Search/ESET Blog
● Source

ESET Blog

Connections
26 relationships
This month in security with Tony Anscombe – October 2026 edition
With October coming to a close, ESET Chief Security Evangelist Tony Anscombe reviews some of the top cybersecurity stories that made the news over the month. At the same time, he offers insights that the latest wave of incidents hold for your company's, or your own, cyber-defenses. Here's Tony's rundown of some of what stood out most in October: Over the past ten days, the cybersecurity news cycle has been dominated by two critical vulnerabilities in PHP, a server-side scripting language powering some 70 percent of websites globally. Here's what you should know more about the pair of flaws known as PHP2Shell, as well as related topics – and we also dived into the issues in this video and this blog post. Cyberattackers broke into refrigeration controllers at up to 700 French supermarkets and raised the temperature of refrigerators and freezers to as high as 50 degrees Celsius. A data breach at Denmark’s Central Population Register (CPR) has exposed the personal information of 8.8 million people. The person claiming responsibility for the hack claims to have gained access through a third-party provider's account that was 'protected' by the password ‘123456’. Malware preinstalled on thousands of cheap Android phones that can generate fraudulent ad revenue and potentially turn infected devices into parts of larger botnets – a problem that we also wrote about a while back. What are the lessons that businesses should take away from the latest cybersecurity news? Watch Tony's video to find out, and be sure to check out the September 2026 edition of his monthly security news roundup for more insights.
welivesecurity.comOct 30, 2026source
Lessons from Cybersecurity Black Week
The week leading up to October 25, 2026, was every cyber defender’s nightmare: a barrage of critical vulnerabilities in PHP and the Linux kernel, a compromised LiteLLM release, a maximum-severity flaw in GitLab, a Node.js Brotli flaw, and a certificate authority incident affecting TLS trust. A few years ago, a backlog like this would have filled a defender’s to-do list for a whole month. In the age of AI, all of this was reported within a single “Cybersecurity Black Week,” an exceptionally eventful seven-day period amid an already turbulent Black October. AI-assisted research helped uncover at least two of these flaws: PHP2Shell and the Brotli bug. On top of the breakneck patching pace, security teams faced mounting pressure to deliver status reports to other departments, management, customers, and the media. Beyond their severity, the issues stood out for the variety of responses they demanded. Some required emergency runtime patches and upgrades; others called for containment, credential rotation, or the replacement of outdated certificate trust bundles. In several cases, installing the latest version was only the beginning of a long remediation process. With so many problems landing at once, the main challenge was separating real risk from hype and prioritizing accordingly. Here is our assessment. PHP2Shell: The loudest headline, with real-world limits PHP2Shell generated the biggest media frenzy of the lot and, with it, a major headache for security teams, prompting many organizations to take their websites offline and look for emergency fixes. The name refers to two memory-safety vulnerabilities in PHP core, CVE-2026-109618 and CVE-2026-109619, affecting PHP versions 5.4 through 8.5.11. Both vulnerabilities received a CVSS v4.0 score of 9.2 and can allow unauthenticated remote code execution (RCE) on vulnerable PHP servers. In the highest-risk setups, a demonstrated exploit can trigger the vulnerability before application code runs, meaning application-level authentication offers no protection. With PHP powering an estimated 69% of websites whose server-side programming language is known, government agencies in the United States, Europe, and elsewhere urged organizations to patch immediately, with some advising that affected internet-facing systems be isolated or taken offline if they could not be patched promptly. Government websites in the Netherlands, Germany, France, and other countries were temporarily taken offline as a precaution. The PHP project released fixes on October 21 in versions 8.5.12, 8.4.27, 8.3.36, and 8.2.35, along with source-only backports for the end-of-life 7.4, 8.0, and 8.1 branches. The rollout was complicated by a premature public leak, which forced the coordinated disclosure timeline forward. While the flaws clearly belonged near the top of the patching list, early evidence suggested they were considerably harder to weaponize than initially feared. Successful exploitation requires specific conditions and knowledge of the target, and public proof-of-concept exploits did not work against a default PHP installation, according to Sansec. On October 22, ESET was among nine security companies that told NBC News they had not seen any successful attacks, even though more than 200 proof-of-concept exploits were circulating. The remediation challenge extends beyond obvious web servers: PHP can also be embedded in appliances, network devices, management interfaces, and third-party products, while end-of-life PHP branches do not receive normal upstream binary releases. Linux kernel: Reports went public before fixes were ready As if the PHP issues weren’t severe enough, on October 21, Trend Micro’s Zero Day Initiative (ZDI) sent several reports on unpatched Linux kernel vulnerabilities to the kernel’s private security address while also copying public mailing lists. As a result, technical details of vulnerabilities for which fixes were not yet generally available became public, prompting stable kernel maintainer Greg Kroah-Hartman to question why the reports had been sent to public lists without patches. The two most severe issues, both rated CVSS 9.0 by ZDI, affect KSMBD, the Linux kernel’s built-in SMB server. ZDI-CAN-33549 is an integer underflow in the SMB Direct path that could lead to RCE, while ZDI-CAN-33774 is a race condition that could also lead to RCE. Exploiting ZDI-CAN-33549 requires an RDMA-capable interface, a writable KSMBD share, and SMB Direct negotiation, significantly narrowing the number of systems exposed to that particular attack. Other reports concerned local privilege escalation vulnerabilities in the networking stack and the VMware graphics driver, as well as an information disclosure issue in the NVMe target code (NVMET). To remediate, security teams had to track each report independently, consult distribution advisories, and restrict exposed services where practical while awaiting verified fixes. Kernel updates also required rebooting into the patched kernel and checking the running version, not merely confirming that a new package had been installed. LiteLLM: Compromised, not merely vulnerable On October 21, attackers published a malicious version of LiteLLM, 1.107.0, to PyPI and pushed compromised images to the project’s official container registries. According to the project’s incident report, its release pipeline was breached using GitHub CLI credentials associated with three developers. One of the stolen tokens belonged to an owner of the project’s GitHub organization, allowing attackers to disable branch protections and modify the release workflow to point to a lookalike repository. LiteLLM, a Python proxy for LLM providers, has about 23 million weekly downloads, and the malicious version remained available on PyPI for roughly two hours before it was quarantined. It was the project’s second supply-chain compromise of 2026, after malicious versions 1.82.7 and 1.82.8 appeared in March. The payload combined a credential stealer with a second-stage remote access trojan (RAT). By abusing Python’s startup mechanisms, including a .pth file, it ran every time Python started, even if an application never imported LiteLLM. Execution therefore did not depend on exploiting an application flaw: simply running Python in an affected environment could be enough. Affected environments had to be treated as potentially compromised, not merely vulnerable. According to LiteLLM, these included environments that ran litellm==1.107.0 and containers that executed affected images distributed under the v1.107.0, latest, main-latest, or main-stable tags during the compromised intervals. Removing or replacing the package was insufficient: teams needed to determine whether the malicious code had run, contain affected systems, rebuild from clean sources, and rotate exposed secrets from a clean environment. Cloud, source-control, and CI/CD logs also had to be reviewed for unauthorized activity. Node2Shell: From Node.js to HTTP2Shell Node2Shell (CVE-2026-111497) is an out-of-bounds write in Google’s Brotli decoder, which Node.js bundles. Malformed compressed input can cause crashes, memory disclosure, or potentially RCE when an application decompresses attacker-controlled Brotli data. This can happen through explicit decompression, compressed request bodies, or automatically decompressed responses. Node.js rates the flaw as high severity, while the broader Brotli vulnerability carries a CVSS v4.0 score of 9.2. According to Aikido, researchers demonstrated code execution against the official, unmodified Node.js 24.22.0 image on Linux ARM64. Although the exploit did not work on every attempt, the demonstration showed that exploitation was practical, if not universally reliable. Node.js shipped fixes on October 21 in versions 22.23.4, 24.22.1, and 26.12.1, and Bun followed with version 1.4.4 or later. Node.js 25.x, 23.x, 20.x, and earlier releases are end-of-life and will not receive the fix. Within days, the story outgrew Node.js. The same decoder is compiled into the Python Brotli and brotlicffi packages as well as into other HTTP software, which led researchers to rename the issue HTTP2Shell. The operational challenge was therefore not simply upgrading a single runtime but identifying every container, serverless deployment, and library that bundles the vulnerable decoder. The issue was fixed upstream in Brotli 1.2.1. Notably, OpenAI’s Codex Security had reported the underlying bug, and the fix had been present in the Brotli source since February without being recognized as a security issue. GitLab: Credential leaks and authorization failures On October 21, GitLab shipped a critical out-of-band patch release (versions 19.5.1, 19.4.3, 19.3.6, 19.2.9, 19.1.9, 19.0.10, and 18.11.13) addressing 12 security vulnerabilities. The headline flaw, CVE-2026-110264, rated 10.0 under CVSS v3.1, is an integer overflow triggered when GitLab compiles a specially crafted regular expression. Under certain conditions, it could allow an unauthenticated attacker to execute arbitrary code on the server. The affected versions date back to GitLab 7.0. The other fixes addressed credential exposure, cache poisoning, authorization bypasses, and failures to enforce membership or merge-request approval policies. Reported unauthenticated attack paths included exposure of sensitive credentials through internal response headers and manipulation of a shared OAuth/OIDC discovery cache, while other issues required authenticated access. GitLab did not publish individual severity scores or exploitability details for these issues. GitLab.com and GitLab Dedicated were already patched, so the work fell on administrators of self-managed instances. GitLab strongly recommended upgrading immediately, but with a catch: the older backport releases (19.2.9, 19.1.9, 19.0.10, and 18.11.13) included only the critical fix, while the remaining 11 issues were fixed only in 19.3.6 and later. Administrators also had to verify that all deployment components had been updated and assess the potential exposure of credentials and private project data due to the fact that patching would not invalidate stolen credentials or reverse unauthorized changes. eMudhra/emSign: Compromised certificate trust The high-priority eMudhra/emSign incident involved no software vulnerability at all. Several eMudhra intermediate certificate authorities (CAs) had insecure private keys, meaning attackers could potentially create certificates that would be trusted through the affected emSign certificate chain. While browsers had updated their trust stores and revocation lists, some Python tooling continued to trust the compromised intermediate CAs through its own certificate bundles. This created a risk of man-in-the-middle (MITM) attacks against TLS connections, including those to python.org and pypi.org. The Python Software Foundation urged users to upgrade immediately to at least certifi 2026.10.19, pip 26.2.2, and uv 0.13.4, and the Python core team shipped expedited releases 3.15.2, 3.14.10, 3.13.18, 3.12.16, and 3.11.18 to update the bundled pip. The updates had to reach every host, application, and existing virtual environment. Container images had to be rebuilt with layer caching disabled to fetch the latest system certificates. The decisive check was whether deployed clients rejected the compromised trust paths; simply upgrading Python was not enough. Cutting through the noise Cybersecurity Black Week showed that a severity score rarely tells the whole story. A 9.2-rated PHP flaw that attackers struggled to exploit, a 10.0-rated GitLab bug, exposed kernel reports with no patches, and a poisoned package that ran at Python startup each called for a different response. What they all had in common was a window of exposure: the time between disclosure and when the last affected system is patched. For the leaked kernel flaws, that window opened before any fix existed, and for end-of-life PHP versions, it may never fully close. This is where ESET’s multilayered protection comes in, using behavioural detection and other security mechanisms to help block exploitation attempts and keep systems protected until patches are applied. At the same time, the signals defenders rely on were buried in noise, from scanning that was largely benign to a single behavioural rule firing hundreds of thousands of times in a week. What mattered most was quickly understanding which incidents actually affected an organization's environment and what each incident required. That is the problem ESET’s correlation engine, ESET LiveCortex, and its specialized cybersecurity model, ESET NeoCortex, are designed to solve: turning an overwhelming volume of raw data into understandable incidents so that defenders can act before attackers achieve their goals. To learn how they work, read Beyond the alert: How ESET is teaching AI to understand security incidents.
welivesecurity.comOct 28, 2026source
ESET APT Activity Report Q2 2026–Q3 2026
ESET APT Activity Report Q2 2026–Q2 2026 summarizes notable activities of selected advanced persistent threat (APT) groups documented by ESET researchers between April and September 2026. The operations presented here reflect the broader threat landscape observed during this period and highlight some of its key trends and developments. They represent only a subset of the cybersecurity intelligence made available to customers of ESET Threat Intelligence APT Reports. During the monitored period, China-aligned groups conducted intelligence-collection operations primarily against governmental targets. FamousSparrow kept up its monitoring in South America, just as in the previous reporting period. In Southeast Asia and Oceania, China-aligned threat actors ran campaigns that appear to support Chinese law enforcement efforts against organized crime in the region. One threat actor targeted hotels in Palau, which seems to be increasingly used as a base for fraud-related organized crime. SlippyNunchucks deployed its DarkVigil toolset in Sri Lanka against Chinese-speaking individuals connected to fraud schemes, likely in the context of broader regional law enforcement efforts against financial crime. As scam compounds have become an issue for countries neighboring China, with direct negative effects on China itself, it is likely that China-aligned APT groups have supported law enforcement efforts in this area. Another, broader cluster of activity targeted Taiwanese organizations operating in the semiconductor industry. This focus on Taiwan’s semiconductor sector is consistent with long-running Chinese industrial espionage priorities, and the strategic importance of semiconductor supply chains to both economic competition and national security. Middle East-aligned activity increased compared with the previous reporting period, driven mainly by the return of several established Iran-aligned groups. In the previous period, activity from these groups had declined, most likely because internet restrictions imposed by the Iranian regime hindered their own ability to operate effectively. FilthyFeline, a subgroup of OilRig, and MuddyWater resumed operations after a quieter period, with most activity directed at entities in Israel across multiple sectors; a smaller number of US companies were also affected. FilthyFeline also appears to have taken over activity against a cellular data service provider in Uzbekistan that had previously been compromised by BladedFeline, another OilRig subgroup. In parallel, we analyzed new or updated malware, suggesting renewed development activity within parts of the Iran-aligned ecosystem. We also discovered activity by Arid Viper, a Hamas-affiliated threat actor, against what appears to be the Palestinian Authority’s Ministry of Foreign Affairs and Expatriates or an affiliated governmental entity. Given the ongoing US-led peace process, as well as the long-standing rivalry between Hamas and the Palestinian Authority, this activity was potentially intended to collect intelligence on the Palestinian Authority’s diplomatic position and internal thinking regarding negotiations to reduce hostilities in the region. North Korea-aligned threat actors did not deviate significantly from the types of targets seen in previous reporting periods. These include primarily technology and cryptocurrency companies, software developers, and other opportunistic targets. Alongside South Korea, Ukraine has become increasingly prominent among North Korea-aligned targets. This is likely related to growing ties between Russia and North Korea, and North Korea’s increasing involvement in Russia’s war against Ukraine, not only through indirect military support (munitions, etc.) but also through the deployment of North Korean soldiers. Alongside these broader trends, Lazarus resumed targeting of the drone sector, typically using recruitment-themed social engineering to gain initial access. The likely motivation is the increasingly important role that drone systems are playing in the war in Ukraine, where new tactics and capabilities are being tested. Access to such knowledge and manufacturing processes could provide North Korea with valuable know-how for developing its own drone-production capabilities. We also detected ScarCruft targeting an African country, which was probably intended either to maintain oversight of a North Korean worker scheme operating in the country or to monitor Russian activities there. Separately, DeceptiveDevelopment moved from fake job-interview lures to broader software supply-chain compromises, harvesting developer credentials and using that access to inject malicious code into public and private source code repositories. Russia-aligned activity continued to support its war objectives, targeting governments, the defense sector, and military entities in Europe, with a substantially higher number of targets in Ukraine. Looking more closely at these activities, similarities between Russia-aligned and Belarus-aligned campaigns stand out, further supporting our hypothesis that threat actors aligned with the two countries are becoming operationally closer. Gamaredon’s operations, which are typically numerous, saw a significant decline. We suspect that this could be related to Ukrainian attacks on Crimea, where the group is suspected to operate, a broader restructuring of its operations, or other factors that we continue to assess. Despite this decline in activity, Gamaredon updated its PteroBox file stealer and started using new tools such as PteroPSDoor and Turla’s VBScript Rezzuar. PhantomCaptcha attempted to deploy spyware, which we named PrincessSpy, on Android devices while impersonating Ukrainian entities. We also identified an extensive campaign in Ukraine exploiting MDaemon mail servers, during which the attackers deployed two backdoor malware families, HELLAGE and HELLTAG. North Korea’s expanding military cooperation with Russia has prompted expectations that North Korea-aligned cyberactivity might be starting to align more closely with Russian objectives in Ukraine. While some of their targeting interests overlap, notably around the drone sector, our telemetry more clearly points to Belarus-aligned groups acting as cyberproxies or force multipliers for Russia. We observed increased and more assertive activity from Belarus-aligned threat actors, with their operations increasingly reflecting Russia’s intelligence priorities in Ukraine in some areas. Most notably, we identified the first destructive cyberattack by a Belarus-aligned threat actor that we have documented in the European theater, marking a significant development in this threat nexus. According to our telemetry, FrostyNeighbor attacked a manufacturing company in Ukraine that plays an important role in the country’s transportation sector. This marks a notable escalation, as threat actors with a well-established history of destructive cyberoperations in Ukraine have, until now, been Russia-aligned. Finally, we identified several cases of APT groups using AI and large language models (LLMs) at various stages of their operations. Threat actors used AI primarily to accelerate tasks that would otherwise require painstaking effort, including malware development and the creation of decoy websites. In some cases, this resembled vibe coding: rapid, prompt-driven drafting of code, with human operators adapting and integrating the results into their toolchains. We also discovered phishing lures containing images that were clearly AI-generated. In one notable case, Russia-aligned UAC-0099 inserted a decoy request for instructions on building a nuclear weapon into a script comment, apparently attempting to trigger the safety guardrails of LLM-based security analysis and prevent the inspection of the malicious code that followed. Despite these developments, we have not noticed AI substantially reshaping the APT landscape over the past six months. While we cannot determine whether some operations may involve a higher degree of AI-driven automation, our telemetry continues to show hands-on-keyboard activity and operations that are not fully automated. Overall, our observations suggest that APT groups are currently using AI much like mainstream software developers: to accelerate software development, automate previously manual tasks, and assist with the analysis of large volumes of data, rather than to conduct fully autonomous malicious operations. ESET products protect our customers’ systems from the malicious activities described in this report. Intelligence shared here is based mostly on proprietary ESET telemetry data and has been verified by ESET researchers. ESET APT Activity Reports contain only a fraction of the cybersecurity intelligence data provided in ESET Threat Intelligence APT Reports. For more information, visit the ESET Threat Intelligence website.
