist.gov/vuln/detail/CVE-2024-20356

 


 

2. CVE-2024-3661

DHCP can add routes to a client’s routing table via the classless static route option (121). VPN-based security solutions that rely on routes to redirect traffic can be forced to leak traffic over the physical interface. An attacker on the same local network can read, disrupt, or possibly modify network traffic that was expected to be protected by the VPN.

https://nvd.nist.gov/vuln/detail/CVE-2024-3661

 


 

3. CVE-2020-13699

TeamViewer Desktop for Windows before 15.8.3 does not properly quote its custom URI handlers. A malicious website could launch TeamViewer with arbitrary parameters, as demonstrated by a teamviewer10: –play URL. An attacker could force a victim to send an NTLM authentication request and either relay the request or capture the hash for offline password cracking. This affects teamviewer10, teamviewer8, teamviewerapi, tvchat1, tvcontrol1, tvfiletransfer1, tvjoinv8, tvpresent1, tvsendfile1, tvsqcustomer1, tvsqsupport1, tvvideocall1, and tvvpn1. The issue is fixed in 8.0.258861, 9.0.258860, 10.0.258873, 11.0.258870, 12.0.258869, 13.2.36220, 14.2.56676, 14.7.48350, and 15.8.3.

https://nvd.nist.gov/vuln/detail/CVE-2020-13699

 


 

4. CVE-2024-27956

Improper Neutralization of Special Elements used in an SQL Command (‘SQL Injection’) vulnerability in ValvePress Automatic allows SQL Injection.This issue affects Automatic: from n/a through 3.92.0.

https://nvd.nist.gov/vuln/detail/CVE-2024-27956

 


 

5. CVE-2024-26304

There is a buffer overflow vulnerability in the underlying L2/L3 Management service that could lead to unauthenticated remote code execution by sending specially crafted packets destined to the PAPI (Aruba’s access point management protocol) UDP port (8211). Successful exploitation of this vulnerability results in the ability to execute arbitrary code as a privileged user on the underlying operating system.

https://nvd.nist.gov/vuln/detail/CVE-2024-26304

 


 

6. CVE-2024-21762

A out-of-bounds write in Fortinet FortiOS versions 7.4.0 through 7.4.2, 7.2.0 through 7.2.6, 7.0.0 through 7.0.13, 6.4.0 through 6.4.14, 6.2.0 through 6.2.15, 6.0.0 through 6.0.17, FortiProxy versions 7.4.0 through 7.4.2, 7.2.0 through 7.2.8, 7.0.0 through 7.0.14, 2.0.0 through 2.0.13, 1.2.0 through 1.2.13, 1.1.0 through 1.1.6, 1.0.0 through 1.0.7 allows attacker to execute unauthorized code or commands via specifically crafted requests

https://nvd.nist.gov/vuln/detail/CVE-2024-21762

 


 

7. CVE-2024-3273

** UNSUPPORTED WHEN ASSIGNED ** A vulnerability, which was classified as critical, was found in D-Link DNS-320L, DNS-325, DNS-327L and DNS-340L up to 20240403. Affected is an unknown function of the file /cgi-bin/nas_sharing.cgi of the component HTTP GET Request Handler. The manipulation of the argument system leads to command injection. It is possible to launch the attack remotely. The exploit has been disclosed to the public and may be used. The identifier of this vulnerability is VDB-259284. NOTE: This vulnerability only affects products that are no longer supported by the maintainer. NOTE: Vendor was contacted early and confirmed immediately that the product is end-of-life. It should be retired and replaced.

https://nvd.nist.gov/vuln/detail/CVE-2024-3273

 


 

8. CVE-2023-49606

A use-after-free vulnerability exists in the HTTP Connection Headers parsing in Tinyproxy 1.11.1 and Tinyproxy 1.10.0. A specially crafted HTTP header can trigger reuse of previously freed memory, which leads to memory corruption and could lead to remote code execution. An attacker needs to make an unauthenticated HTTP request to trigger this vulnerability.

https://nvd.nist.gov/vuln/detail/CVE-2023-49606

 


 

9. CVE-2021-43008

Improper Access Control in Adminer versions 1.12.0 to 4.6.2 (fixed in version 4.6.3) allows an attacker to achieve Arbitrary File Read on the remote server by requesting the Adminer to connect to a remote MySQL database.

https://nvd.nist.gov/vuln/detail/CVE-2021-43008

 


