Search This Blog

Powered by Blogger.

Blog Archive

Labels

About Me

Showing posts with label AWS. Show all posts

Private API Keys and Passwords Discovered in a Popular AI Training dataset

 

The Common Crawl dataset, which is used to train several artificial intelligence models, has over 12,000 legitimate secrets, including API keys and passwords. The Common Crawl non-profit organisation maintains a vast open-source archive of petabytes of web data collected since 2008, which is free to use. 

Because of the huge dataset, various artificial intelligence initiatives, including OpenAI, DeepSeek, Google, Meta, Anthropic, and Stability, may rely on the digital archive to train large language models (LLMs).

Truffle Security researchers discovered legitimate secrets after scanning 400 terabytes of data from 2.67 billion web pages in the Common Crawl December 2024 database. They uncovered 11,908 secrets that were successfully authenticated and were hardcoded by developers, highlighting that LLMs could be trained on insecure code.

It should be noted that LLM training data is not used in its raw form; instead, it is cleaned and filtered to remove extraneous content such as useless data, duplicate, malicious, or sensitive data. Despite these efforts, removing confidential data is challenging, and the method does not guarantee that all personally identifiable information (PII), financial data, medical records, and other sensitive content will be erased from the huge dataset. 

Truffle Security discovered legitimate API keys for the WalkScore, MailChimp, and Amazon Web Services (AWS) services after examining the scanned data. In the Common Crawl dataset, TruffleHog found 219 different secret kinds in total, with MailChimp API keys being the most prevalent. 

Cybersecurity researchers explain that the developers made a mistake by hardcoding them into HTML forms and JavaScript snippets rather than using server-side environment variables. An attacker could exploit these keys for nefarious purposes like phishing and brand impersonation. Furthermore, disclosing such knowledge could result in data exfiltration. Another feature of the paper is the high reuse rate of the uncovered secrets, with 63% found on several pages. 

However, a WalkScore API key "appeared 57,029 times across 1,871 subdomains." The researchers also discovered a homepage with 17 unique live Slack webhooks, which should be kept private because they allow apps to submit messages to Slack. After conducting the research, Truffle Security got in touch with the affected suppliers and collaborated with them to remove the keys belonging to their users. 

The researchers claim to have "successfully assisted those organisations collectively in rotating/revoke several thousand keys." Truffle Security's findings are a warning that insecure coding mistakes can affect the LLM's behaviour, even if an AI model uses older archives than the dataset the researchers analysed.

PowerSchool Data Breach Exposes Millions

 


An American education technology company, PowerSchool, is the latest giant to fall a victim of hacking and data breaches, which probably compromised millions of records of students and teachers in North America. As one of the leading providers of school records management software, PowerSchool serves 18,000 schools who manage data over 60 million students.


How the breach happened

The compromise was discovered on December 28 and was traced to a subcontractor's account. The new report said, however, that another incident of hacking-a compromise of the access of a PowerSchool software engineer-may have had something to do with the breach. Malware infected the engineer's computer and exfiltrated login credentials for internal systems, such as Slack, AWS, and other tools.

According to the logs retrieved by researchers, the infostealing malware known as LummaC2 was used to steal the engineer's passwords. The malware extracted saved passwords and browsing histories from the web browsers of the engineer and uploaded them to a server run by cybercriminals. The stolen credentials were shared in cybercrime groups, which further exposed PowerSchool's systems. 


What Data Was Stolen?

The hackers accessed a range of sensitive personal information, including:  

  • Social Security numbers  
  •  Student grades and demographics  
  •  Medical information  
  •  Parental access details, such as restraining orders  
  •  Records of students’ medication schedules  

School districts impacted by the breach reported that the attackers stole all historical data stored in PowerSchool’s systems.  

The lack of multi-factor authentication (MFA) on a compromised maintenance account was one key vulnerability. PowerSchool has implemented MFA and reset passwords across its customer support portal. Many of the employee credentials discovered were weak and have been exposed in other breaches.

The breach, which has underlined the threats of infostealing malware in hybrid work setups where employees operate company systems using personal devices, has left much to be expected from PowerSchool.