welivesecurity.comOct 28, 2026source
MythicRatatoskr: An Yggdrasil-enabled backdoor deployed in a UAC-0099-Sandworm collaboration
Executive summary This report analyzes activity during which UAC‑0099 gained initial access on July 16th, 2025 and handed it off to Sandworm, which then conducted a data-wiping attack on September 17th, 2025. It particularly focuses on the MythicRatatoskr backdoor, which is unique in its use of the Yggdrasil mesh network to communicate with a C&C server hosted within it. We presented our findings about MythicRatatoskr at LABScon on September 18th, 2026, and at the Virus Bulletin conference on October 16th, 2026. Key points of this blogpost: On July 16th, 2025, UAC-0099 gained initial access to a grain logistics company in Ukraine via a spearphishing email with a malicious attachment, leading to the execution chain of the MATCHWOK backdoor. One day later, the attackers deployed MythicRatatoskr, a previously undocumented backdoor that communicates with a C&C server hosted inside the Yggdrasil mesh network. We have not seen the backdoor used anywhere else, and we found indications that a Linux variant may also exist. Yggdrasil is a decentralized mesh network that implements an experimental routing scheme in which the IPv6 communication between the nodes is end-to-end protected. Ultimately, UAC-0099 handed off access to Sandworm, which then carried out a data-wiping attack. Cooperation among Russia-aligned actors The handover documented in this report is not an isolated phenomenon. As noted above, Sandworm is attributed to GRU Unit 74455, and UAC-0099’s targeting leads us to assess that it is aligned with Russian interests. Earlier in 2025, we uncovered that Russia-aligned Gamaredon collaborated with Turla, another Russia-aligned threat actor; we documented our findings in our blogpost Gamaredon X Turla collab. Both groups are linked to the Federal Security Service of the Russian Federation (FSB). This kind of cooperation underscores the potential for coordinated cyberespionage campaigns among Russia-aligned groups, likely intended to amplify their operational impact. In the past, Gamaredon collaborated with a threat actor that we discovered and named InvisiMole. Against that backdrop, a UAC‑0099-to-Sandworm handoff is unusual in the specific actors involved, but no longer surprising as a pattern. Overview of the intrusion According to ESET telemetry, the events unfolded as follows: at 06:59 on July 16th, 2025, a scheduled task named pdfopentask was executed on one machine in the target organization. It executed a PowerShell payload from %temp%\temporarydoc.txt. A minute later, MATCHBOIL was deployed. MATCHBOIL is a downloader that deployed, in this case, the MATCHWOK backdoor (C&C: https://uztickets[.]com) from https://egyptanimals[.]com/animals/photoalbum. The following day, at 09:28, a new binary was delivered to the machine: the MythicRatatoskr backdoor. Due to the close proximity of time, we believe with high confidence that the MATCHWOK execution chain led to the deployment of MythicRatatoskr. Our telemetry further shows that additional MythicRatatoskr samples were subsequently deployed to other machines within the organization, this time using an OpenSSH SFTP server. Later, ESET telemetry detected a SOCKS5 server on some of the machines, with one sample having been extracted from an unknown archive using WinRAR. We believe that the SOCKS5 server was deployed using MythicRatatoskr. Approximately two months later, on September 17th, 2025, we detected data-wiping activity at the compromised grain logistics company. The data-wiping attack was executed by deploying the SDelete utility via Active Directory Group Policy – standard Sandworm tactics, techniques, and procedures (TTPs) that we have observed numerous times in ESET telemetry. Therefore, we assess that UAC‑0099 can act as an initial access group for Sandworm. This blogpost focuses primarily on the analysis of the MythicRatatoskr backdoor, the Yggdrasil network that hosts its C&C server, and the accompanying SOCKS5 server. The first versions of MATCHBOIL and MATCHWOK have already been analyzed by CERT‑UA in its advisory, with MATCHBOIL and its evolution further covered in our blogpost MATCHBOIL: New tricks, same old evil intentions. These tools are therefore omitted here. A squirrel in the world tree The word Yggdrasil originates in Norse mythology, where it refers to a sacred tree connecting the various realms of the cosmos. Within that tree lives a squirrel named Ratatoskr that carries messages between its top and its roots. Given the backdoor’s role as a messenger scurrying through the Yggdrasil network, we decided to name the backdoor MythicRatatoskr. More technically, Yggdrasil is a decentralized mesh network that connects devices directly to each other without a central hub, building an encrypted IPv6 overlay on top of the internet so that participating nodes can exchange end-to-end-protected traffic. Legitimate users adopt such meshes for resilient connectivity, research, and hobby networking. Because no single server runs the mesh, the network is robust and hard to disrupt – properties that also make it attractive to adversaries. Importantly, Yggdrasil is not designed to provide anonymity. Attribution The post-compromise execution chain relied on known malware already attributed to UAC‑0099 – namely the MATCHBOIL downloader and the MATCHWOK backdoor – which leads us to attribute this stage of the attack to UAC‑0099 with high confidence. We attribute the data-wiping activity observed two months later to Sandworm, also with high confidence, based on the TTPs involved. Moreover, we assess with high confidence that UAC‑0099 handed off access to Sandworm at some point during the intrusion, although we have not yet been able to determine whether the MythicRatatoskr backdoor was operated primarily by UAC‑0099 or by Sandworm. UAC‑0099 profile UAC‑0099 is a cyberespionage group targeting governmental organizations, financial institutions, and media, all in Ukraine. Based on the targeting, we believe with medium confidence that the group is aligned with Russian interests. UAC‑0099 can act as an initial access broker for Sandworm, a Russia-aligned group best known for its destructive attacks in Ukraine. The group has been active since at least 2022 and was first reported by CERT‑UA in June 2023. Apart from MATCHBOIL, the group also typically deploys LONEPAGE, a PowerShell downloader named after the presence of the word page in its C&C URLs. Sandworm profile Sandworm is a Russia-aligned threat group that performs destructive attacks. It is mostly known for attacks against Ukrainian energy companies in December 2015 and December 2016, which resulted in power outages. In June 2017 Sandworm launched the NotPetya data-wiping attack that used a supply-chain vector by compromising the Ukrainian accounting software M.E.Doc. In February 2018, Sandworm launched the Olympic Destroyer data-wiping attack against organizers of the 2018 Winter Olympics in Pyeongchang. The Sandworm group uses such advanced malware as Industroyer, which is able to communicate with equipment at energy companies via industrial control protocols. In April 2022, CERT‑UA thwarted an attack against an energy company in Ukraine where the Sandworm group tried to deploy a new variant of this malware, Industroyer2. Since February 2022, Sandworm has been very active in Ukraine, by deploying data wipers against strategic organizations. In October 2020, the US Department of Justice published an indictment against six Russian computer hackers that it alleges prepared and conducted various Sandworm attacks. As we already mentioned, the group is commonly attributed to GRU’s Unit 74455. Compromise chain Initial access and execution According to ESET telemetry, initial access was gained via a spearphishing email with a malicious attachment – a finding that matches CERT‑UA blogpost. We assess that the MATCHWOK backdoor execution chain (already documented in the CERT‑UA blogpost) then led to the execution of MythicRatatoskr, a backdoor that communicates with its C&C server over the Yggdrasil network. Subsequently, the attackers probably used MythicRatatoskr to deliver a SOCKS5 server. Persistence MATCHBOIL persists MATCHWOK by adding the registry value UpdateMonitor to the registry key HKCU\Software\Microsoft\Windows\CurrentVersion\Run\, with data pointing to the dropped MATCHWOK path. MythicRatatoskr is started as a Windows service with the name ieupdate. Collection and exfiltration During check-in, MythicRatatoskr sends information about the current user, hostname, and local network settings to the C&C server. In addition, the operators can use MythicRatatoskr to exfiltrate files and screenshots to the C&C server by using HTTPS routed through the Yggdrasil mesh network. The data is not staged prior to exfiltration. Toolset MythicRatatoskr MythicRatatoskr is a backdoor written in Go that uses the Yggdrasil network (described in the Network protocol section) to reach its C&C server. As noted earlier, we named the backdoor after Ratatoskr, the squirrel from Norse mythology that scurries through Yggdrasil – the sacred tree connecting the realms of the cosmos – carrying messages between its top and its roots. We felt this was a fitting parallel for a backdoor that ferries data across the Yggdrasil network. There are signs in the binaries suggesting that the development started sometime between December 2024 and February 2025. Function names left in the binaries – following a mix of mixed caps (CheckInAgent) and snake case (agent_run) – suggest either code merged from another project or inconsistent naming by the authors. Key functions handle service registration and Yggdrasil network initialization (main, runService), check-in and beaconing (agent_run, init_agent, CheckInAgent), configuration handling (saveConfig, Encrypt, Decrypt), command polling and dispatch (getTasks, doTask), and the backdoor commands themselves (upload, download, screencat, shell, manageProxy, etc.). Some samples also include a 10-minute startup delay, likely to avoid immediately connecting to the Yggdrasil network on execution, helping to evade detection. Host-based artifacts Table 1 presents MythicRatatoskr host-based artifacts that we observed during our analysis. Note the filename collision for krbtman.exe: it is MythicRatatoskr on some machines, and the SOCKS5 server on others. Table 1. Host-based artifacts associated with MythicRatatoskr Persistence According to Windows event logs obtained from the compromised machines, updater.exe and WindowsDefenderSvc.exe were started, and thus likely persisted, as Windows services. Network protocol To communicate with the C&C server, MythicRatatoskr relies on the Yggdrasil network – a decentralized mesh network in an alpha state that implements an experimental routing scheme. Each node in the network generates its own Ed25519 key pair, which allows protected end-to-end IPv6 communication between the peers. The IPv6 address is derived from the public key in the 200::/7 subnet range, which was deprecated by IETF in 2004. Anonymity is not among the goals of the project. We explore this topic more in the Network infrastructure section. The yggdrasil-go repository already provides the necessary types and methods to join the Yggdrasil network. There’s also another project, called yggstack, that enables the Yggdrasil network to communicate via a network stack in the user space. This has the benefit of not needing to create a tunnel interface, which might be an issue if the user does not have the required privileges to create one, or if such an interface is not supported on the system. The MythicRatatoskr code is based on yggstack, making the malware portable. There’s a publicly available network that one can join by adding at least one public peer into the Yggdrasil configuration. MythicRatatoskr specifically contains an embedded base64-encoded blob that contains a list of public peers in the URI format. Most peers on the list are still operational. MythicRatatoskr randomly picks three peers from the hardcoded list, which is the recommended number, and connects to them in order to join this public Yggdrasil network. Once MythicRatatoskr is part of the Yggdrasil network, it communicates with its C&C server, which is also part of the network, via HTTPS on port 443. Communication uses JSON-encoded HTTP POST requests containing a uuid (explained in Table 4, in the MythicRatatoskr main configuration section), a message type, and optional data. Four message types exist: checkin – register the backdoor and send victim metadata such as username, hostname, and local IPs; the server responds with a fresh uuid and psk), tasking – poll for new commands, result – return command output, base64-encoded when present, and getfile – retrieve file content for the upload command. Although the C&C server is hosted in the Yggdrasil network – which may prevent network security tools from detecting outbound connections to it – blocking the public peer addresses is a viable mitigation. Configuration We distinguish between two types of configuration: a backdoor-specific configuration used by the malware itself; and an Yggdrasil configuration used to join the mesh network. The MythicRatatoskr configuration is stored in a file at a hardcoded path. The path can differ in each sample; we identified the following paths: C:\Windows\Temp\.updatecache (in krbtman.exe, Act64.exe, ie.exe, ieupdate.exe, and windowsdefendersvc.exe) C:\Windows\system32\.updatecache (in updater.exe, and windowsdefendersvc.exe) %LocalAppData%/Temp/.crond (only in OneDriveLiveUpdater.exe) The %LocalAppData%/Temp/.crond path is notable, since such a filename (.crond) along with forward slashes is expected for Unix-like systems. This indicates that a Linux variant (or macOS, FreeBSD, etc., though less likely) of MythicRatatoskr may also exist. Moreover, OneDriveLiveUpdater.exe contains an unused configuration path /tmp/.test_config; the filename suggests that this configuration path may have been used during testing and accidentally left in a Windows deployment. All of the configuration filenames start with a dot. On Unix-like systems, programs used for listing files, such as ls, typically exclude these entries by default, effectively hiding them. This is not the case on Windows, where these files are not hidden. This further suggests the existence of a Linux variant. The backdoor-specific configuration stored on disk is a JSON file encrypted via AES‑256‑CFB using a hardcoded key embedded in the binary, US3pM9KFxrZBSgc1pGl2wO4Kir1XjZP5Z9kBXBMYd2s=, which is decoded using base64 before being used. The first 16 bytes of the file represent the IV. The configuration file is not required for the malware to start. MythicRatatoskr contains an embedded default configuration, which it encrypts and writes to disk during initialization. This allows operators to deploy the malware without a separate configuration file and update the configuration later via C&C commands. Backdoor commands are described later in the Capabilities section. Table 2 presents the embedded MythicRatatoskr configuration common for all samples. Table 2. Embedded MythicRatatoskr configuration Although not part of the above configuration, MythicRatatoskr contains the following hardcoded C&C address for the proxy command: 201:b479:2771:8df4:b5eb:d007:63b0[:]2838:8080. It is not possible to modify the value during runtime. Yggdrasil configuration Initially, when MythicRatatoskr starts, it generates a new Yggdrasil configuration, including a new Ed25519 private key. Because the Yggdrasil IPv6 address is derived from the public key, the victim machine’s Yggdrasil IPv6 address changes each time the malware runs. MythicRatatoskr then uses empty NodeInfo data and sets the NodeInfoPrivacy flag. As a result, other nodes in the Yggdrasil network won’t be able to get more information about the node, such as the Yggdrasil version or the operating system on which the node is running. The backdoor uses a publicly available Yggdrasil network to reach the C&C server. To join the network, it needs to connect to at least one of the publicly available peers. All MythicRatatoskr samples contain the same embedded list of 32 peers from which three are randomly picked. The peer list is base64 encoded in the binary and was likely compiled by the backdoor authors from publicly available sources. We provide the full peer list at the end of the blogpost. MythicRatatoskr supports standard backdoor functionality, including command execution, file system navigation, file transfer, screenshot capture, and proxying. Table 3 contains all of the backdoor commands; the command IDs are strings that are hardcoded in the code. Table 3. MythicRatatoskr commands Unlike the standalone SOCKS5 server described below, MythicRatatoskr’s proxy command is used to create a reverse SOCKS5 proxy connection. The standalone tool creates a local SOCKS5 server instead. The SOCKS5 server exclusively used by UAC‑0099 or Sandworm during the August 2025 activity is a simple tool written in Go that uses go-socks5 for managing proxy connections. The