 

10. CVE-2023-27997

A heap-based buffer overflow vulnerability [CWE-122] in FortiOS version 7.2.4 and below, version 7.0.11 and below, version 6.4.12 and below, version 6.0.16 and below and FortiProxy version 7.2.3 and below, version 7.0.9 and below, version 2.0.12 and below, version 1.2 all versions, version 1.1 all versions SSL-VPN may allow a remote attacker to execute arbitrary code or commands via specifically crafted requests.

https://nvd.nist.gov/vuln/detail/CVE-2023-27997

 


"SOS
CVE Top 10

The SOS Intelligence CVE Chatter Weekly Top Ten – 06 May 2024

 

This weekly blog post is from via our unique intelligence collection pipelines. We are your eyes and ears online, including the Dark Web.

There are thousands of vulnerability discussions each week. SOS Intelligence gathers a list of the most discussed Common Vulnerabilities and Exposures (CVE) online for the previous week.

We make every effort to ensure the accuracy of the data presented. As this is an automated process some errors may creep in.

If you are feeling generous please do make us aware of anything you spot, feel free to follow us on Twitter @sosintel and DM us. Thank you!

 


 

1.  CVE-2024-20356

A vulnerability in the web-based management interface of Cisco Integrated Management Controller (IMC) could allow an authenticated, remote attacker with Administrator-level privileges to perform command injection attacks on an affected system and elevate their privileges to root. This vulnerability is due to insufficient user input validation. An attacker could exploit this vulnerability by sending crafted commands to the web-based management interface of the affected software. A successful exploit could allow the attacker to elevate their privileges to root.

https://nvd.nist.gov/vuln/detail/CVE-2024-20356

 


 

2. CVE-2024-21893

A server-side request forgery vulnerability in the SAML component of Ivanti Connect Secure (9.x, 22.x) and Ivanti Policy Secure (9.x, 22.x) and Ivanti Neurons for ZTA allows an attacker to access certain restricted resources without authentication.

https://nvd.nist.gov/vuln/detail/CVE-2024-21893

 


 

3. CVE-2023-23752

An issue was discovered in Joomla! 4.0.0 through 4.2.7. An improper access check allows unauthorized access to webservice endpoints.

https://nvd.nist.gov/vuln/detail/CVE-2023-23752

 


 

4. CVE-2023-43770

Roundcube before 1.4.14, 1.5.x before 1.5.4, and 1.6.x before 1.6.3 allows XSS via text/plain e-mail messages with crafted links because of program/lib/Roundcube/rcube_string_replacer.php behavior.

https://nvd.nist.gov/vuln/detail/CVE-2023-43770

 


 

5. CVE-2024-1709

ConnectWise ScreenConnect 23.9.7 and prior are affected by an Authentication Bypass Using an Alternate Path or Channel

vulnerability, which may allow an attacker direct access to confidential information or

critical systems.

https://nvd.nist.gov/vuln/detail/CVE-2024-1709

 


 

6. CVE-2023-7028

An issue has been discovered in GitLab CE/EE affecting all versions from 16.1 prior to 16.1.6, 16.2 prior to 16.2.9, 16.3 prior to 16.3.7, 16.4 prior to 16.4.5, 16.5 prior to 16.5.6, 16.6 prior to 16.6.4, and 16.7 prior to 16.7.2 in which user account password reset emails could be delivered to an unverified email address.

https://nvd.nist.gov/vuln/detail/CVE-2023-7028

 


 

7. CVE-2023-41266

A path traversal vulnerability found in Qlik Sense Enterprise for Windows for versions May 2023 Patch 3 and earlier, February 2023 Patch 7 and earlier, November 2022 Patch 10 and earlier, and August 2022 Patch 12 and earlier allows an unauthenticated remote attacker to generate an anonymous session. This allows them to transmit HTTP requests to unauthorized endpoints. This is fixed in August 2023 IR, May 2023 Patch 4, February 2023 Patch 8, November 2022 Patch 11, and August 2022 Patch 13.

https://nvd.nist.gov/vuln/detail/CVE-2023-41266

 


 