Response and Investigation

PowerSchool, the company concerned, is reportedly working with a cybersecurity firm named CrowdStrike for the investigation into the incident. According to them, no signs of malware have been found neither has any sign of system-layer access. But they are analyzing the stolen data.


Effects on Schools

Many school districts are operating independently to gauge the scope of the breach, relying on collective knowledge from other administrators. As the investigation continues, there are questions about PowerSchool's security measures and how it managed this extensive breach. 

Schools, parents, and educators are urged to be vigilant and ensure additional layers of security are put in place to prevent future incidents.


Volkswagen Cybersecurity Breach Exposes Sensitive Vehicle Data

 


A recent cybersecurity lapse within Volkswagen’s Cariad unit, which manages the company’s cloud systems, exposed sensitive data from hundreds of thousands of vehicles. The breach, attributed to a misconfiguration in a cloud environment hosted on Amazon Web Services (AWS), was uncovered by a whistleblower and investigated by the Chaos Computer Club, a cybersecurity association. The incident has sparked significant concerns about data privacy and the security of connected vehicles.

The exposed dataset reportedly included detailed information on approximately 800,000 electric vehicles. Notably, location data was exceptionally precise for 460,000 cars. For Volkswagen and its subsidiary Seat, the data pinpointed vehicles to within 10 centimeters, while data from Audi and Skoda vehicles were accurate to within six miles. In some instances, the leaked information was linked to personal details of car owners, such as names, contact information, and vehicle operational statuses. Alarmingly, the breach also disclosed the locations of prominent individuals, including German politicians, raising concerns about potential misuse.

Volkswagen’s Cariad unit is responsible for integrating advanced technologies into the automaker’s vehicles. This incident highlights vulnerabilities in cloud environments used by automakers to store and manage vast amounts of vehicle and customer data. According to Volkswagen, accessing the exposed information required bypassing multiple security layers, which would have demanded advanced expertise and considerable effort. Despite this, the data remained publicly accessible for several months, drawing criticism and prompting calls for stronger cybersecurity measures.

Existing Security Measures and Gaps

Automakers generally follow industry standards such as ISO/SAE 21434, which outline best practices for securing systems against breaches and mitigating vulnerabilities. Many vehicles are also equipped with cybersecurity hardware, including network switches and firewalls, to protect data within a car’s subsystems. However, the Volkswagen incident underscores critical gaps in these measures that require urgent attention.

Company Response and Moving Forward

The leaked dataset, spanning several terabytes, reportedly did not include payment details or login credentials, according to Volkswagen. The company has since patched the vulnerability and emphasized its commitment to data security. While Volkswagen stated that there was no evidence hackers had downloaded the information, the breach serves as a stark reminder of the risks inherent in managing sensitive data within interconnected systems.

This incident underscores the need for stricter regulations and enhanced cybersecurity frameworks for cloud-based infrastructures, especially as connected vehicles become increasingly prevalent. Moving forward, automakers must prioritize robust security protocols to safeguard consumer data and prevent similar breaches in the future.

Ledger Phishing Scam Targets Cryptocurrency Wallets

 


A sophisticated phishing email campaign has emerged, targeting cryptocurrency users by impersonating Ledger, a prominent hardware wallet provider. These fraudulent emails claim that the recipient’s Ledger wallet seed phrase — also known as a recovery or mnemonic seed — has been compromised. In an attempt to secure their funds, users are directed to a so-called “secure verification tool” where they are asked to confirm their seed phrase. The phishing emails appear convincing, offering a “Verify my recovery phrase” button. Clicking this button redirects victims through an Amazon Web Services (AWS) website to a fake domain, “ledger-recovery[.]info.”

Once users enter their seed phrase on this page, the attackers capture the information, granting them full access to the victims’ cryptocurrency wallets. A recovery phrase, typically consisting of 12 or 24 random words, acts as the key to accessing a wallet’s funds. The importance of keeping this phrase private and offline cannot be overstated. By stealing these phrases, the attackers gain control of the wallets and can siphon all funds, leaving victims with no recourse.

To increase the scam’s credibility, the phishing site includes several deceptive features. For example, it accepts only valid seed phrase words from a predetermined list of 2,048 options. Regardless of the entered data, the site falsely informs users that their phrase is incorrect, encouraging them to re-enter their information multiple times and ensuring the attackers receive accurate details.