tool is internally named bindProxy. Because MythicRatatoskr already includes SOCKS5 proxy functionality (see Table 3), it is unclear why the attackers deployed a separate SOCKS5 server. It is possible that MythicRatatoskr’s proxy command did not work as expected, so the operators installed this additional tool. Nevertheless, together they support both forward and reverse proxying. Host-based artifacts Table 4 presents SOCKS5 server host-based artifacts that we observed during our analysis. Again, note the filename collision for krbtman.exe: it is MythicRatatoskr on some machines, and the SOCKS5 server on others. Table 4. Host-based artifacts associated with the SOCKS5 server Capabilities The tool uses go-socks5 to create a SOCKS5 proxy server that listens on 127.0.0.1, port 33690. The address is specifically used to allow only the local machine to access the service. To make the proxy server reachable from outside the victim’s network, the attackers could have used MythicRatatoskr’s expose command to expose it through the Yggdrasil network. Alternatively, they may have used another existing tunnel to access the machine. As stated in its FAQ, Yggdrasil is not anonymous, so it is possible to locate nodes in the Yggdrasil network. First, we tried to find the C&C server 201:b479:2771:8df4:b5eb:d007:63b0[:]2838, which we extracted from the MythicRatatoskr configuration, in the publicly available map of the Yggdrasil network at http://51.15.204.214/ to no avail because the map was last updated on August 8th, 2024, nearly a year before the intrusion took place. We therefore generated our own map of the network’s spanning tree, which does include the C&C server, as shown in Figure 1. The C&C server is situated near the center of the tree, making it easily reachable from other peers. The colors of each node indicate how far the node is from the C&C server. For example, as indicated by the red color, the connection from/to the nodes at the edge can be unreliable. The coordinates (Coords) for the C&C are rather simple: the server is reachable via the node with ID 4, which is the root of the spanning tree. Using Yggdrasil’s command line utility yggdrasilctl, we identified that the C&C server is configured with three public peers, shown in Figure 2. The C&C server’s port 443 uses TLS certificates that are rotated regularly. This pattern is characteristic of certificates generated by revsocks, on which the server code is probably based. Conclusion In this campaign, we observed cooperation, or at least operational handoff, between two Russia-aligned threat groups with different roles. UAC‑0099 appears to have gained initial access to a Ukrainian logistics company using its known tooling. That access was later used in activity culminating in a destructive Sandworm operation. In addition, we identified a new backdoor, MythicRatatoskr, which is attributed to either UAC‑0099 or Sandworm. The C&C server for this backdoor is hosted within the Yggdrasil mesh network, and the use of such infrastructure made determining the server’s physical location particularly challenging. The backdoor was probably used to deploy the SOCKS5 server to route connections through the compromised network. To our knowledge, this is the first documented case of Yggdrasil being abused in this way by a state-aligned threat actor, and it may signal a broader trend of adversaries adopting experimental peer-to-peer networks to increase the resilience and stealth of their infrastructure. Ultimately, Sandworm carried out a data-wiping attack by deploying the SDelete utility via Active Directory Group Policy, a technique consistent with previous Sandworm destructive operations. IoCs A comprehensive list of indicators of compromise (IoCs) and samples can be found in our GitHub repository. Files Network MITRE ATT&CK techniques This table was built using version 19 of the MITRE ATT&CK framework. tls://37.205.14[.]171:993 tls://109.176.250[.]101:65534 tls://65.21.57[.]122:61995 tls://95.217.35[.]92:1337 tls://51.255.223[.]60:54232 tls://51.15.204[.]214:54321 tls://s2.i2pd[.]xyz:39575 tls://62.210.85[.]80:39575 tls://vpn.ltha[.]de:443?key=0000006149970f245e6cec43664bce203f2514b60a153e194f31e2b229a1339d tls://mlupo.duckdns[.]org:9001 quic://ip6.fvm.mywire[.]org:443?key=000000000143db657d1d6f80b5066dd109a4cb31f7dc6cb5d56050fffb014217 tls://helium.avevad[.]com:1337 tls://yggdrasil.neilalexander[.]dev:64648?key=ecbbcb3298e7d3b4196103333c3e839cfe47a6ca47602b94a6d596683f6bb358 tls://bode.theender[.]net:42169?key=f91b909f43829f8b20732b3bcf80cbc4bb078dd47b41638379a078e35984c9a4 tls://g.ygg[.]yt:443 tls://b.ygg[.]yt:443 tls://n.ygg[.]yt:443 tls://ygg1.grin[.]hu:42444 tls://[2a03:90c0:85:[:]28a]:443 tls://srl.newsdeef[.]eu:59999 tls://23.137.249[.]65:444 tls://ygg-nl.incognet[.]io:8884 quic://0.ygg.l1qu1d[.]net:11102?key=0000000998b5ff8c0f1115ce9212f772d0427151f50fe858e6de1d22600f1680 tls://london.sabretruth[.]org:18472 tls://yggdrasil.sunsung[.]fun:4443 tls://78.27.153[.]163:33166 tls://193.93.119[.]42:443 tcp://sysop[.]link:555 tls://sto01.yggdrasil.hosted-by.skhron[.]eu:8884 tls://ygg.ace.ctrl-c.liu[.]se:9999?key=5636b3af4738c3998284c4805d91209cab38921159c66a6f359f3f692af1c908 quic://spain.magicum[.]net:36900 tcp://y.zbin[.]eu:7743
welivesecurity.comOct 22, 2026source
The quest for simplicity: Why SMBs want advanced protection without the complexity
Cybersecurity has never been easy. Threats evolve at breakneck speed. Corporate attack surfaces continue to expand with each new digital investment. Regulatory requirements add extra pressure on stretched teams. And behind it all, the unpredictability of human nature means there is no such thing as 100% security. But the market for security solutions doesn’t make things any easier for SMBs. Smaller organizations usually have fewer resources and less expertise to throw at the problem than their larger counterparts. But they absolutely understand the need for advanced protection. It’s just that they also want it to feel simple, guided and manageable. Solutions and services that foreground such an approach are in short supply. But they do exist. SMS understand the importance of cybersecurity “Good enough” security is simply not good enough at a time of persistent geopolitical uncertainty and rapid technological change. ESET data reveals that three-quarters (75%) of SMBs are concerned about global conflicts and cyber warfare. And nearly half have experienced at least one cyber incident in the past year. Whether it’s Iran-nexus threat actors targeting British power and US water plants, or, as shown by recent ESET research, Russia-aligned hackers hitting Polish energy facilities, state-backed attacks are more frequent and more ambitious than they’ve ever been. SMBs are very much in the crosshairs, as providers of critical infrastructure themselves or suppliers to firms operating in critical infrastructure. But that’s only half the picture. Financially motivated groups continue to pose an outsized threat, usually via ransomware or data theft extortion. Often they operate with impunity from hostile states, where they elude the reach of Western law enforcement agencies. There’s growing evidence to suggest they even collaborate with government hackers on some campaigns. Yet the identity of the adversary is arguably less important than the tactics they’re using to target global SMBs. AI is empowering threat actors everywhere to upskill and scale attacks more cost effectively than before. They’re moving through the attack chain at speed, from reconnaissance to exfiltration, collapsing the vulnerability exploitation window, and building highly convincing personalized phishing messages. A new wave of autonomous agent-powered attacks will come next. We know what they’re capable of, because they’ve already broken free of vendor testing environments to socially engineer victims, create fake personas, exploit weak passwords, and deploy malicious code. If rogue agents can do this, then deliberately manipulated agentic AI can too. AI is also the target. As SMBs embrace coding agents, chatbots and AI-assisted workflows, these tools are creating a new attack surface. Do you know what you’re running in your organization? Can you be sure it’s not being manipulated by malicious third parties? And are your users following policy when they interact with the technology? A third (32%) of North American and even more (42%) European organizations admit to having no rules in place to restrict the use of AI applications outside approved processes or platforms. Overwhelmed and overpriced Against this backdrop, it’s perhaps not surprising that ESET finds that 58% of SMBs believe they are more vulnerable to cyber attacks than larger enterprises. It’s not just that they’re facing an unprecedented threat landscape. The pressure is operational as well as technical. SMBs often struggle to make sense of a market crammed with industry jargon, competing marketing claims, and point solutions which address only narrow problems. They buy more technology, but the outcomes don’t improve. The average company might use scores of siloed security tools, all pushing out alerts and impressive-looking dashboards. That does nothing but create more operational overhead, alert fatigue, and training gaps. Without the time, skills and confidence to manage all that technology, SMBs can quickly become overwhelmed. Real threats slip through the net while teams waste valuable time chasing false positives. Critical security updates are left undeployed while teams patch non-essential systems. Important tools are misconfigured. Reducing the burden Smaller organizations want better protection, not another security puzzle to solve. They need tools and services they understand; that are simple to operate and deliver plain-language guidance and explanations on how they work and what they’re doing under the covers. They want 24/7/365 monitoring to ensure that any threat at any time can be caught and contained as soon as possible, reducing the impact on the organization. And they want expert help to understand what threats to prioritize when alerts are flying in and context is in short supply. For many, this will mean outsourcing some parts of the security function to a trusted partner. They may not know the search term. But for many, the destination should be managed detection and response (MDR) which reduces the burden on in-house teams while ensuring no alert is missed. Yet historically, the SMB market has not necessarily been well served by MDR providers. Smaller organizations may have felt locked out of a market that often charges a premium for their capabilities. After all, in-house 24/7 monitoring and threat hunting doesn’t come cheap, especially during a time of skills shortages. Then there’s onboarding. Despite offering simplicity on the surface, many MDR offerings don’t work out like that in practice. They require integration with multiple systems and data sources, the granting of the correct permissions to outsourced SOC staff, the fine-tuning of detection rules, incident response planning, and many other complex, time-consuming tasks. What most SMBs want is continuous monitoring, detection of suspicious activity, clear guidance on next steps, and experts on hand to help with containment and remediation as required. But in reality, it can be much more complicated. For SMBs, this is about finding value for money: security that delivers the outcomes it promises. This matters more than ever at a time of continued business uncertainty, and the threat of cyber disruption which could have lasting consequences. All organizations have to assume that they’ll be breached at some point in the future, no matter how effective their preventative controls. The key to avoiding potentially crippling financial and reputational damage is to contain and recover from any attack as quickly as possible. That means finding security that helps them make better decisions, not security that creates more work. What SMBs want Against this backdrop, SMBs are looking for several complementary capabilities from their security partners. They want products and services that help to: Lower the operational burden by absorbing the day-to-day work that in-house teams are struggling to get through. A provider might do this by automating routine processes such as remediation of common issues. By offering teams AI assistance to improve productivity. And by ensuring there’s always expert help for things like onboarding and intervening when AI and automation don’t deliver. Detect and contain threats faster to maintain resilience. This capability must work round the clock, given that threat actors increasingly look to strike at weekends and during holidays to maximize their efforts. By minimizing dwell time, organizations can reduce the costs associated with an incident. IBM calculates that data breaches last year which took under 200 days to identify and contain incurred costs of $4.32 million, while those with a lifecycle exceeding 200 days had an average cost of $5.65 million. Encourage clearer understanding and better decision making, by not bombarding the customer with too much technical information. They need to know what happened, how serious the incident is, what the service provider has done to remediate, and what else needs to be done by their organization. Too much information can cripple response efforts. Deliver stronger operational continuity with a system which will support resilience by preventing incidents where possible, limiting the impact when they do occur, and ensuring the company can maintain minimum viable operations while recovering swiftly. SMBs have a smaller margin of error than large enterprises, making rapid recovery key to preserving customer trust and minimizing the financial impact. Drive faster time to value through rapid onboarding and deployment, ensuring the SMB can start benefitting from the service as soon as possible. AI and automation can help here, and ensure the exposure window is kept as short as possible. Encourage safer AI adoption by ensuring that all AI in use in the organization is managed and governed securely. That means first understanding what is in use in the organization, what data is being fed into it, and (in the case of agents) what it can access. Then working out policies and controls to minimize risk. A secondary layer of protection works to reduce leakage risks and prompt injection or data poisoning so that existing systems can’t be hijacked by malicious third parties. Increase protection across the corporate attack surface, which could include identity, collaboration tools, remote access, AI usage, SaaS apps, cloud infrastructure and more. A trusted provider will help to manage risk across this distributed and complex IT environment, which could stretch from remote employee laptops to cloud servers. Threat actors are past masters at finding gaps in protection, so coverage must be comprehensive. Deliver the right approach for each individual organization, because no two SMBs are the same. They have different risk profiles, attack surfaces, IT maturity levels and budgets. For some, automated protection is sufficient. For others, a white-glove security relationship with comprehensive support is a better fit. The key is offering choice without confusion, and services aligned to the customer’s risk appetite. The case for supervised security What this translates into is friction-less, supervised security for the AI era. Friction-less because it should reduce unnecessary friction in terms of purchasing, deployment and use. And offer robust protection without the need to interpret complex technical alerts or integrate disconnected point solutions. Supervised, because it should offer a blend of AI, automation and dedicated human oversight, ensuring the customer is in the driving seat but with an expert navigator by their side. And “security for AI” because it should leverage AI to enhance prevention, detection, response, correlation, investigation, and efficiency. But it should also have capabilities to reduce risk across the customer’s AI attack surface. AI is changing the game for security. But it should never be treated as a feature, or a marketing message. It should be embedded into the fabric of what a trusted security partner delivers. Not to replace human expertise, but to enhance it where necessary, making teams more productive and catching threats which their eyes might miss. Ultimately, SMBs want confidence that they are better protected, supported, and prepared. That they are doing everything they can to minimize an intrusion. And that, if a breach does occur, that they can recover as quickly as possible. Beyond this, they need to feel this confidence without the operational burden that is hurting so many smaller organizations. Because security should at the end be a business enabler. Not a burden on growth and innovation. The market is evolving. A better approach is available for SMBs. You just need to find the right security partners; those which focus specifically on meeting the needs of the mid-market with simpler solutions that deliver supervised cybersecurity outcomes.