8. CVE-2023-41265

An HTTP Request Tunneling vulnerability found in Qlik Sense Enterprise for Windows for versions May 2023 Patch 3 and earlier, February 2023 Patch 7 and earlier, November 2022 Patch 10 and earlier, and August 2022 Patch 12 and earlier allows a remote attacker to elevate their privilege by tunneling HTTP requests in the raw HTTP request. This allows them to send requests that get executed by the backend server hosting the repository application. This is fixed in August 2023 IR, May 2023 Patch 4, February 2023 Patch 8, November 2022 Patch 11, and August 2022 Patch 13.

https://nvd.nist.gov/vuln/detail/CVE-2023-41265

 


 

9. CVE-2024-23222

A type confusion issue was addressed with improved checks. This issue is fixed in iOS 17.3 and iPadOS 17.3, macOS Sonoma 14.3, tvOS 17.3. Processing maliciously crafted web content may lead to arbitrary code execution. Apple is aware of a report that this issue may have been exploited.

https://nvd.nist.gov/vuln/detail/CVE-2024-23222

 


 

10. CVE-2023-6548

Improper Control of Generation of Code (‘Code Injection’) in NetScaler ADC and NetScaler Gateway allows an attacker with access to NSIP, CLIP or SNIP with management interface to perform Authenticated (low privileged) remote code execution on Management Interface.

https://nvd.nist.gov/vuln/detail/CVE-2023-6548

 


"SOS
Investigation, Product news

Cracking CAPTCHAs for fun and profit

Through synthetic training sample dataset generation and ML training.

Preface

Cracking CAPTCHAs is already a well-documented and established process which this article looks to expand on. We will approach this article with a general view of how we’ve cracked CAPTCHAs within undesirable conditions. This article is not meant to be a how-to or detailed guide to replicate our steps. However, it may give you some inspiration for your specific challenge. 

We believe that the methods laid out in this article are novel and significantly improve the efficiency of automated CAPTCHA solving in contrast to traditional approaches. Especially when considering a target CAPTCHA system with poor sample harvesting opportunities.

Ethics

We bypass human verification checks to maintain automatic information collection pipelines. The use of the methods we have developed only extends as far as what is required to automate our collection process. 

If a CAPTCHA or other human verification check system is poorly designed and not adequately rate limited, condition checked etc. bypassing it on scale may lead to a DDoS (Distributed Denial of Service) attack in the worst of cases. But with correctly implemented human verification systems, you should mitigate this even with the system bypassed. At best, unethical manipulation of these verification systems can lead to spam posts/comments and otherwise undesirable automated “bot” interaction. We do not condone this type of use. 

The Problem

There are several well-established methods to automate the solving of CAPTCHAs, depending on the complexity of the CAPTCHA, and if we start at the easy end of the spectrum we are presented with a fairly basic alphabetical captcha. 

With a simple distortion background, one might choose to apply a straightforward process of applying denoise filters or Gaussian blurring to an image to reduce or remove the amount of “stars” or random dot pixels present in its background that are applied at random. 

This process can give us a less noisy picture and we can further convert the image to grayscale.  If the source sample is a colour image doing so improves edge detection. 

The image can then be processed through a standard OCR (Optical Character Recognition) library and in our experience can result in a 0.1% failure rate yielding excellent stable solutions. 

In some cases, a good test of CAPTCHA ease of solvability is to feed it to Google Translate as an image; have Google Translate attempt to read the text and translate the letters back into English. If it can, then you have a very good chance that rudimentary OCR libraries will also work for you.

But this article is not about the easy end of the challenge…

What we are dealing with is a CAPTCHA that is both alphanumeric, upper and lower case with random character placement and rotation, and random disruption lines across the image and characters.  Furthermore, most importantly, a point that we will discuss in more detail is where the target source is a Tor Onion website that, at the best of times loads slowly and at the worst of times is offline or responds with backend timeout errors. 

The image complexity of the source CAPTCHA means it’s nearly impossible to effectively read it by OCR. This is made challenging due to the disruption patterns provided by the background random line arrangement (an outward star pattern) and each of our characters are independently disrupted with seemingly random lines of various length and width. Combining all that with offset angles of each character it’s beyond what most OCR or OpenCV methods can handle. 

Therefore, for more complex CAPTCHAs image manipulation (removing noise, grey scaling etc.) is typically not sufficient. These challenges usually require machine learning to get a reasonable failure rate and sufficient solving speed. 

The biggest factor in achieving a good model that will solve accurately is having a large enough sample base. In some cases, many thousands of samples are required for training. Certainly, when dealing with a CAPTCHA that may have upper, lowercase and numerical characters with randomisation of all these points plus randomisation on disruption patterns or lines the larger the sample set, the more accurate a model the training will produce. 