The Evolving Nature of Phishing Scams

This phishing attempt highlights the evolving sophistication of such scams. In the past, phishing emails were often marred by poor grammar or clumsy wording, making them easier to spot. However, with advancements in generative artificial intelligence, scammers can now produce polished and professional-looking messages. In this instance, one of the few red flags was the use of the SendGrid email marketing platform and the redirection through an AWS website, which sharp-eyed recipients might notice.

While it remains unclear how many individuals fell victim to this scheme, any user who shared their seed phrase likely lost their funds permanently. This incident underscores the importance of exercising caution and maintaining strict security protocols when handling sensitive information like recovery phrases.

How to Protect Your Cryptocurrency Wallet

Cryptocurrency users are advised to verify communications directly through official sources and avoid clicking on links in unsolicited emails. Recovery phrases should never be shared online, as doing so compromises the entire wallet’s security. With scams becoming increasingly sophisticated, vigilance and education are crucial in safeguarding digital assets.

Massive Credit Card Breach Puts Millions at Risk


A significant credit card breach has been uncovered, threatening to disrupt holiday shopping for millions of Americans. The breach stems from an Amazon Web Services (AWS) S3 bucket left unsecured online, which contained sensitive customer data, including credit card details, names, addresses, and emails. This exposed data belongs to approximately five million individuals who fell victim to phishing scams, with one notable scheme promoting fake offers for a free iPhone.

The breach poses immediate risks such as fraud, unauthorized transactions, and identity theft. Cybersecurity experts are advising affected individuals to contact their financial institutions promptly to mitigate potential damage. Although the perpetrators remain unidentified, Amazon’s AWS Abuse team has initiated an investigation into the incident. According to researchers at Leakd.com, the breach originated from a phishing campaign orchestrated by a fraudulent company named “Braniacshop.” This group deceived victims with false promises of winning an iPhone 14, leveraging social engineering tactics like fake emails and websites to obtain personal information.

Researchers warn that the stolen data, now potentially available on the dark web, is estimated to be worth $85 million. Each stolen credit card detail could fetch up to $17. The timing of this breach during the busy holiday season intensifies its impact, as millions of Americans could face financial challenges while preparing for Christmas. To mitigate these risks, experts recommend carefully monitoring financial statements for any unusual activity and immediately notifying banks or credit card issuers of suspicious transactions to freeze compromised accounts.

Setting up fraud alerts with financial institutions can add an additional layer of security by flagging unauthorized actions. Taking a proactive stance, such as initiating a credit freeze, can prevent scammers from opening new accounts in an individual’s name. Strengthening online account security by using multi-factor authentication, encrypted password managers, and longer passphrases is another critical step to safeguard personal information.

For comprehensive protection, investing in identity theft monitoring services is highly recommended. These services provide ongoing alerts about potential misuse of personal data, helping users act swiftly in the event of a security breach. The incident serves as a crucial reminder to remain vigilant, particularly during the holiday season when phishing scams and fraudulent offers become more frequent. Proactive measures taken now can safeguard financial security and ensure peace of mind during this critical period. 

Amazon Identified Internet domains Exploited by Russian APT29

 

The leading advanced persistent threat group in Russia has been phishing thousands of targets in businesses, government agencies, and military institutions. 

APT29 (also known as Midnight Blizzard, Nobelium, and Cozy Bear) is one of the world's most prominent threat actors. It is well known for the historic breaches of SolarWinds and the Democratic National Committee (DNC), which are carried out by the Russian Federation's Foreign Intelligence Service (SVR). It has recently breached Microsoft's codebase and political targets in Europe, Africa, and beyond. 

"APT29 embodies the 'persistent' part of 'advanced persistent threat,'" notes Satnam Narang, senior staff research engineer at Tenable. "It has persistently targeted organizations in the United States and Europe for years, utilizing various techniques, including spear-phishing and exploitation of vulnerabilities to gain initial access and elevate privileges. Its modus operandi is the collection of foreign intelligence, as well as maintaining persistence in compromised organizations in order to conduct future operations.”

In the same vein, the Computer Emergency Response Team of Ukraine (CERT-UA) recently found APT29 phishing Windows credentials from government, military, and commercial sector targets in Ukraine. After comparing notes with authorities in other nations, CERT-UA discovered that the campaign had expanded across "a wide geography."