welivesecurity.comOct 6, 2026source
Timeshare exit scams: From fake buyers to recovery fraud
Timeshares don’t have the best reputation in the world. But many consumers still invest in them. Estimates suggest the global marketplace was worth nearly $20 billion in 2025, and could exceed $38 billion by 2034. Yet many holidaymakers eventually want out, whether because of rising maintenance fees, life changes, or concerns over long-term liabilities. If you want out, there are several legitimate exit paths. You could contact your developer to see if there are surrender or give-back schemes. Or you could list the property on specialized marketplaces or via licensed brokers. Some of the big-name chains might even buy back your timeshare for a percentage of the original value. However, online scammers have caught wind of a growing market of unhappy timeshare owners. And they’re cashing in. What do exit scam companies look like? The way these companies operate is almost a mirror image of a classic timeshare scam. But instead of luring you with the promise of a beautiful timeshare property, they reel you in with the guarantee of a hassle-free exit. Although legitimate companies exist to help you with this process, the market is saturated with scammers. Here are some tell-tale signs: Unsolicited calls made after they find your details by searching through real estate records. Free dinner seminars where you will be subject to high-pressure pitches or fearmongering designed to amplify your concerns over long-term debt. Large upfront fees that can run into thousands of dollars, before any work has even been completed, in schemes similar to advance fee fraud. Guaranteed results or your money back promises, or claims that they know secret legal hacks that can cancel contracts. Unsound advice such as to stop paying your annual fees, which could wreck your credit score. They may also discourage you from checking with an attorney. 5 common schemes Take the fee and vanish: The exit scam company charges a large upfront fee in the thousands of dollars. Once the funds clear they do little or no work and spend months stalling, before the company winds up and disappears, only to reemerge under a new brand later. Fake title transfer to a shell LLC: The scam company promises to transfer your deed out of your name. But instead of conveying it back to the resort, or a legitimate buyer, they transfer it to a shell company. Once the resort finds out, you’ll still be legally liable for any fees, penalties and foreclosure proceedings. Fake mass tort/class action solicitation: The scammers pose as a legal entity claiming to organize a "mass-tort" or "class-action" lawsuit against your timeshare developer to invalidate all contracts. They will demand large legal retainer fees and other associated costs, but usually no lawsuit exists. Fake resale buyer: You get a cold call or email from a broker claiming a buyer is ready to purchase your timeshare in cash and/or at above the market rate. The only catch is that you have to pay money first to cover taxes, wire transfer charges, escrow setup or some other made-up reason. The whole backstory is a pretext to trick you into paying more fees. Recovery scam: Scammers obtain lists of defrauded exit scam victims from other fraudsters and get in touch claiming to work for a government agency or specialized recovery firm. They promise to get your money back or to complete the exit. But this is also a scam. All they want is an upfront fee for processing your case, which is of course an invention. Once paid, they disappear. How to vet a timeshare exit company Although the market is awash with scammers, there are some legitimate companies that are able to help you offload your timeshare. The best way to make sure you’re not dealing with scammers is to: Check that payment to the company is made via escrow once the work has been completed, so that there are no upfront fees. Demand proof of legal credentials, such as the name of the licensed law firm or attorney they use and/or their real estate license number. Be sure to check the veracity of any details they send you. Run some background checks on the company by searching for customer complaints about them. The Better Business Bureau has a ratings system that can help here. Or you could check whether the state attorney general (in the US) has taken legal action against the firm. Other places to check are trade body the American Resort Development Association (ARDA) or ResponsibleExit.com. Review your contract for any suspect language, including vague language versus explicit outcomes, small print about the ‘money-back guarantee,’ and any advice to default. If an exit company cold calls you or makes 100% guaranteed promises of a swift exit, they’re likely to be a scammer. What happens if I’ve already been scammed? If you’ve already paid a scam timeshare exit company, there’s still a chance you could get your money back if you act quickly. Follow these steps: If you paid by card, contact your bank or card issuer and dispute the charge under “fraud” or “services not rendered” or similar. Deadlines vary by card network and issuer, so don’t wait. If you paid by wire transfer, ask your bank to try to recall it immediately. At the same time, contact the bank’s fraud team and make sure any automatic debits are cancelled. Collect evidence of the scam, including all of your communications (emails, letters, texts etc), proof of payment (bank statements and card receipts), contracts and other relevant documents File a report with the FTC, the FBI, and your state attorney general Be prepared for more scams, so don’t engage with any entity promising to get your money back, even if they claim to be from the government Check with the resort to see if the scam company contacted them, and to check if you’re up to date on maintenance fees Frequently asked questions (FAQs) What are the legitimate ways to exit a timeshare? You could contact your developer to see if there are surrender or give-back schemes. You could also list the property on specialized marketplaces or contact licensed brokers. Some of the big-name chains might also buy back your timeshare. What are the warning signs of a timeshare exit scam? Cold calling, free lunch/dinner seminars with high-pressure tactics, large upfront fees, and guaranteed results are all red flags. Scammers may also advise you to stop paying your annual fees, and they will not want you speaking to an attorney. What are some common timeshare exit scams? The company may take an upfront fee off you and vanish. They might transfer your title deeds to a shell company. They could claim to be organizing a “mass tort” or class action lawsuit. Or they may pretend there’s a buyer ready to purchase your timeshare. Beware also for recovery scammers who claim they can get your money back for a small fee. How can I check whether a timeshare exit company is legitimate? Check that payment is made via escrow once work has been completed, rather than an upfront direct transaction. Demand proof of legal or real estate credentials. Run background checks with the BBB, ARDA, the state attorney general, ResponsibleExit.com, or just a simple Google search. What should I do if I've already paid a timeshare exit scam company? Contact your bank and begin chargeback proceedings (you usually have 120 days to do so). Collect all the evidence of the scam and file a report with the FTC, FBI and state attorney general. Be on the lookout for follow-on recovery scams. Could I be targeted by another scam after losing money? Yes. Recovery scammers will claim to work for government agencies or specialized recovery firms. They promise to get your money back or to complete the exit. But they will require a fee or charge to do so, and disappear once you’ve paid it.
welivesecurity.comSep 29, 2026source
Been told to pay at a Bitcoin ATM? Read this first
Many of us still view cryptocurrency as a niche asset. Yet some estimates claim that nearly one in 10 people globally own some. That’s why you may have noticed Bitcoin or crypto ATMs springing up in retail stores, transport hubs, and gas stations over recent years. Today there are tens of thousands in the US alone. Their job is simple: allow you to buy or sell crypto using cash or card. But scammers also have them in their sights. If you ever receive an unsolicited call urging you to deposit money into a crypto ATM, hang up immediately, no matter how serious the allegations. No real government agency, bank, or company will ever tell you to pay them via a Bitcoin ATM. How do Bitcoin ATM scams work? According to the FBI, 13,460 of these scams made fraudsters $389 million in 2025, a 23% increase in volume and 58% in losses over 2024. Most of those complaints came from people over 50, who lost more than $302m between them. Victims over 60 accounted for $257.4m of that total. Older victims are over-represented partly because fewer of them know what crypto ATMs are. But the machine is only the payment method. What makes the scam work is a phone call designed to frighten you and keep you on the line. The scam itself is a variation of an “impersonation” attack. That is, you will usually be called up out of the blue by a fraudster masquerading as a serious-sounding police officer, government employee or other trusted official. Or you might get a popup on your computer telling you to call a number as your machine is infected. Once you’re on the phone, they will try to: Frighten you by saying someone has accessed your bank account, or your identity or account is being used for criminal activity. They may alternatively confirm that your computer is infected. Tell you that your funds are at risk and that you need to move them somewhere ‘safe’. Keep you on the phone and discourage you from seeking help from friends, family or legitimate authorities. Tell you to withdraw your cash from your bank. Direct you to a specific Bitcoin/crypto ATM, and tell you to deposit the cash and purchase Bitcoin/crypto with it. Text you a QR code to scan. In fact, the QR code tells the ATM to send those funds to a crypto account under the criminal’s control. Because it’s crypto, it can be extremely difficult to reverse the transaction. Can I get my money back from a crypto ATM scam? Crypto ATMs are deliberately used by cybercriminals precisely because the funds are not like card transactions: they’re very difficult to reverse and there’s no chargeback facility. Although transaction flows can be traced, wallet holders are able to stay anonymous. All of which makes it very unlikely that you’ll get your money back. How do I stay safe from Bitcoin ATM scams? Steer clear of these scams in the first place by spotting the warning signs. If someone calls you up purporting to be an official, slow them down, get their details, and put the phone down. Don’t fall for their attempts to frighten you and create a sense of urgency. Then call a friend or family member to sanity check the story you’ve just been told. Or verify it by contacting the organization or individual directly. But always use official contact information rather than a number/email address given to you by the caller. Likewise, never respond to unsolicited emails, messages or computer popups by calling the number provided. Always check via official channels only. If you're already standing at the machine with a caller in your ear, end the call and walk away. It goes without saying that you should never take large sums of money out of your bank account and deposit them in a crypto ATM at the request of a stranger who calls you on the phone. Real businesses, government agencies, and law enforcers will never request that you do so. What can I do straight after a Bitcoin ATM scam? If you’re in the middle of being scammed or the worst has already happened, follow these steps immediately: Stop communicating with the scammers. Don’t reply to their messages and hang up Save any evidence you might have including ATM receipts, texts, emails, call logs, and bank transactions Find out the contact details of the ATM operator (which are usually found on the receipt or the machine itself) and tell them immediately what has just happened Contact your bank to tell them what happened and ensure your accounts are safe In the US, report to your local police, and the FTC as well as the FBI’s IC3 The chances of recovery are extremely low. But that doesn’t mean you shouldn’t report what just happened. Reporting is important to help the authorities better understand fraud patterns and provide the best preventative advice to the populace. You should also be on the lookout for recovery scams, where the same or other groups of fraudsters will call you up pretending to be able to get your funds back for a fee. They can’t, and they will simply make off with more of your money. Frequently asked questions (FAQs) Can a Bitcoin ATM transaction be reversed? No, your money usually can’t be reversed as if it were a card payment. The Bitcoin ATM operator doesn’t hold the funds and can’t reverse the transaction. Once it’s gone, it’s gone. Would police or government officials ever investigate Bitcoin ATM fraud? Usually not. They will log your report so you have an official record of the crime if needed. But in most cases, it will not be investigated. Why do scammers demand Bitcoin ATM transfers? It’s harder for the authorities to trace than wire fraud and there’s no chargeback mechanism for victims to get their funds back. It’s also fast and anonymous, while social engineering tricks play on our anxieties and predispose us to believing what we’re being told over the phone. What should I do if I’ve fallen for a Bitcoin ATM scam? Stop communicating with the scammers. Save any evidence you might have including ATM receipts, texts, emails, call logs, and bank transactions. Find out the contact details of the ATM operator and your bank and tell them immediately what has just happened. In the US, report to your local police, the FTC and the FBI’s IC3. Someone says they can get my money back after a Bitcoin ATM scam. Is this for real? Unfortunately not. They’re likely to be a “recovery scammer,” preying on victims of these scams. They’ll ask you for an admin fee or similar which they’ll simply take without getting your money back.