So how do you get thousands of samples from a source that is slow to load and has poor availability, both conditions of the source being a Tor website? Harvesting samples this way would be far too inefficient and we can’t hang around! 

Even with a target source that responds reasonably quickly, has good availability, and can be harvested without aggressively hitting rate limits, who would want to sit there endlessly solving eight thousand captchas to feed to an optical character recognition model? 

I know that’s not going to be me! Sure, there are options to outsource these problems and crowdsource them, but those options take time, money and are likely to introduce errors in our training sample data. Neither of these is desirable, so how do we get 100% accurate sample data cheaply without human solving, without having to harvest the source, and that can scale? 

The Solution

The solution we came up with was first to not focus on the solving of the CAPTCHAs, or the training of our model, or anything that was a direct result or outcome of the end goal we are driving towards. Instead, we looked at how the CAPTCHAs are constructed; what do they look like and what are their elemental parts. 

We know harvesting is not an optimal option, so we have put that aside. Doing so leaves us with a handful of maybe 20 or so harvested solved CAPTCHA samples. Nowhere near enough to start training but it’s enough to start focusing on the sample set we have.

If we look at how the CAPTCHA is constructed and try and break its construction down piece by piece, in a way “reverse engineering” the construction of the CAPTCHA we might either: 1) be able to generate our own `synthetic` CAPTCHAs on demand and at scale all 100% accurately pre solved, or 2) sufficiently understand the method of construction to identify the library or process in which the CAPTCHA is constructed and reimplement it for ourselves with the same 100% accurately pre-solved outcome. 

In our case and the example, we are writing this article from the path of the former option. This option was chosen as some time was spent trying to identify the particular CAPTCHA library but no exact match was found, and in the interest of not burning too much time, and depending on external factors we decided to attempt to create our own synthetic CAPTCHA generation process.

To create our CAPTCHAs, we used Pillow (a PIL Python Fork), a Python Image Manipulation Library that offers a wide range of features all well suited for the job at hand. 

We start by defining a few values that we have observed to be fixed, such as a defined image size (in our case, 280 by 50 pixels) and use this to create a simple image. 

Then we define our letter set (a to z, A to Z, 0 to 9) as we know these to be fixed. 

Using `random.choice` we can pick a required amount of characters.  In our case, the CAPTCHA uses a fixed length of 6 characters. 

The text font is also important and from our source samples we see it is fixed: therefore we try to match the font type as closely as possible. Font size also remains constant. This will be important in ensuring that our training is as accurate as possible when our model is presented with real sample data.

To kick things off, the process carefully establishes the dimensions of the image canvas, akin to laying out a pristine piece of paper before beginning a drawing. Then, with a deft stroke, we construct a blank background canvas, pristine and white, awaiting the arrival of the CAPTCHA characters. But here’s where the true artistry takes centre stage; the process methodically layers complexity onto the character, 

With each character in the CAPTCHA text, our process doesn’t simply slap it onto the canvas; instead, it treats each letter as an individual brushstroke, adding specific characteristics at every turn. We begin by precisely measuring the width and height of each character, ensuring that characters will not be chopped off the edges, correctly fit and fill the CAPTCHA, and that they resemble the source CAPTCHA text. Then, like with the source samples, we introduce randomness into the mix, spacing out the letters with varying degrees of separation, akin to scattering scrabble pieces.

We are also introducing a touch of chaos by randomly rotating each character, giving them a tilt that defies conventional alignment. This clever sleight of hand resembles the source samples accurately and adds to the difficulty level of solving this CAPTCHA. 

Yet the process doesn’t stop there. No, it goes above and beyond, adorning our canvas with a riotous display of crisscrossing lines, as if an abstract artist had gone wild with a brush. These random lines serve as a digital labyrinth, obscuring the text beneath a veil of confusion and intrigue.

We then add and overlay lines of random length and weight across each character, aligned to the character’s angle closely matching that of the source sample. 

Now that we have a way to populate our image canvas, we have a working framework with which we can iterate to get an output that resembles the source samples as closely as possible. 

For now, we generate a few hundred samples, each image file is named the randomly selected CAPTCHA text, assisting us by essentially generating a sample set that has already been solved. 