It is not surprising that APT29 would target sensitive credentials from geopolitically influential and diversified organisations, according to Narang. However, "the one thing that does kind of stray from the path would be its broad targeting, versus [its typical more] narrowly focused attacks.” 

AWS and Microsoft

Malicious domain names that were intended to seem to be linked to Amazon Web Services (AWS) were used in the August campaign. The emails received from these domains simulated to give recipients advice on how to set up zero trust architecture and combine AWS with Microsoft services. Despite the charade, AWS stated that neither Amazon nor its customers' AWS credentials were the target of the attackers.

The attachments to those emails revealed what APT29 was really looking for: configuration files for Remote Desktop, Microsoft's application for implementing the Remote Desktop Protocol. RDP is a common remote access technique used by regular consumers and hackers. 

"Normally, attackers will try to brute force their way into your system or exploit vulnerabilities, then have RDP configured. In this case, they're basically saying: 'We want to establish that connection [upfront],'" Narang added. 

Launching one of these malicious attachments would have resulted in an immediate outbound RDP connection to an APT29 server. But that wasn't all: the files contained a number of other malicious parameters, such that when a connection was established, the perpetrator gained access to the target computer's storage, clipboard, audio devices, network resources, printers, communication (COM) ports, and more, as well as the ability to execute custom malicious scripts.

Bling Libra Shifts Focus to Extortion in Cloud-Based Attacks

 


It was observed during an incident response engagement handled by Unit 42, that the threat actor group Bling Libra (which was responsible for distributing ShinyHunters ransomware) had shifted from extortion to extortion of victims rather than its traditional tactic of selling/publishing stolen data in an attempt to increase their profits. 

During this engagement, it was also demonstrated how the group was able to acquire legitimate credentials, which were accessed from public repositories, to gain initial access to an organization's Amazon Web Services (AWS) environment through its public username and password. The compromised credentials had limited impact due to the limited permissions associated with them, but Bling Libra managed to infiltrate the organization's AWS environment and conduct reconnaissance operations on it during this time. 

The threat actor group used various tools for gaining information and accessing S3 bucket configurations, interacting with S3 objects, as well as deleting files from the service using tools such as the Amazon Simple Storage Service (S3) Browser and WinSCP. As a result of previous jobs with high-profile data breaches, including the Microsoft GitHub and Tokopedia incidents in 2020, Bling Libra has developed a special part of their business model that enables them to monetize stolen data through underground marketplaces. 

There has, however, been a significant change in the methods that Unit 42 implements, which have been reported in a recent report. As of 2024, Bling Libra has revitalized its business model from data theft to extortion, primarily targeting vulnerabilities within cloud-based environments to heighten its revenue. As Unit 42 explained in its latest report, Bling Libra obtained AWS credentials from a sensitive file that was exposed online to perform the latest attack. 

AWS account credentials were obtained from an Identity and Access Management (IAM) user, which would have provided the attackers with access to the victim's account on Amazon Web Services (AWS). While the permissions for accessing Amazon S3 resources were restricted, Bling Libra exploited them to gain a foothold in the cloud environment even though they were limited. Even though Bling Libra uses the same method of accessing victims for the first few minutes, it has instead instigated the double-extortion tactics normally associated with ransomware gangs - they initially steal data from victims and threaten to publish it online if they do not pay the ransom. 

According to the researchers, Bling Libra used credentials from a sensitive file exposed by the attacker on the Internet as a way of stealing the credentials, even though this file contained a variety of credentials. Aside from these exposed AWS access keys, the group also alleged that it "targeted a few other one-time credentials that were exposed by this individual as well as a few other exposed AWS access keys belonging to this individual.". 

Using these credentials, it is possible for the threat actors to gain access to the AWS account where the IAM user resides and to use the AWS API call to interact with the S3 bucket under the context of the AmazonS3FullAccess policy, which allows all permissions to be granted to users. The attackers in this case sat on the network and lurked for about a month before launching an attack that led to the exfiltration of information, its deletion from the environment, and the recovery of an extortion note demanding ransom payment. 