welivesecurity.comSep 24, 2026source
Beware the SparroWock: The backdoor that bites, the commands that catch
ESET Research’s ongoing monitoring of FamousSparrow has borne fruit once again. Our previous public report on FamousSparrow revealed that this China-aligned APT group had developed two new versions of its custom backdoor named SparrowDoor. This time, we discovered that FamousSparrow has switched to a new backdoor, SparroWocky, and has been deploying it to several countries in Latin America since at least August 2025. In what was probably China’s reaction to the US showing increased interest in Latin America, FamousSparrow increased its targeting of the region to almost exclusively targeting it in July 2025. A month later, we noticed that the group had started using the new SparroWocky backdoor, which then quickly replaced SparrowDoor as FamousSparrow’s main implant. SparroWocky is a modular, C++ backdoor. Its architecture and the techniques used by its authors indicate strong knowledge of anti-analysis tricks and Windows internals. We chose to name the backdoor SparroWocky because the first samples we collected all contain the first stanza of Jabberwocky, a nonsense poem by Lewis Carroll. Fortunately, while advanced, SparroWocky’s inner workings are much less arcane than a gyre and gimble in the wabe, so a through and through [of] the vorpal blade allowed us to bring you a detailed analysis of the backdoor. Key points of the blogpost: FamousSparrow is extensively targeting governmental organizations in Latin America. Since August 2025, the group appears to be abandoning SparrowDoor in favor of SparroWocky, a new custom C++ backdoor. With the switch to SparroWocky, FamousSparrow started to incorporate code from open-source projects directly into its malware. SparroWocky is a full-featured backdoor that manipulates low-level structures in memory, and patches code at runtime in order to avoid detection. SparroWocky has the capability to load and execute Beacon Object Files, a special type of executable file supported by many red-teaming and penetration-testing tools. FamousSparrow is a China-aligned cyberespionage group believed to have been active since at least 2019. We first publicly documented the group in a blogpost from September 2021 when we observed it exploiting the ProxyLogon vulnerability. The group was initially known for targeting hotels around the world but has also targeted governments, international organizations, trade groups, engineering companies, and law firms. FamousSparrow is the only known user of the SparrowDoor backdoor. We analyzed two versions of SparrowDoor in a 2025 blogpost, in which we also discussed the attribution claims around the group. As mentioned by Trend Micro, FamousSparrow is linked to Earth Estries; however, the exact nature of the link is not fully known. FamousSparrow has also been publicly linked to Salt Typhoon, but, due to the absence of any technical indicators, we track them as separate. Based on our investigation, we attribute the latest campaign and the SparroWocky backdoor to FamousSparrow with high confidence, since in some of the first attacks involving this backdoor, SparroWocky was deployed by the FamousSparrow-exclusive SparrowDoor. Moreover, not only does the victimology match FamousSparrow’s previous targeting, we have also recorded attempts to deploy SparroWocky at many of the same organizations that had previously been targeted with SparrowDoor. Latin America in the crosshairs As previously mentioned, FamousSparrow currently appears to be focused on high-profile targets in Latin America. This trend started at the latest in July 2025 and has continued with the introduction of SparroWocky. In fact, from mid-2025 and into 2026, 90% of the group’s targets registered in our telemetry have been located in the region. As depicted in Figure 1, we’ve seen the new backdoor deployed against governmental entities in Argentina, Ecuador, Guatemala, Honduras, Panama, Peru, Puerto Rico, and Venezuela. This represents a rare occurrence among the China-aligned APT groups that we currently track, which are generally observed throughout various world regions within such an extended time frame. We believe that this focus is not coincidental and likely reflects China’s reaction to various recent US initiatives in the region. Indeed, Donald Trump’s second presidential term has brought about an aggressive reaffirmation of US interests in Latin America, which threatens various long-term investments that China has cultivated throughout the continent in the last decade, in domains such as energy, mining, and telecommunications. We suspect that FamousSparrow’s activities are intended to help China better monitor and anticipate the reaction of local governments to current US pressures. In some cases, we have observed elements that clearly seem to confirm this hypothesis. For instance, one of the Panamanian entities we’ve seen being targeted is directly involved in the ongoing commercial dispute regarding two major ports located in the canal area, which were, until recently, operated by a China-based company. As the concession granted to this company was legally challenged by the Panamanian government in early 2025, it seems highly likely that FamousSparrow’s operation was intended to gain early, privileged knowledge of local authorities’ intentions on this issue. It is not clear whether the group’s apparent focus on Latin America may reflect a formal, geographical mandate, or whether this focus is only temporary and dictated by the current geopolitical circumstances. Examining SparroWocky SparroWocky is a full-featured, modular C++ backdoor built with modularity and stealthiness in mind. It appeared shortly after FamousSparrow started focusing on Latin America and quickly became the group’s new flagship implant, replacing SparrowDoor. It should be noted that SparroWocky is not a variant of SparrowDoor, but is rather a distinct malware family. The transition to this new backdoor also came with a greater level of integration of open-source tooling into FamousSparrow’s workflow: while previously, standalone versions of these tools were deployed side by side with SparrowDoor, with SparroWocky, some have been incorporated directly into the malware. Some of SparroWocky’s notable features include the ability to execute arbitrary files, to act as a TCP proxy, and to execute commands. The backdoor also collects general information about the compromised machine, such as the computer name, the username, domain name, Windows version, and the IP addresses of its network interfaces. SparroWocky is also capable of exfiltrating files and taking screenshots periodically. Exfiltrated information is encrypted using RC4 and sent over the TLS protocol. Depending on its configuration, SparroWocky can establish persistence either by creating a dedicated service or an entry in a registry Run key. SparroWocky is deployed using the common trident loader scheme, which consists of a legitimate executable, a malicious DLL standing in for one required by that executable, and a file containing an encrypted payload (see Figure 2). The loader resides in the aforementioned DLL and is executed via DLL side-loading. We have seen FamousSparrow use a wide range of side-loading targets; in most cases, a patched version of the legitimate DLL that the executable is supposed to load. While most of the file is left untouched, an arbitrary portion of the .text section is replaced with the malicious code, and the entry point header is changed to point inside this patched region. This has some defense-evasion properties: having the metadata and exported function list of the malicious DLL be the same as that of the legitimate version allows it to more easily blend in. Since the code in the patched region does not align with the exported functions and calls in the untouched portion, automated analysis tools may have trouble recognizing function boundaries. The loader’s main role is to extract and decrypt its payload from a file. These files, which typically have the same name as the executable but with a .dat extension, have a specific structure detailed in Figure 3. The file has a custom header that begins with a four-byte magic value of 0x11328712, followed by the size of the configuration data, the size of the payload, and a 16-byte RC4 key. This RC4 key is used to decrypt the remainder of the file, which contains the configuration for SparroWocky (detailed in the Configuration section) and the backdoor itself. We provide a script to decrypt SparroWocky payload files in our GitHub repository. The plaintext backdoor payload is formatted as a portable executable (PE) file with the MZ and PE magic values deleted. This executable payload is reflectively mapped directly into memory without being written to disk. Thus, we believe that stripping these magic values is possibly an attempt to evade in-memory defense mechanisms that use simple pattern recognition to identify or dump suspicious sections of memory. SparroWocky Our analysis of SparroWocky is mainly based on a sample compiled on November 17th, 2025 according to its PE timestamp (SHA-1: 44F0A22B143B79FA760BF31E14C8FFF714C8A2A1). The version of this backdoor appears to be 1.8, based on the information collected by its fingerprint command explained in Table 3. As we already mentioned, we chose the name SparroWocky because we found the first stanza of Lewis Caroll’s Jabberwocky in several samples we collected. We believe that this stanza comes from the test vectors in RFC 7539, which defines the ChaCha20-Poly1305 encryption algorithm. The samples of SparroWocky also contain other strings that are used as test vectors in that RFC. However, SparroWocky does not use ChaCha20-Poly1305. While we don’t know the exact version of Mbed TLS used in the backdoor, the test vectors were present in that library prior to version 4.0.0. Notably, SparroWocky relies at least on the following public projects: Mbed TLS, a C library it uses to establish a secure communication channel with its C&C server, MinHook, a Windows API hooking library it uses to hide the start address of newly created threads from security products, and COFF Loader (or a similar project) that it uses to enable dynamic loading and execution of in-memory plugins in the form of COFF objects. Additionally, our analysis revealed that the developers implemented various techniques to evade monitoring tools. This includes a variant of a technique called SilentMoonwalk (or StackMoonwalk), which allows SparroWocky to spoof the call stacks originating from MinHook routines. The backdoor also uses a custom API-hashing algorithm to dynamically resolve Windows API functions. These are explained in greater detail in the Anti-analysis techniques section. The SparroWocky loader extracts and decrypts its configuration from the payload .dat file located in the same directory, as explained in the Loader section. The RC4 key stored in the payload header is used to decrypt the configuration, which is provided in the form of a tab-separated string that is then parsed and stored in a structure. The fields and their values are described in Table 1 in order of appearance. Table 1. SparroWocky configuration Capabilities After parsing its configuration, the backdoor checks the command line of the process in which it is running and behaves differently based on the number and value of the arguments passed. If no arguments are present, SparroWocky simply sets up persistence and executes the core logic of the backdoor. Otherwise, the value of the first argument directs the malware to follow specific instructions, as described in Table 2. Table 2. SparroWocky command line arguments and their meaning Explained in the Backdoor commands section. When SparroWocky is executed with the c option, it reads an additional comma-separated list of parameters from standard input (stdin): a command string, a timeout delay (in seconds), and optionally, the body of a PE file. If this last parameter is absent, the backdoor reads the executable specified in the command string from C:\Windows\System32\ and loads the associated English MUI (Multilingual User Interface) file (from C:\Windows\System32\en-US\). This process is described in the Host process camouflage for dynamically loaded PEs section. Otherwise, the PE file is executed by SparroWocky’s reflective loader, and the command string is passed as a command line. This functionality is likely meant to allow the backdoor to easily execute system utilities. SparroWocky loads the specified executable into memory and executes it with the provided command. At the same time, the backdoor creates a new thread that calls ExitProcess to kill the process when the timeout delay expires. The loading process involves setting up hooks and forging structures in memory to camouflage the host process before running the target executable. These anti-analysis tricks are explained in greater detail in the dedicated Host process camouflage for dynamically loaded PEs section. Additionally, when the malware is executed without arguments or with the p option, an instance synchronization mechanism is started. This feature prevents multiple instances of the backdoor from running concurrently by leveraging a custom interprocess communication (IPC) mechanism. When a new instance is launched, the currently running instance stops and, if the new instance is launched from a different location than the current one, the files and persistence configurations set by the currently running instance are deleted. This is achieved by using three types of global objects: a mutex, an event, and a shared memory block named MyMutexName, MyEventName, and MySharedMemName, respectively. The backdoor first establishes communication with its C&C server, then executes its core logic in an infinite loop, within which it processes received commands. These are handled by a custom class named WinHandler (derived from a ServerHandler custom class), according to the runtime type information (RTTI) present in the malware. Handlers for a minimal set of commands are hardcoded in the command loop itself. ServerHandler has a dedicated virtual method to handle more commands. This method is implemented in WinHandler. While we have not observed other implementations of this method, this architecture would make it easy for its developers to change the set of commands that the backdoor can handle. The list of supported commands is shown in Table 3. Table 3. SparroWocky commands Hardcoded command. Command 0x14 uses a slightly modified version of RunCOFF from the open-source COFF Loader project to load and execute a Beacon Object File (BOF). A BOF is a position-independent Common Object File Format (COFF) executable that is meant to be run within the memory of an implant. BOFs were first introduced in Cobalt Strike and have since been adopted by other popular red-teaming frameworks such as Brute Ratel, Metasploit, and Sliver. The change to RunCOFF resides in the resolution of imported symbols. SparroWocky redirects calls to external libraries in the BOF to a stack-spoofing subroutine. This effectively hides and proxies calls made by the BOF object. Once the object is loaded, the BOF loader finds and executes function_name, passing the arguments provided in function_arguments. The ability to load BOFs allows FamousSparrow to use existing modules and tools designed to work with this file type. Self-deletion As described in the Argument-controlled behavior section, SparroWocky can delete itself entirely from the system. This can be done from the C&C server via command 0x13. First, the persistence mechanism previously set is removed and then the batch file shown in Figure 4 is created and executed. @echo off timeout /t 2 del " " /f /q del " " /f /q del " " /f /q del "%%0" /f /q\n Figure 4. Batch file for self-deletion This deletes the files used by the backdoor: the legitimate executable, the side-loading library, and the payload file. The batch file deletes itself at the end of the script. SparroWocky employs a few techniques to complicate its analysis and to evade security software that may be in place. A common technique that the backdoor uses is dynamic API resolution via API hashing, but the backdoor also uses more interesting ones, described below. SilentMoonwalk The first noteworthy technique is called SilentMoonwalk, which essentially provides a way to forge fake call stacks. Its purpose is to prevent analysis tools and products from inspecting the true caller of specific functions that are frequently monitored, such as Windows API functions. This method requires a few initialization steps: Finding the offset of RtlUserThreadStart and BaseThreadInitThunk, two functions that are usually found at the start (or bottom) of any call stack. Finding a JOP (jump-oriented programming) and a ROP (return-oriented programming) gadget in the legitimate kernel32.dll library to restore the original call stack. Once these requirements are met, when SparroWocky makes an obfuscated call to a Windows API function, it first saves the current context (registers); next, it forges a fake stack using the gadgets found previously, and then inserts the address of a stack and context restoration routine. This makes it appear as if the calls to Windows API functions are originating from RtlUserThreadStart and BaseThreadInitThunk. Figure 5 shows the call stack view from a debugging session using WinDbg. In the case of SparroWocky, this technique is used to obfuscate calls made by BOF-formatted plugins (command 0x14) or by the statically linked MinHook hooking library. Concealing the thread start address SparroWocky uses the MinHook library to hook the CreateThread function in order to conceal the original lpStartAddress parameter from security products. Essentially, any thread created by SparroWocky would have AnimateWindow as the starting address, which would likely be considered legitimate by a security product. The patch applied to AnimateWindow turns it into a trampoline that simply executes the original start address, as illustrated in Figure 6. The last notable piece of code from SparroWocky is its custom PE loader, used when executed with option c. While implementing PE loaders is pretty much routine for malware authors, SparroWocky authors took it a step further and integrated host process camouflage. As described in Table 2, when SparroWocky is executed with the c option, it loads a PE file in memory and executes it. If the file is not passed as an argument, the PE loader parses the specified command line to extract the file’s name. It searches for that filename in the C:\Windows\System32\ directory, but most importantly it retrieves the English localization MUI file associated with the target PE (stored as C:\Windows\System32\en-US\ .mui). In that case, the PE file is loaded in memory, and a few hooks are set to make sure any calls made by the loaded PE file to retrieve resource data, such as RtlLoadString or RtlFindMessage, are redirected to the .mui data. This process mirrors normal behavior of Windows when loading PEs, and reduces the risk of unexpected errors. The command line retrieved from Stdin is parsed and SparroWocky hooks the following functions, which are used to retrieve information about command line arguments, to make them point to this command line: GetCommandline[AW] __(w}getmainargs p_argc p_{w}argv The PE loader is also able to register the exception handlers of the newly loaded executable – an unusual, yet critical, addition – since it allows exceptions to be handled correctly. Finally, before calling the entry point of the loaded PE file, SparroWocky forges and inserts a fake LDR_DATA_TABLE_ENTRY structure in the doubly linked list of the PEB_LDR_DATA structure. This doubly linked list is used by Windows to keep track of loaded modules and is usually monitored by security products. Figure 7 shows a snippet of the code used to set some of its fields. This last technique shows that SparroWocky authors possess a deep understanding of the Windows PE loading mechanism and are willing to go the extra mile to camouflage the host process and confuse monitoring software. Network protocol To communicate with its C&C server, SparroWocky uses the TLS encryption protocol. Under the hood, the backdoor uses the Mbed TLS library and the only element worth mentioning is that it uses the personalization string acdbenus when initializing the deterministic random bit generator, as seen in Figure 8. Before the initial TLS handshake, a TCP connection is established with the C&C server using one of three connection modes: A connection mode of 0 means that SparroWocky uses the proxy currently configured on the machine or a direct TCP connection if no system proxy is configured. This configuration is retrieved by querying the ProxyServer registry value located under the registry key HKCU\Software\Microsoft\Windows\CurrentVersion\Internet Settings. If the connection to the proxy server is not successful, SparroWocky tries to connect via mode 1, then mode 2. Connection mode 1 represents a connection via an HTTP proxy. This connection uses either the Negotiate (Kerberos or NTLM) or Basic authentication scheme with the username and password provided in the configuration. Both authentication methods use generic HTTP headers with the User-Agent string set to Mozilla/5.0. Connection mode 2 uses a SOCKS5 proxy without authentication (AUTH field set to 0x00) or with a username and password (AUTH field set to 0x02). The values used by the latter are provided in the configuration. Command messages Once the TLS handshake is complete, SparroWocky sends the bytes 0x11223344 (big-endian) to indicate that it is ready to receive commands in the main session. The backdoor uses a simple format to receive commands and send results, as illustrated in Figure 9. If the command_arg_size field does not equal 0, then additional data is to be received or sent after the header. In that case, the data (command arguments or results) is encrypted via RC4, and each command message uses a newly generated eight-byte key, which is sent in the header. Network infrastructure SparroWocky uses the IP address of its C&C servers, which is generally running on port 443, to connect directly. We have also seen it running on port 8080 in some cases. While we have observed some self-signed certificates being reused across multiple servers, we do not have a reliable generic fingerprint. Conclusion Over the latter half of 2025 and the first half of 2026, FamousSparrow had been focusing on targets in Latin America. This represents a shift from its previous global targeting. To go along with this change, the group has developed SparroWocky, which replaced SparrowDoor as its main implant. While it doesn’t appear to be based on the same codebase, we can see that SparroWocky still shares some of the functionality and concepts that were present in the group’s previous backdoor, which we analyzed in our previous blogpost. SparroWocky uses more complex defense evasion techniques to stay under the radar. FamousSparrow still uses open-source offensive tooling for its own malicious ends. Previously, these tools were mainly used side by side with the group’s backdoor. With SparroWocky, we can observe that it also has the development capabilities to integrate open-source code directly into its own custom backdoor. IoCs A comprehensive list of indicators of compromise (IoCs) and samples can be found in our GitHub repository. Files Network MITRE ATT&CK techniques This table was built using version 19 of the MITRE ATT&CK framework.