After that, we compared each iteration’s output closely to the source and made tweaks and adaptations. For each iteration of the CAPTCHA generator we looked closely at just one specific attribute to simplify the synthesis process. We adjust the random scattered background lines, adjusting their length, width and count.  Moving then onto tweaking the letter placement and random angles, to closely match the apparent pseudo randomness of the sample data set.

Following sufficient tweaking and iterations, we are producing a CAPTCHA that is at least visually very closely matching our source samples. It matches so closely that if mixed with real samples it’s difficult to distinguish. This is the ideal level of synthesis we are looking to achieve. 

Example synthetic captcha on the left, real on the right

Next steps

Now that we have a way to produce synthetic CAPTCHAs that very closely match our target, it’s time to produce a few thousand of them. This is easily and quickly done by specifying the total count in our process loop and out pops 5,000 freshly generated pre-solved captchas all nicely labelled and ready for shoving into our training process. 

For model training, we’ve chosen to use the TensorFlow framework alongside the ONNX Runtime machine learning model accelerator. This combination worked well for us for both training accuracy and efficiency. All training was conducted with the use of a Nvidia GPU.

Following initial training, using just our best-produced synthetic CAPTCHA samples as our data set, we achieved a CER (Character error rate) of 3.26%. For a first batch run of a model trained against a synthetic data set was not too bad at all. But we knew we could do better. 

Now that we had a model to work with, we could use it to start solving actual real target CAPTCHAs.  This would allow us to generate a larger pool of real CAPTCHA samples, with a solve set, and mix those in with our synthetic set.  We were looking to generate 5k synthetic and 1k real harvested CAPTCHAs with our newly trained, albeit unoptimized model. 

With a framework in place that would interface with the target website, collect CAPTCHAs, generate a text prediction, check that with the website and if solved, store the solved and labelled CAPTCHA image we generated about 1,000 samples over a short time.

Feeding this back into the mix of training model data we dropped the CER down to 2.77%.

A screen shot of a black screen

Description automatically generated

We were confident that even with 2.7% it was a rate better than a human could achieve, and we were also confident that our methodology was working. 

Our remaining tasks were to reiterate the model once more, using this slightly more optimised model and generate a slightly larger set of labelled real CAPTCHAs. 

We were able to go from the initial model, with a worse CER (orange line) to the best model (green line) in only a few training iterations.

The model training improvements are best shown in the graph below with each improvement yielding a lower CER, for longer (more stable) and at a sooner point in time. 

At which point we settled on a final model, with a CER of 1.4%, opting for an optimal  mix real CAPTCHAs to synthetic. 

Our final ML model diagram: 

Once the efficacy of this model was validated it was then a task of simply plugging it into the collection pipeline process and enlivening it into our production collection system. The automated solver process has been running stable ever since and most of the disruption we’ve observed has solely been to the target source going offline and being unavailable. 

Bias and Variance

A key consideration during the training process was to be aware of and mitigate where possible Overfitting and Overtraining our model. Instead of using the terms `overfitting` and `overtraining` I like to instead use Bias and Variance as two potential pitfalls of ML training as they better explain undesirable conditions that may occur. Without diving into too many details around these ML concepts as to fully understand them you would probably need a PhD. The best way I can describe what my simple mind can understand is as follows.

Due to the nature of our novel, one might say clever iterative process to train a CAPTCHA solver on a very low original source data set we are by virtue potentially adding bias into our training process. For example, from the first model any solved data sets will be solved by a model that has a predefined bias to solving a particular set, style or character combination potentially resulting in a new data set that is biassed towards what that previous model was good at solving thereby amplifying the bias in our next model’s training. 

This bias would result in a real world regression of CER as the model is unoptimised to solve a wider range of character combinations and randomisation characteristics. 

Our second pitfall: overfitting slides at both ends of the extreme in terms of providing an overly varied training set or an insufficiently varied training set, i.e. creeping into bias. Whereby we must consider that although we could train a model to solve many different types of CAPTCHAs, beyond just this one example, from one model using a very varied data set doing so and if not carefully tuned could result in `overfitting` our data set thereby introducing an unoptimised CER as our model is essentially training on more noise than signal. 

We therefore considered both Bias and Variance closely, ensuring a healthy mix of varied real correctly labelled CAPTCHAs harvested from source to a ratio of synthetically generated CAPTCHAs with a randomly distributed character set. An optimal CER band was then discovered through iterative AB testing of data set mix, training iterations until a stable plateau was identified. 