Their ransom note gave them a week to make their payment. It has been reported that Bling Libra also created new S3 buckets in the aftermath of their attack, presumably to mock the organization about the attack, as well. Ticketmaster's attack in June was notable because of how much data Bling Libra was able to obtain during this attack. At the time, the organization claimed that a total of more than half a million records were stolen, some of which contained Personal Identifiable Information (PII) such as names, emails, addresses, and partial credit card information. 

In May, the same group also claimed responsibility for several other attacks on other companies, including Ticketek Entertainment Group (TEG), in Australia, that occurred around the same period as Ticketmaster. Like Ticketmaster, TEG was attacked at the beginning of May. This group has been associated with several significant data breaches that have affected millions of records of data, and the implications have been severe. 

In the final phase of the attack, Bling Libra created new S3 buckets with mocking names to signify their control over the environment, illustrating their ability to manipulate the system. The threat group known as Bling Libra has adopted a new tactic, pivoting to extortion as a primary method for monetizing their cyber breaches. 

Following their recent cloud-based attacks, the group sent out extortion emails demanding payment in exchange for the return of stolen data and the cessation of further malicious activities. This shift in strategy underscores their focus on using extortion as a central means to profit from their operations. A recent report by Unit 42 offers a comprehensive analysis of Bling Libra's operational tools, particularly emphasizing their use of S3 Browser and WinSCP. 

These tools enable the threat actors to interact seamlessly with Amazon Web Services (AWS) environments. The report provides in-depth insights that assist incident responders in distinguishing between legitimate tool usage and activities indicative of a security breach. To counteract such threats, Unit 42 strongly advises organizations to adhere to the principle of least privilege, ensuring that users have only the minimal level of access necessary to perform their functions. 

Additionally, they recommend implementing robust security measures, including the use of AWS IAM Access Analyzer and AWS Service Control Policies. These tools are essential for mitigating the risks associated with similar attacks on cloud infrastructure. As businesses increasingly depend on cloud technologies, maintaining a proactive and vigilant cybersecurity posture is critical. Organizations must be diligent in their efforts to protect their cloud environments from sophisticated threat actors like Bling Libra.

AWS Employs MadPot Decoy System to Thwart APTs and Botnets

 

Amazon Web Services (AWS), a prominent player in cloud computing, has unveiled its internal defense system, MadPot, which has proven effective in luring and trapping malicious activities, including those orchestrated by nation-state-backed Advanced Persistent Threats (APTs) such as Volt Typhoon and Sandworm.

Conceived by AWS software engineer Nima Sharifi Mehr, MadPot is described as an advanced network of monitoring sensors equipped with automated response capabilities. This system ensnares malicious actors, monitors their actions, and generates protective data for various AWS security products.

MadPot is ingeniously designed to mimic numerous plausible targets, thwarting Distributed Denial of Service (DDoS) botnets, and preemptively blocking formidable threat actors like Sandworm from compromising AWS customers.

According to AWS, the sensors are vigilant over a staggering 100 million potential threat interactions and probes daily worldwide. Out of these, about 500,000 are identified as malicious activities, and this colossal trove of threat intelligence is meticulously analyzed to provide actionable insights on potentially harmful online activities. 

The response capabilities automatically shield the AWS network from identified threats, and they also reach out to other companies whose infrastructure is being exploited for malicious purposes.

In the case of Sandworm, the honeypot effectively intercepted the actor's attempt to exploit a security vulnerability in WatchGuard network security appliances. AWS not only identified IP addresses but also other distinct attributes linked to the Sandworm threat involved in the attempted breach of an AWS customer.

MadPot's remarkable capability to simulate a range of services and engage in extensive interactions enabled AWS to gather additional insights about Sandworm campaigns. This included specific services targeted by the actor and post-exploitation commands initiated by them. Armed with this intelligence, AWS promptly informed the affected customer, who took swift action to rectify the vulnerability.

Furthermore, AWS highlighted that the data and insights gathered by MadPot are harnessed to enhance the efficacy of their security tools, including AWS WAF, AWS Shield, AWS Network Firewall, and Amazon Route 53 Resolver DNS Firewall. These are complemented by detective and reactive services like Amazon GuardDuty, AWS Security Hub, and Amazon Inspector.