welivesecurity.comSep 17, 2026source
AI-assisted reconnaissance: Why everyone could be a viable target for fraud
AI is changing the threat landscape, one model at a time. The pace at which the technology is evolving means activities that once took a skilled cybercriminal several hours or even days to achieve can now be done in a fraction of that time by someone with little or no prior knowledge. This matters, because it’s lowering the barrier to entry for cybercriminals. We’re seeing it with vulnerability exploitation. We’re seeing it with basic malware generation. And we’re seeing it with victim reconnaissance. This should put even regular internet users on high alert. If attacks are getting cheaper and easier for fraudsters to carry out, everyone could be a target. How AI is changing the game There was a time when open-source intelligence gathering (OSINT) was something only used on high-profile and potentially high-value targets. That’s because it took skill and time to carry out. The term itself originates outside of the cyber world, referring to military and intelligence efforts to collect publicly available information which could be used to achieve strategic goals. With the advent of the web, and more recently social media, these efforts have become a familiar part of the digital world, used by researchers and threat actors alike. But until AI came along, there was often still a major bottleneck. An adversary needed to spend time finding their targets’ public-facing accounts, interests, friends, family and colleagues. They might have to use dedicated tools or trawl through countless websites, accounts and posts, correlating and cross-checking information, working out relationships between individuals, and deciding what to use in an attack. AI tools have largely removed that bottleneck by processing information at machine speed. They can find publicly available material from across the web, link it to their potential victims, and map relationships between them and others. It’s particularly good at hunting down unstructured information – especially images and videos. You don’t have to be an OSINT specialist any longer to do this kind of work. Putting the pieces together at scale This is troubling news for a variety of reasons. Fraudsters can now easily collect publicly available information on you to make social engineering more convincing and scams more scalable. Consider: A phishing email designed to trick you into sharing your logins or clicking through and unwittingly installing malware. It would be much more convincing if it contained personal details such as your workplace or child’s school, a recent event you were at, or some contextually relevant news you’ve posted like a birthday or a recent holiday. All of that information may be in the public domain and easy to gather at scale. A video or phone call impersonating your voice/face might be used to trick a loved one into believing you’re in danger. Scammers can use these deepfakes to ask for money in your “voice”, or to pretend that you have been kidnapped and need them to pay a ransom to ensure your safety. They might even use a deepfake of you to create an obscene video which they threaten to send to your contacts unless you pay up. This kind of sextortion scheme is becoming more commonplace. In the UK, hundreds of under-18s have reported being victimized this year. Unfortunately, large language models (LLMs) are expert at piecing together the kind of information that fraudsters need to make their scams work. They can profile large numbers of potential victims in little time, collecting relevant pictures, videos, and info on personal interests, work, and family and friends that could be leveraged. AI can also help to design the social engineering scripts used by fraudsters, enabling them to sound convincing over email/social media or other channels even if they’re non-native speakers. It offers an end-to-end fraud pipeline. Trouble at work It’s not just your personal life that malicious actors can trawl through in this way. The boundary between work and home has become increasingly blurred in recent years, especially as many of us work in a hybrid setup. We might use personal devices and home addresses for corporate activities. And of course, linking our professional and personal social media accounts is a simple task for AI. All of which means that reconnaissance efforts can have an impact on your professional life. For example, fraudsters could use personal information to craft a social engineering attack designed to harvest your work credentials or information. Or they could target your colleagues when you’re on holiday, knowing that you may not be contactable to verify fraudulent details. This is a useful time to launch a business email compromise (BEC) attack. Losing your personal information is one thing. But OSINT efforts which have a corporate dimension could potentially have a serious impact on your professional reputation and career. What you can do about AI-powered OSINT When it comes to AI-powered reconnaissance, there are things you can control and things you can’t. Once information about you is in the public domain, it can be extremely challenging to request its removal, especially as it may have been republished in other places. Old social media photos are easier to remove, but few of us have the time to trawl through our entire digital life to remove anything potentially useful to criminals. It’s better therefore to focus on the things that are within your power to change. Consider the following: Ensure your social profiles can’t be publicly accessed, to limit AI’s ability to find any information or images/videos contained within Be judicious in what you share on social media; things like birth dates, children’s schools, holidays and similar events should be off limits Be aware that photos may contain information in them that could be used to identify addresses, vehicle details etc Avoid posting anything that may be used to link your personal and professional lives Use multi-factor authentication and strong, unique passwords to add an extra layer of security on your accounts Don’t accept friend/follower requests from anyone you don’t know. Or if you’re curious, approach them via a separate channel AI is changing many facets of our lives for the better. But it comes with risks we’re only just waking up to. Caution is always the best policy. It pays to work under the assumption that anything you publish could be read by AI. Act accordingly, and encourage your friends and family to.
welivesecurity.comAug 27, 2026source
Black Hat USA 2026: Will vulnerability discovery eventually decline in the AI era?
The accelerated discovery of previously unknown software vulnerabilities has been making headlines for months. It’s an issue that has even led the US government to create a vulnerability clearing house named Gold Eagle to coordinate research efforts in vulnerability discovery, mitigation and fixes. An indication of the broader pressure facing cyber-defenders can be drawn from the sheer number of patches being delivered in Microsoft’s Patch Tuesday through the last four months: 169 CVEs in April, 118 CVEs in May, 571 CVEs overall in June (including 208 direct Microsoft CVEs) and another 622 vulnerabilities in July that included zero-days under active exploitation. A keynote at Black Hat USA 2026 detailed research by associate professor Yan Shoshitaishvili and his undergraduate students at Arizona State University on the expanding use of AI models for vulnerability discovery. Mr. Shoshitaishvili referred to a Washington Post article from June that stated that Anthropic’s next-generation model Claude Mythos had discovered 479 vulnerabilities in the Linux kernel, which the university team used as a benchmark. Using previous generations of GPT models, meanwhile, the team had discovered ‘just’ around 300 flaws. The difference was attributed to the use of workflows in Mythos, so the team set about integrating similar workflows into three GPTs, which resulted in the discovery of around 600 vulnerabilities. The team then trained the GPTs using the properties of previously known vulnerabilities and discovered approximately 1,000 vulnerabilities. They hit the barrier of discovering vulnerabilities at such speed that they could not keep pace reporting them; for clarity, reporting means detailed research and proposed fixes, rather than just the issue itself. The scale calls into question the whole process of responsible disclosure, which in the team’s view was already broken as disclosure often creates increased risk. Patching software in a timely fashion in production environments was already a stress point for many cybersecurity teams. Exponential growth like this could be the breaking point that causes either more unpatched software and greater opportunities for cybercriminals or patching without testing, which, in turn, could cause compatibility issues in many environments. If I take a logical view of this issue and adopt an optimistic mindset, then it could be that we are heading towards a peak in discovery – and that somewhere over this peak is a meadow of peace and calm with an improved normality. Humans researching vulnerabilities has traditionally been a resource-intensive process, producing a steady stream of discoveries that have been increasing year on year. This is potentially due to there being more researchers, more software and more motivation to discover the vulnerabilities for financial gain through bug bounty programs and such like. Switch from humans to AI, and it’s like a quantum approach to discovery, but note that AI is still in a learning phase: as detailed by the Arizona team, tweaking the model and its workflow potentially uncovers more vulnerabilities. Then there’s also legacy software. Consider the enormous volume of software written over the past 30 years – no amount of human effort could possibly uncover all the vulnerabilities in the current and back catalogues of software. The scale of AI-assisted discovery, however, could potentially reach the end of the catalogue at some stage, and then new discoveries would only be through improvements to the model being used to unearth the vulnerabilities. Let’s not forget that new software is being developed all the time, of course. Here, too, of course, logic should suggest that any development team today would use the same available AI capabilities to remove any potential vulnerabilities prior to releasing their software. And as the models improve, the prospect of virtually flaw-free software could become a reality. If this logic prevails, we may reach the calm and peaceful meadow with very few new vulnerabilities being found. This view could, of course, be just a dream, or my misplaced optimism.
welivesecurity.comAug 13, 2026source
Why the tech industry needs to stand firm on preserving end-to-end encryption
The UK Government wants access, when requested, to the end-to-end encrypted messages and data for everyone in the UK. The reasons are to specifically tackle serious crimes, such as terrorism and child sex abuse. The UK Government is not alone in this, of course, as other countries are also grappling with how to address similar problems in their own jurisdictions. To enforce such a requirement, however, tech companies would need to provide a backdoor – something that is either highly unlikely or never going to happen, at least according to the current stance of most tech companies. The alternative would be to have specific app developers comply with the requirement, but this would only work for local apps tied to a country’s app store location settings. Demanding the impossible Put simply, restricting end-to-end encryption on a single-country basis is inherently unenforceable. What happens when someone from another country visits a restricting country? Would they need to unencrypt, download a new app, delete the encrypted content, or use some other method to comply? The only method to enforce such a law would be at the border… can you imagine the lines at ‘device immigration’? This issue was highlighted when Apple withdrew Advanced Data Protection (ADP) from the UK marketplace back in February. It transpired that the UK Government had issued a non-public notice to Apple under the investigatory Powers Act, asking for access to such data, which would have required a backdoor to be built into Apple’s encryption service. Apple’s response was unequivocal, however: “We have never built a backdoor or master key to any of our products or services and we never will.” ADP uses end-to-end encryption, meaning only the account holder can decrypt files. Recently, WhatsApp threw their support behind Apple in its fight. The issue of breaking encryption with a backdoor should not be shrouded in secrecy like the non-public notice issued to Apple, as this concerns a fundamental privacy and security issue. There are times for secrecy, and I am sure there will be specific cases when data is accessed using the legislation that could, depending on circumstances, be kept secret. Currently, the tech industry continues to stand by their principles of providing customers privacy and security products without backdoors, which, in my opinion, they should continue to do. The UK government’s stance, though, is that all people, when physically in the UK and regardless of citizenship, should be answerable to a UK court. Apple’s removal of ADP for UK users does not fulfill the requirement. If you are a UK iPhone user, then ADP has been removed and is now greyed out and no longer available to you. The method used to determine if a user is in the UK seems not to be based on their location – it appears to rely on the ‘country and region’ you have set in your Apple account. Simply switching your country and region to somewhere other than the UK re-enables the option to turn on ADP. There are some downsides to this, such as the App Store only offering apps from the selected country and region, so you may not be able to download all the apps you need. You can then enable ADP and then switch countries again and ADP remains active. But, if the UK courts and legal system should apply to all those in the UK, then it will need to include visitors and not be based on ‘country and region’. This is not so simple, however: once you enable encryption, to disable it you need to decrypt the data before switching off the encryption, otherwise the encrypted data remains encrypted and unreadable. Border chaos It’s not realistic to force everyone entering a country to provide access to their encrypted messages, especially when they’re carrying a device from a country and region where there is no legislation requiring government access to encrypted data. To enforce it at the border, each person entering the country would need to unencrypt end-to-end encrypted data and disable any apps or features that use end-to-end encryption where there is no backdoor. Every border agent will need to be a tech wizard, and if every visitor is carrying two or three devices, the agent will need to go through each device meticulously to ensure compliance. In other words, each border agent might be able process one individual every few hours. Again, can you imagine the chaos and lines at border control? And then there are people like me. I have two phones, both are on a UK carrier network, one has a country and region setting of the United States and the other to the United Kingdom. ADP is only available to activate on one of them. This means circumventing the current restriction is remarkably simple, and for those who wish to use ADP, whether for legitimate privacy concerns or for criminal activity, there really is no barrier – they just need to seek out this very simple solution. I am assuming there will never be a requirement that forces all visitors to stop using end-to-end encryption services as they enter the country, especially as the services are legal in the countries they reside in. It’s just too complicated to enforce. And, because it’s far too easy to make yourself appear to be located somewhere other than the UK, then those with criminal intent who wish to use end-to-end encryption will continue to use services designed for use in other countries or will find alternatives that strengthen their security even further. This results in just the law-abiding residents of countries enforcing this type of legislation being subject to government and law enforcement access to their data if required. The demonstrable ease of bypassing the requirement, coupled with the impossible logistical burden of its enforcement, make that requirement, at least in my mind, fundamentally unfit for purpose. ESET believes that strong encryption is essential for protecting personal privacy, securing sensitive data, and preventing cybercrime. When one government mandates weakened encryption, others may follow, including those with fewer safeguards for citizens.We must strike the right balance: protecting privacy while ensuring law enforcement has the necessary legal tools to uphold public safety. Instead of backdoors that risk weakening security for everyone, we support a system where law enforcement can access data through court warrants, backed by robust oversight mechanisms in place to ensure both security and safeguards for users.
welivesecurity.comAug 1, 2025source
This month in security with Tony Anscombe – July 2025 edition
With another month behind us, it's time for ESET Chief Security Evangelist Tony Anscombe to look at cybersecurity stories that made an impact and offered vital lessons in July 2025. Here's Tony's rundown of some of what stood out most over the past 30 or so days. attacks targeting on-premises Microsoft SharePoint servers by exploiting the ToolShell zero-day vulnerabilities – and the implications of these attacks for businesses, Lumma Stealer, previously disrupted in a global operation that also relied on ESET's expertise, has returned, as also confirmed by ESET research, KNP, a UK transport company with a history of 158 years, has gone out of business after cybercriminals compromised its computer systems with ransomware in an attack that began by guessing an employee's password, a vulnerability in McHire, McDonald's chatbot job application platform, that exposed the chats of more than 64 million job applications across the United States, with the platform's admin panel using "123456" as both its username and password, so-called "PerfektBlue" critical vulnerabilities in a widely used Bluetooth stack that could have allowed remote code execution on millions of cars, the UK government's proposal to ban public sector bodies and critical infrastructure organizations from paying ransoms after ransomware attacks. Don't forget to check out the June 2025 edition of Tony's monthly security news roundup for more insights.
welivesecurity.comJul 31, 2025source
The hidden risks of browser extensions – and how to stay safe
What would we do without the web browser? For most of us, it’s our gateway to the digital world. But browsers are such a familiar tool today that we’re in danger of giving them a free ride. In fact, there are plenty of rogue extensions masquerading as legitimate ad blockers, AI assistants, or even security tools that are designed to steal our data, send us to malicious sites and flood our screen with popups. For example, earlier this year, a malicious campaign was uncovered that may have impacted dozens of extensions and compromised nearly three million users. Next time you’re thinking about downloading a web browser add-on, think through the following risks. Why extensions matter Browser extensions are an increasingly popular vehicle for threat actors. They give attackers access to a vast amount of sensitive information, with people often trusting these add-ons, especially if they’re downloaded from official sources. Also, extensions provide multiple avenues for monetization and malicious activity and generally give attacks a better chance of success and are a threat also in corporate settings, where they may often stay under the radars of security teams and tools. However, by installing and granting an extension permissions, you could unwittingly be enabling malicious actors to access your most sensitive data – everything from browsing history to saved logins and session cookies, which could be abused to hijack your accounts. When browsers go bad A 2023 risk assessment of 300,000 browser extensions and third-party OAuth applications used in corporate environments revealed that half (51%) of the former were high risk and could potentially have caused “extensive damage.” So how could they end up on your machine? Malware may be hidden in legitimate-looking browser extensions like those purporting to be ad blockers or PDF converters or even security enhancements. They could be packaged up and placed on browser stores for unwitting users to download, bundled with other software, shared through deceptive links or uploaded to platforms outside your official web store, where hackers rely on users “sideloading” in order to target them. Sideloading is particularly dangerous because third-party stores don’t feature the kind of security reviews and other checks that official marketplaces have in place. That means they’re more likely to feature harmful add ons spoofed to appear as if legitimate. Alternatively, threat actors could hijack or acquire a legitimate extension and use it to send malicious updates to its entire user base. Sometimes, extensions can seem legitimate, but on activation will be programmed to install new payloads with malicious capabilities. What can malicious extensions do? The nefarious actions run the gamut and include: Stealing data, including usernames and passwords, browsing history, session cookies (which can be used to access your accounts without needing a password) and financial information. This may be sourced from your clipboard, browser or obtained via keylogging as you type it in. The end goal is usually to either sell that data on the dark web, or use it directly to hijack accounts and commit identity fraud. Directing you to malicious or risky websites that may harbor malware including infostealers and banking Trojans. Other sites may be spoofed to appear as if a legitimate brand, but are actually designed to harvest your personal and financial information and/or logins. Injecting unwanted ads and possible malware into your browsing experience. Ads could be monetized by threat actors, while malware may be designed to steal credentials or harvest other lucrative personal data for identity fraud. Backdooring your browser so that they can access your machine at any time in the future. Mining for cryptocurrency without your knowledge, something that can slow down or even wear out your machine completely. Staying safe To mitigate these risks, caution is always advised when you’re on the hunt for a new extension. First of all, stick to legitimate web stores and closely scrutinize any new add-on. That might include checking the developer’s credentials, reading reviews of the product and searching separately for it to see if it has been connected to any suspicious or malicious behavior in the past. Look closely too at its permissions. If it requests any that seem to go beyond what is needed for the product, it should be a red flag. As is the case with, for example, mobile apps, not many extensions should need access to your passwords or browsing data. Additional tips to keep yourself safe include: Keep your browser updated so it’s on the latest, more secure version at all times. This means it will be better protected against potential malware. Switch on multi-factor authentication on all your online accounts – that will go a long way toward keeping you safe even if a malicious browser extension does steal your passwords. To make your web browsing experience safer in general, consider using a secured browser mode that is offered together with other security-enhancing features by some security vendors. This mode comes in particularly handy when you perform financial and crypto transactions in your browser. Enhanced Safe Browsing in some common web browsers can also help you steer clear of malicious sites. Importantly, use security software from a reputable vendor, and perform periodic scans to check for anything suspicious running on your computer. It will go a long way towards preventing you downloading malware from third-party sites, or redirecting to a phishing site. Every piece of software we install, no matter how small, comes with an element of trust; indeed, this trust may be particularly significant with browser extensions, as they operate directly within your gateway to the internet. Think carefully about the value or convenience that an extension provides versus the potential risk. Ultimately, the goal is to make informed choices about the add-ons you allow into your digital space. be sure to source your browser extensions and, indeed, all other software from reliable providers.