Conclusion

We deploy a final model, incorporating a mix of synthetic and real CAPTCHAs, achieving a CER of 1.4%. The automated solver process seamlessly integrates into our production collection system, ensuring stability and efficiency.

By leveraging synthetic sample training data generation, we’ve advanced CAPTCHA cracking. Our approach offers an effective and efficient solution for CAPTCHA cracking without significant human involvement or effort allowing for effective automated data collection.

With this capability, we are able to add value to our customers by automating the collection from otherwise programmatically inaccessible sources, where we would have to manually have a human solve the CAPTCHA access the page, insert any updates and then alert our customers. Automation is key to what we do at speed and at scale especially when dealing with many hundreds of collection sources as we do.

Photo by Kaffeebart on Unsplash.

"SOS
CVE Top 10

The SOS Intelligence CVE Chatter Weekly Top Ten – 29 April 2024

 

This weekly blog post is from via our unique intelligence collection pipelines. We are your eyes and ears online, including the Dark Web.

There are thousands of vulnerability discussions each week. SOS Intelligence gathers a list of the most discussed Common Vulnerabilities and Exposures (CVE) online for the previous week.

We make every effort to ensure the accuracy of the data presented. As this is an automated process some errors may creep in.

If you are feeling generous please do make us aware of anything you spot, feel free to follow us on Twitter @sosintel and DM us. Thank you!

 


 

1.  CVE-2024-29986

Microsoft Edge for Android (Chromium-based) Information Disclosure Vulnerability

https://nvd.nist.gov/vuln/detail/CVE-2024-29986

 


 

2. CVE-2024-29981

Microsoft Edge (Chromium-based) Spoofing Vulnerability

https://nvd.nist.gov/vuln/detail/CVE-2024-29981

 


 

3. CVE-2024-29991

Microsoft Edge (Chromium-based) Security Feature Bypass Vulnerability

https://nvd.nist.gov/vuln/detail/CVE-2024-29991

 


 

4. CVE-2024-29987

Microsoft Edge (Chromium-based) Information Disclosure Vulnerability

https://nvd.nist.gov/vuln/detail/CVE-2024-29987

 


 

5. CVE-2024-29049

Microsoft Edge (Chromium-based) Webview2 Spoofing Vulnerability

https://nvd.nist.gov/vuln/detail/CVE-2024-29049

 


 

6. CVE-2020-13699

TeamViewer Desktop for Windows before 15.8.3 does not properly quote its custom URI handlers. A malicious website could launch TeamViewer with arbitrary parameters, as demonstrated by a teamviewer10: –play URL. An attacker could force a victim to send an NTLM authentication request and either relay the request or capture the hash for offline password cracking. This affects teamviewer10, teamviewer8, teamviewerapi, tvchat1, tvcontrol1, tvfiletransfer1, tvjoinv8, tvpresent1, tvsendfile1, tvsqcustomer1, tvsqsupport1, tvvideocall1, and tvvpn1. The issue is fixed in 8.0.258861, 9.0.258860, 10.0.258873, 11.0.258870, 12.0.258869, 13.2.36220, 14.2.56676, 14.7.48350, and 15.8.3.

https://nvd.nist.gov/vuln/detail/CVE-2020-13699

 


 

7. CVE-2024-21412

Internet Shortcut Files Security Feature Bypass Vulnerability

https://nvd.nist.gov/vuln/detail/CVE-2024-21412

 


 

8. CVE-2022-38028

Windows Print Spooler Elevation of Privilege Vulnerability

https://nvd.nist.gov/vuln/detail/CVE-2022-38028

 


 

9. CVE-2024-0519

Out of bounds memory access in V8 in Google Chrome prior to 120.0.6099.224 allowed a remote attacker to potentially exploit heap corruption via a crafted HTML page. (Chromium security severity: High)

https://nvd.nist.gov/vuln/detail/CVE-2024-0519

 


 

10. CVE-2023-1671

A pre-auth command injection vulnerability in the warn-proceed handler of Sophos Web Appliance older than version 4.3.10.4 allows execution of arbitrary code.

https://nvd.nist.gov/vuln/detail/CVE-2023-1671

 


"SOS
Uncategorized

Ransomware – State of Play March 2024

SOS Intelligence is currently tracking 183 distinct ransomware groups, with data collection covering 368 relays and mirrors.