welivesecurity.comJul 29, 2025source
ToolShell: An all-you-can-eat buffet for threat actors
On July 19th, 2025, Microsoft confirmed that a set of zero-day vulnerabilities in SharePoint Server called ToolShell is being exploited in the wild. ToolShell is comprised of CVE-2025-53770, a remote code execution vulnerability, and CVE‑2025‑53771, a server spoofing vulnerability. These attacks target on-premises Microsoft SharePoint servers, specifically those running SharePoint Subscription Edition, SharePoint 2019, or SharePoint 2016. SharePoint Online in Microsoft 365 is not impacted. Exploiting these vulnerabilities enables threat actors to gain entry to restricted systems and steal sensitive information. Starting from July 17th, ToolShell has been widely exploited by all sorts of threat actors, from petty cybercriminals to nation-state APT groups. Since SharePoint is integrated with other Microsoft services, such as Office, Teams, OneDrive, and Outlook, this compromise can provide the attackers a staggering level of access across the affected network. As part of the attack, the threat actors often chain together four vulnerabilities: the previously patched CVE‑2025‑49704 and CVE-2025-49706, alongside the already mentioned CVE-2025-53770 and CVE-2025-53771. As of July 22, CVE‑2025‑53770 and CVE-2025-53771 have also been patched. Webshell payloads Exploiting ToolShell allows the attackers to bypass multi-factor authentication (MFA), and single sign-on (SSO). After getting inside the targeted server, attackers were seen deploying malicious webshells to extract information from the compromised system. One of the scripts frequently used for this purpose is named spinstall0.aspx, which we track as MSIL/Webshell.JS. Additionally, on July 22nd, 2025, we observed that attackers attempted to deploy other simple ASP webshells capable of executing attacker-supplied commands via cmd.exe. These webshells were deployed using the following filenames: ghostfile346.aspx, ghostfile399.aspx, ghostfile807.aspx, ghostfile972.aspx, and ghostfile913.aspx. ESET products first detected an attempt to exploit part of the execution chain – the Sharepoint/Exploit.CVE-2025-49704 vulnerability – on July 17th in Germany. However, because this attempt was blocked, the final webshell payload was not delivered to the targeted system. The first time we registered the payload itself was on July 18th on a server in Italy. As seen in Figure 1, we have since observed active ToolShell exploitation all over the world, with the US (13.3% of attacks) being the most targeted country according to our telemetry data. Attack monitoring Our monitoring of the ToolShell attacks from July 17th to July 22nd revealed that they were coming from the IP addresses shown in Table 1 (all times are in UTC). Table 1. Attacker IP addresses Figure 2 shows the timeline of the attacks coming from the three most active IP addresses. Concerningly, Microsoft has reported that several China-aligned threat actors have joined in on the exploitation attempts. From our side, we detected a backdoor associated with LuckyMouse – a cyberespionage group that targets mainly governments, telecommunications companies, and international organizations – on a machine in Vietnam targeted via ToolShell. At this stage, it remains unclear whether the system had been previously compromised or if the backdoor was introduced during the current attack. Nevertheless, China-aligned APT groups have certainly seized the opportunity to add the exploit chain to their arsenals: according to our telemetry, the victims of the ToolShell attacks include several high-value government organizations that have been long-standing targets of these groups. Since the cat is out of the bag now, we expect many more opportunistic attackers to take advantage of unpatched systems. The exploit attempts are ongoing and will surely continue. Therefore, if you are using SharePoint Server, the following is recommended (as per guidance from Microsoft): use only supported versions, apply the latest security updates, make sure that Antimalware Scan Interface is turned on and configured properly, with an appropriate cybersecurity solution, and rotate SharePoint Server ASP.NET machine keys. For any inquiries about our research published on WeLiveSecurity, please contact us at [email protected]. ESET Research offers private APT intelligence reports and data feeds. For any inquiries about this service, visit the ESET Threat Intelligence page. IoCs A comprehensive list of indicators of compromise (IoCs) and samples can be found in our GitHub repository. Files Network MITRE ATT&CK techniques This table was built using version 17 of the MITRE ATT&CK framework.
welivesecurity.comJul 24, 2025source
Rogue CAPTCHAs: Look out for phony verification pages spreading malware
Bots have got a lot to answer for. They now make up over half of all internet traffic, and while some, such as Google’s web crawlers and fetchers, have legitimate purposes, nearly two-fifths are considered malicious. Their power can be harnessed for everything from posting inflammatory social media posts to launching distributed denial-of-service attacks and hijacking online accounts using, for example, previously breached passwords. So when we’re presented with what’s known as a CAPTCHA challenge, which websites commonly use to deter bots, many of us follow the instructions and click through. All the better to keep the bots out, right? Well, not always. Sometimes the page itself is a fake and could land you in deep trouble. This could, for example, be the case with ClickFix, a social engineering technique that has taken the threat landscape by storm recently. Using fake CAPTCHA images, ClickFix spreads all manner of threats, including infostealers, ransomware, remote access trojans, cryptominers, and even malware from nation-state-aligned threat actors. Why the CAPTCHA threat works CAPTCHA threats work for several reasons: It exploits our familiarity with the process and our trust in CAPTCHA as a way to keep the digital world safe and secure. It also exploits the fact that many of us are impatient when we’re browsing: we just want to access the content we came for and CAPTCHA is in our way, hence we’re more likely to do what it says. It takes advantage of the fact we’re used to clicking through multiple steps to verify ourselves online; for example when paying online. It hides the malicious activity from us, and our security software, and uses legitimate Windows tools to stay under the radar. What do CAPTCHA threats look like? There are various ways you may be exposed to a malicious CAPTCHA. It may be that you are tricked into clicking on a malicious link in a phishing email, text or social media message. Thanks to AI, this threat is growing. Generative AI is helping threat actors to scale social engineering attacks while improving the quality of the language to near-perfect, in multiple languages at once. Alternatively, you may visit a random, legitimate site that hackers have injected malicious ads or other content into. These are particularly dangerous as no user interaction is required for the download. And you may not realize that anything untoward has happened until it’s too late. When the CAPTCHA box pops up it will look legitimate enough. But what it asks you to do should set alarm bells ringing. Instead of the usual CAPTCHA task, like identifying similar images or typing text that has been obfuscated in some way, it will ask you to perform specific commands like: Clicking through to verify you are a human Pressing Windows key + R to open “Run” Pressing CTRL + V to paste a command secretly copied to the clipboard by the malware Pressing ENTER to execute the above command This command often triggers legitimate Windows tools like PowerShell or mshta.exe to download additional malicious payloads from an external server. The end goal is usually to install infostealer malware on your device. Infostealers are designed to scour the computer or mobile phone for logins, photos, contacts, and other sensitive data to sell on the dark web and/or use to commit identity fraud. They target browsers, email clients, crypto wallets, apps and the operating systems itself to do so – taking screenshots, keylogging and harvesting data in other ways. According to one study, there were at least 23 million infostealer victims in 2024, most of which were Windows systems. They managed to steal over two billion victim credentials. One of the most popular strains of infostealing malware, Lumma Stealer, compromised as many as 10 million devices before an international effort, which involved also ESET, disrupted this prolific malware-as-a-service (MaaS) threat. A CAPTCHA threat could also install a remote access trojan (RAT), another type of malware but this time designed to provide remote access to your machine. According to one study, AsyncRAT was seen in 4% of incidents during 2024. This RAT has been operational since 2019 and carries out activities including like data theft and keylogging. Staying safe from CAPTCHA threats To stay safe from infostealers, RATs and other nasties, consider the following: Stay alert to unusual CAPTCHA requests like the ones listed above. Be cautious of any CAPTCHA challenge that seems to pop up out of nowhere. Keep your OS and browser software up to date, to minimize the risk of malware exploiting vulnerabilities in older versions. Install security software from a reputable vendor and keep it updated. This will go a long way toward blocking any malware or suspicious activity. Don’t download pirated software, as this could contain malware of the sort that delivers fake CAPTCHAs. Consider using an ad blocker, which would stop you from being exposed to any content delivered via a malicious online ad. What happens if you fall for a fake CAPTCHA If the worst happens and you unwittingly execute the hidden commands described above: Run a malware scan to find and hopefully purge the machine of any malware that may have been covertly downloaded. Disconnect from the internet and backup any important photos and/or files. Perform a factory reset on your computer or device. Change all your passwords, using strong, unique credentials stored in a password manager. Switch on multi factor authentication (MFA) for all accounts, so that even if a hacker has stolen your passwords they can’t access your accounts. Falling for a CAPTCHA threat isn’t the end of the world. But it pays to act fast if you find yourself in a worst-case scenario. Be safe out there.
welivesecurity.comJul 24, 2025source
Unmasking AsyncRAT: Navigating the labyrinth of forks
AsyncRAT has cemented its place as a cornerstone of modern malware and as a pervasive threat that has evolved into a sprawling network of forks and variants. While its capabilities are not that impressive on their own, it is the open-source nature of AsyncRAT that has truly amplified its impact. This blogpost provides an overview and analysis of the most relevant forks of AsyncRAT, drawing connections between them and showing how they have evolved. Key points of this blogpost: We provide unique insights into the landscape of AsyncRAT and its numerous variants in order to navigate the labyrinth of forks easily. In the effort to map the vast hierarchy of AsyncRAT’s forks, we uncover their unique interconnections and document how these variants can be distinguished. We explore less common variants that feature unique plugins, ranging from a basic screamer plugin to a USB malware spreader. Origins of AsyncRAT You may have heard of AsyncRAT, short for asynchronous remote access trojan. This open-source RAT was released on GitHub in 2019 by a user going by the name of NYAN CAT. Developed in C#, it offers a wide range of typical RAT functionalities, including keylogging, screen capturing, credential theft, and more. Its simplicity and open-source nature has made it a popular choice among cybercriminals, leading to its widespread use in various cyberattacks. But where does it come from? We believe that the groundwork for AsyncRAT was laid earlier by the Quasar RAT, which has been available on GitHub since 2015 and features a similar approach. Both are written in C#; however, their codebases differ fundamentally, suggesting that AsyncRAT was not just a mere fork of Quasar, but a complete rewrite. A fork, in this context, is a personal copy of someone else’s repository that one can freely modify without affecting the original project. The main link that ties them together lies in the custom cryptography classes used to decrypt the malware configuration settings. Specifically, these are classes Aes256 and Sha256, which fall under the Client.Algorithm namespace for AsyncRAT and the Quasar.Common.Cryptography namespace for Quasar. Figure 1 shows identical code being used in both implementations of Aes256. The same code is mostly copied and pasted, including the same salt value and decryption settings. This class, together with Sha256, leads us to believe that AsyncRAT was to some degree influenced by the Quasar RAT. Apart from these similarities, AsyncRAT introduced significant improvements, particularly in its modular architecture and enhanced stealth features, which make it more adaptable and harder to detect in modern threat environments. Its plugin-based architecture and ease of modification have sparked the proliferation of many forks, pushing the boundaries even further. Fork labyrinth Ever since it was released to the public, AsyncRAT has spawned a multitude of new forks that have built upon its foundation. Some of these new versions have expanded on the original framework, incorporating additional features and enhancements, while others are essentially the same version in different clothes. Fork hierarchy Figure 2 illustrates how some of the more prevalent AsyncRAT forks have evolved from one another over time. In the middle of the tree are DcRat and VenomRAT. Our analysis has shown that they are the most widely deployed variants, together accounting for a significant number of campaigns. Other lesser-known forks occupy smaller but nonetheless notable portions of the pie. Figure 3 depicts the distribution of the most prevalent forks according to our telemetry. DcRat offers a notable improvement over AsyncRAT in terms of features and capabilities. One of the more obvious changes is the data structure used for transferring data back and forth. It utilizes MessagePack, a well-known open-source library for more efficient binary data serialization. DcRat also implements evasion techniques like AMSI and ETW patching, which work by disabling security features that detect and log malicious behavior – AMSI patching prevents script scanning, while ETW patching blocks event tracing. Additionally, it features an antiprocess system whereby processes whose names match those in a denylist are terminated. Blocklisted programs include Taskmgr.exe, ProcessHacker.exe, MsMpEng.exe, Taskkill.exe, etc. It’s also worth noting that DcRat’s plugin base builds upon AsyncRAT and further extends its functionality. Among the added plugins are capabilities such as webcam access, microphone recording, Discord token theft, and “fun stuff”, a collection of plugins used for joke purposes like opening and closing the CD tray, blocking keyboard and mouse input, moving the mouse, turning off the monitor, etc. Notably, DcRat also introduces a simple ransomware plugin that uses the AES-256 cipher to encrypt files, with the decryption key distributed only once the plugin has been requested. Apart from that, there appear to be many small changes like a different choice of salt (a string instead of a binary value), deliberately changed variable names to further evade detection, dynamic API resolution, and many more. VenomRAT, on the other hand, was likely inspired by DcRat, as evidenced in the Identifying versions section. The malware is packed with so many features that it could be considered a separate threat on its own. We have chosen to group it under AsyncRAT as their client parts are very similar to each other. VenomRAT’s features and plugins have been documented in more detail by other vendors, so we won’t dive deep into them in this blogpost. Not all RATs are serious in nature though, and this applies equally to AsyncRAT forks. Clones like SantaRAT or BoratRAT (see Figure 4) are meant to be jokes. In the case of the former, its authors have themselves acknowledged that the project was basically “shamelessly ripped off of DcRat”. Yet, despite this, we have found instances of real-world usage of them in the wild. While doing the analysis, we used various methods to identify and categorize each sample. It should be noted that the research was primarily on the client part of the malware, as this binary is what ends up on victims’ machines. It contains useful information such as malware configuration and where information about the C&C can be found. The quickest and most straightforward way to identify a fork is to peek directly into the malware’s configuration, which can usually be found in the InitializeSettings function. The configuration values are encrypted with AES-256 and stored as base64 strings in the Settings class. In most cases, the correct fork name is readily available and conveniently labeled as Version. In about 90% of our analyzed samples, the Version field contains some meaningful description of either the fork’s name or the malware author’s pseudonym. The remaining samples had this field intentionally left blank. Figure 5 illustrates the typical configuration initialization procedure found in DcRat and its derivatives (VenomRAT in this case). If the Version field is empty, sometimes it’s possible to get another clue by looking at the Salt value used for encrypting the configuration. Attackers often neglect this parameter when copy-pasting their own fork. The Salt value can be found in the Client.Algorithm.Aes256 class, as seen in Figure 6. Yet another way to get more insight is to look for the embedded certificate used to authenticate the C&C server. It’s also located in the configuration as a base64-encoded value. Unpacking this value often reveals further information about the server, such as common name, organization, and organizational unit. If a particular fork has its own name in the Version field, it is often possible to trace back the previous fork upon which it was likely based by looking at the CN field. Figure 7 shows a DER-encoded certificate that reveals the BoratRAT fork, after extraction and decoding. The techniques mentioned above primarily apply to trivial cases where malware authors either did not bother to remove traces or used a default certificate. A more sophisticated method for identifying AsyncRAT servers exists, which involves sending a specially crafted packet to the C&C server. This approach is explained in detail in this Axel Mahr blogpost. Should everything else fail, determining the sample origin can ultimately be done the old-fashioned way, by manually inspecting the code. This involves a detailed analysis of the code’s structure, syntax, and functionality, comparing