In the reporting period, SOS Intelligence has identified 439 instances of publicised ransomware attacks.  These have been identified through the publication of victim details and data on ransomware blog sites accessible via Tor.  Our analysis is presented below:

LockBit has maintained its position as the most active and popular ransomware strain, despite law enforcement activity against the group in February 2024.  However, we are seeing a significant decrease in their activity level, which is to be expected.  The impact of law enforcement activity against the group is still being monitored, but it has already been seen that the group has suffered significant reputation damage.  Many affiliates have lost trust in the group to keep their data safe and their identities anonymous.  

March also saw the sudden exiting of ALPHV/BlackCat from the scene, in what appeared an exit scam.  Affiliates were left stunned when the group shut up shop shortly after receiving a significant ransom from UnitedHealth Group.  As previously reported, the code for ALPHV/BlackCat was purported to have been sold, so a new group is expected to emerge using similar TTPs in due course.

As such, we have seen increases in activity amongst other high-profile groups.  Most groups have seen small increases in activity over the last month. Still, BlackBasta, Medusa, Play, and RAGroup seem to have profited most from LockBit’s misfortune and ALPHV/BlackCat’s sudden disappearance.  All have been operating for at least 12 months and have carved their own niche in the space vacated by these high-profile group.

Group targeting continues to follow familiar patterns in terms of the victim’s country of origin.

Attacks have increased in South American countries, particularly in Argentina, which may be a response to presidential elections in November 2023 in which the far-right libertarian Javier Milei was elected.  Brazil remains a popular target, as the most economically developed country in the region

Targeting continues to follow international, geopolitical lines.  Heavy targeting follows countries that have supported Ukraine against Russia.  Attacks against Sweden continued as it pressed ahead with preparations to join NATO.   This highlights the level of support ransomware groups continue to show towards the Russian state, and they will continue to use cyber crime to destabilise and weaken Western and pro-Ukrainian states.

Manufacturing and Construction & Engineering have remained the key targeted industries for March.  These industries would be more reliant on technology to continue their business activities, so it logically follows that they would be more likely to pay a ransom to regain access to compromised computer systems.  The Financial, Retail & Wholesale, Legal, and Education sectors have also seen increased activity over the period.  Health & Social Care has seen a significant increase over the period.  This is likely in response to several groups, reacting to law enforcement activity and allowing their affiliates to begin targeting these industries.

We are seeing a shift in tactics for certain industries, particularly those where data privacy carries a higher importance (such as legal or healthcare), where threat actors are not deploying encryption software and instead relying solely on data exfiltration as the main source of material for blackmail and extortion.

Significant Events

Targeting against the UK took an aggressive turn, with NHS Scotland (INC Ransomware) and media outlet The Big Issue (Qilin) amongst the most high-profile victims.  This highlights ransomware groups’ apathy towards who they target, and the secondary impacts that that targeting can have.

The Oceania arm of Nissan suffered a significant data breach, which was associated with the Akira ransomware.  The attack was limited to operations in Australia and New Zealand but did have a significant impact on distribution, marketing, sales, and services.

New Groups

March saw the emergence of three new groups; Donex, Kill Security (5 victims each) and RedRansomware (12 victims).  Kill Security has shown some aggressive public-sector targeting, including police services in India and Romania.

Vulnerability Exploitation

BianLian and Jasmine groups have been observed exploiting CVE-2024-27198 (CVSS 9.8).  This is a vulnerability in JetBrains TeamCity CI/CD server products up to version 2023.11.4, which allows a remote unauthenticated attacker to execute arbitrary code to take complete control of affected instances.  This would allow threat actors to gain access and maintain permanence within an affected network, while conducting reconnaissance, exfiltrating data, and uploading ransomware payloads.

JetBrains has implemented a fix for the impacted system, so it is advised to update to the latest available version.

1 2 4 5 6 7 8 22 23
Privacy Settings
We use cookies to enhance your experience while using our website. If you are using our Services via a browser you can restrict, block or remove cookies through your web browser settings. We also use content and scripts from third parties that may use tracking technologies. You can selectively provide your consent below to allow such third party embeds. For complete information about the cookies we use, data we collect and how we process them, please check our Privacy Policy
Youtube
Consent to display content from - Youtube
Vimeo
Consent to display content from - Vimeo
Google Maps
Consent to display content from - Google
Spotify
Consent to display content from - Spotify
Sound Cloud
Consent to display content from - Sound