them against the patterns of previously categorized samples. Extensive fork list We have highlighted here some of the more prominent AsyncRAT forks. Due to the sheer number of available forks, it is not feasible to cover every single one. For completeness, Figure 8 provides an extended list of AsyncRAT forks known to be used for malicious purposes, as seen in ESET telemetry to date. Exploring lesser-known variants So far, we’ve mentioned some of the major forks that dominate the landscape. In this section, we have cherry-picked some lesser-known forks that enhance AsyncRAT’s functionality beyond the features included in the default versions. These exotic forks are often the work of one person or group and they make up less than 1% of the volume of AsyncRAT samples. NonEuclid RAT This fork stands out primarily for its inclusion of new plugins, on top of the default ones. While some plugins might seem trivial or geared towards “fun stuff”, others, like WormUsb.dll, have distinctly malicious purposes. Table 1 lists a selection of NonEuclid RAT plugins that deviate from the standard plugin base seen in regular forks. Table 1. Selection of NonEuclid RAT plugins we deemed interesting Screamer.dll There are five jump scare images built into the plugin. An attacker sends a command that indicates which image they want to use, along with the WAV file to be played, and the delay after which the jump scare is triggered. Figure 9 shows the first three prebundled images an attacker may choose from. Piano.dll This plugin plays arbitrary WAV files. All sound files are stored in %appdata%\Piano. piano.dll supports three commands: SetSound – adds a new sound file to %appdata%\Piano, PlayMisc – plays a requested sound file from %appdata%\Piano, and ClientAdd – retrieves multiple sound files from the C&C server. Service.dll This plugin facilitates managing Windows services, such as starting, stopping, and pausing services. Maps.dll This is a simple plugin to collect geolocation information from the victim. It uses the .NET GeoCoordinateWatcher class to register a callback function to collect data each time the location is available. Among the collected information are latitude, longitude, username, and computer name. WormUsb.dll This plugin compromises PE files with an arbitrary payload specified by the attacker. Despite the term Usb in WormUsb.dll, this plugin targets several locations based on the command provided: InfectExe – compromises an individual PE file, InfectExeInWindows – targets PE files in personal folders (Desktop, Documents, Downloads, My Music), and InfectUsbExe – targets PE files in all drives excluding the C drive. Under the hood, it works by moving the original file to a temporary location. Then it drops a small stub in place of the original file. This stub’s resource section is then populated to contain both the original file and the specified payload, both of which are compressed and encrypted with a per-file key, generated at the time of construction. Following this, the malware then obfuscates the stub by introducing proxy methods, adding custom control flow obfuscation, and variable renaming. As a final touch, it embeds the original icon and metadata in the modified stub. Figure 10 shows the function, with the original method names, responsible for compromising a single file. When such a compromised file is executed, it first decrypts, unpacks and runs the payload program, then proceeds to do the same with the original program. Brute.dll This plugin supports brute forcing of both SSH and FTP protocols from the client side. The attacker feeds it three parameters: host, login, and password, and the plugin will try to connect using those credentials. If the connection succeeds, the credentials are sent back to the attacker with a flag indicating success. It’s not difficult to imagine a scenario whereby an attacker might use this sort of attack to distribute brute forcing across a large pool of compromised machines, thus circumventing restrictions based solely on the IP address. Signature Antivirus.dll The name of the plugin implies it might have something to do with antivirus functionality. While this is technically true, it is also a case of the most primitive, manual antivirus solution ever created. The plugin receives a list of MD5 hashes from the attacker and compares them to the hashes of all EXE files it finds on every disk. If a matching file is found, it triggers the oddly named DetectVirus function, which merely deletes the file without any further analysis. This makes the name of the plugin very dubious at best. In the hands of the malware author, it may have been used to delete competitor malware, or really just any arbitrary file. cliper.dll This is a standalone clipper that continuously monitors the victim’s clipboard, and if a cryptocurrency wallet address is detected, it is replaced with one provided by the attacker. Attacker-provided wallets are only sent when the plugin is first requested; they are not hardcoded in the plugin. Additionally, in Figure 11, we can also see some credit card entries. This plugin contains an extensive list of regexes that can detect both cryptocurrency wallets and credit cards, and in the case of the latter they just get sent back to the attacker. JasonRAT Identified in 2024, this variant shows continued signs of activity. It is interesting in that it employs obscure variable-naming conventions reminiscent of “satanic” terms from what the malware author refers to as the Book of Jason. In Figure 12, you can see typical AsyncRAT configuration values (in base64), but with renamed variables, while Figure 13 shows the logic of the main entry point of the malware. Besides the usual configuration values, this variant further extends the client by introducing country targeting. Another strange feature is the choice of string obfuscation. A subset of the strings employs an extra layer of obfuscation by utilizing an extended variant of Morse code. Both uppercase and lowercase letters are included, as well as some special characters. Figure 14 shows the encoded registry key string using an extended mapping. XieBroRAT This is a RAT with Chinese localization. It introduces a new plugin, BrowserGhost.dll, which is a browser-credential stealer. Another plugin, Abstain.dll, provides interaction with Cobalt Strike servers by making a reverse connection. To increase the coverage, the malware provides the delivery chain in several different languages. The standard .NET client binary can be wrapped and distributed via shellcode, VBS, or JavaScript. Finally, the author further extended the malware by borrowing heavily from open-source projects, integrating tools like mimikatz, SharpWifiGrabber, SharpUnhooker, etc. Conclusion AsyncRAT’s rise and its subsequent forks highlight the inherent risks of open-source malware frameworks. Our analysis revealed a diverse and evolving ecosystem of derivatives, ranging from persistent threats like DcRat and VenomRAT to lesser-known novelty forks like JasonRAT and BoratRAT, which seem to serve more as curiosities than credible threats. All of these forks not only extend AsyncRAT’s technical capabilities but also demonstrate how quickly and creatively threat actors can adapt and repurpose open-source code. The widespread availability of such frameworks significantly lowers the barrier to entry for aspiring cybercriminals, enabling even novices to deploy sophisticated malware with minimal effort. This democratization of malware development – especially considering the rising popularity of LLMs and potential to misuse their capabilities – further accelerates the creation and customization of malicious tools, contributing to a rapidly expanding and increasingly complex threat landscape. In light of these trends, it is reasonable to anticipate that future forks may incorporate more advanced obfuscation, modularity, and evasion capabilities. This potential evolution underscores the importance of proactive detection strategies and deeper behavioral analysis to effectively address emerging threats. IoCs A comprehensive list of indicators of compromise (IoCs) can be found in our GitHub repository. Files MITRE ATT&CK techniques This table was built using version 17 of the MITRE ATT&CK framework.
welivesecurity.comJul 15, 2025source
Task scams: Why you should never pay to get paid
Many of us have been experiencing a cost-of-living crisis for years, and the news headlines remain filled with doom-laden predictions of what the future might hold. Against this backdrop, it’s understandable why many of us are looking for a side hustle or for even a new, better-paid job. But the scammers know this, and are ready to take advantage. In 2024 alone, employment scams reported to the FBI made fraudsters over $264 million. Many of these are so-called “task scams,” where victims are actually tricked into paying a “deposit” in order to get paid. It might sound unbelievable. But it’s easier to fall for than you think. Here’s what you need to know to make sure your dream job doesn’t turn into a nightmare. How big is the problem? Employment or recruitment scams in general are the result of various factors: the normalization of more “anonymous” types of remote work pure reliance on digital recruitment platforms and tech convincing AI-generated scam sites and schemes, lowering the barrier to entry for criminals the pseudonymous nature of cryptocurrency a large volume of breached data, which can be used to personalize scams a lack of awareness on the part of the victims, coupled with poor economic conditions One thing is certain, the scale of the problem is certainly under-reported. Even the 20,000 cases reported to the FBI last year are likely just to be the tip of the iceberg. What are task scams and how do they work? Task scams are a specific type of employment fraud that use gamification techniques to string victims along. They’re on the rise. The FTC recorded 20,000 cases in the first half of 2024 alone, versus 5,000 in the whole of 2023. They may differ in small details, but all work something like this: You are approached via an unsolicited text message (WhatsApp, Telegram etc), SMS or social media message. The scammer promises easy money performing straightforward online tasks in return for a handsome renumeration. According to the FTC, they may mention “product boosting” or “app optimization tasks,” but in general remain hazy about the details. If you sign up, you’ll be asked to perform mundane tasks such as “liking” online content or clicking various combinations of buttons on a fake site/app. The site/app itself may even be a spoofed version of a legitimate website like Temu. You will be able to view your ‘earnings’ as you go. Although these earnings will initially increase, there will be a point at which the app or your scammer requires you to “charge up” or “level up” your account, or “unlock your earnings” by paying into the fund, usually via cryptocurrency. This is what the whole scam is really about. Once you pay, your money disappears and you’ll have no way to recoup those fake earnings. Reports suggest that, to persuade you to hand crypto over to the scammers, they may invite you to a group chat to hear from peers who have successfully made money this way. In fact, these people are also fraudsters. The bigger picture: other employment scams Task scams are not the only trick employment fraudsters can play on you. Look out for the following ways scammers try to get hold of your personal information and/or money: Fake job ads (sometimes on legitimate recruitment sites) which are designed to lure you into applying. They want your personal info and may also ask for a fee for made-up reasons like admin, training, police checks etc Unsolicited messages reaching you in the same way as the task scam above, but with attractive job opportunities from someone posing as a recruitment agent. They will also try to harvest personal info and money If you post your CV on a recruitment site, the scammers could also find and reply to you that way. Once again, they want your money and personal information Keeping the scammers at bay Fortunately, it’s not hard to keep these fraudsters at arm’s length, if you bear in mind a few things. Remember the warning signs from the examples above, and: Never reply to unsolicited job offers. Legitimate recruiters will reach out via proper channels If you are approached after posting a CV to a job site, research the company/agent and search for any online mentions they may have in connection with scams Remember that if a job offer is too good to be true, it probably is If there’s no formal interview process, that should also raise some red flags Never pay any money upfront for a prospective job. Legitimate recruiters will never ask you to do so Never share sensitive personal or financial information with ‘recruiters’ that reach out to you on unofficial channels. They may want it to sell to fraudsters, or use themselves for identity fraud Falling for a task scam might seem far-fetched, but the truth is that these con tricks usually play out over a period of days, and the gamification element can really suck you in. It’s not uncommon for some victims to end up losing a fortune. Make sure you’re not one of them.
welivesecurity.comJul 4, 2025source
How government cyber cuts will affect you and your business
We often hear about cybersecurity fatigue, the mental and emotional strain that weighs on individuals and teams on the frontlines and leads to decreased productivity, burnout and, ultimately, increases the risk of a successful cyberattack. Add staffing and funding cuts to the mix, and the problem is only likely to get worse. In fact, the impacts will be felt not only by those directly involved, but will extend to cybersecurity vendors and service providers, who will have to either innovate their products or adapt their service offerings to changing market dynamics. The shifting ground Recent cuts in federal budgets and workforce reductions in key organizations such as the Cybersecurity and Infrastructure Security Agency (CISA) will undoubtedly weaken the cybersecurity posture of not only the federal government, but of all businesses and institutions – whether they utilize CISA’s threat intelligence and notifications or are reliant on best practice guidance through cybersecurity frameworks. Beyond the agencies that are directly funded by the U.S. government, there are many companies that provide specialized cybersecurity services and technology to both federal and state-level entities. Governments are among the largest consumers of cybersecurity services, and private companies are often reliant on the revenue from these contracts. Thus, any reduction in contracts may lead to reductions in headcount and in investment in research and development. At the same time, it may further accelerate demand for automated features and AI support – perhaps even beyond what is currently proven efficacious. If this all seems far away from your day-to-day, real world, then you may need to think again. Consider the direct impact of initiatives like the U.S. State and Local Cybersecurity Grant Program, which provided almost $700 million in funding between 2023 and 2024, giving a much-needed boost for states that needed to refresh and improve their cybersecurity posture. Many states used the funding to centralize some elements of their cybersecurity, allowing all state-funded entities to benefit from volume licensing of modern, advanced cybersecurity technologies. For example, if your local school district or regional government benefited from these types of funding programs, any change in future funding could put you and your family at risk should there be a cyber-incident. Stifling innovation, straining talent Some vendors latched onto these federally-funded initiatives and grabbed market share, dominating the opportunity. It’s vendors such as these that are likely to fall victim to funding cuts, either through reduced service contracts or future grant funding. This market dominance also led to single-vendor monoculture issues (you can read more about my concerns on this in this article). As affected cybersecurity vendors take stock of the situation, they will implement their own reductions in headcount, which some have already done, and will make cuts to their R&D budgets. This directly affects the innovation of future technologies, which, in turn, may reduce cybersecurity defense effectiveness. There is an upside – or is there? As companies reduce headcounts, the talent shortage in cybersecurity teams should be alleviated to some degree as additional talent becomes available. At the same time, those left in smaller, leaner teams will likely suffer increased cybersecurity fatigue to the point where they may decide to leave the industry and look for less stressful opportunities. And if the market has more talent to choose from, then salaries being offered could plateau, maybe even decrease, making the industry less attractive to new talent and those considering a career in cybersecurity. Lower investment may also see education establishments removing or reducing the opportunity for students to participate in courses, further shrinking the pool of future talent. Filling the void There may be a silver lining. Federal cuts to CISA could create new opportunities for Managed Service Providers (MSPs) and cybersecurity vendors offering Managed Detection and Response (MDR) services. With reduced federal funding, organizations may seek alternative solutions from operational budgets to maintain their cybersecurity posture, turning to private-sector providers for their expertise and resources. This shift could lead to increased demand for MSPs and MDR services, as businesses look for cost-effective and reliable ways to protect themselves. The reduction in funding may also be felt in other ways; for example, in the evolution of standards and dissemination of intelligence and awareness that is often gained from public-private collaborations. Even critical resources like the MITRE CVE database hosting recently faced a funding challenge, and while the issue did get resolved, at least for now, it served as a stark reminder of how quickly even foundational elements can be threatened. Agencies such as the National Institute of Standards and Technology (NIST), who are responsible for the development of cybersecurity frameworks that are the backbone of many companies’ cybersecurity policies, may struggle to develop new frameworks and delay crucial updates to existing ones. These are examples of how funding issues may materialize; however, in reality, the impact is likely to be felt across all agencies, institutions, businesses and even by consumers who become the victims of breaches that could have been avoided. The true impact of a reduction in federal funding that affects the cybersecurity sector will not be immediately apparent; the underinvestment it causes could take years to materialize. Slowing innovation and the adoption of new technologies will play out over time and the issue caused will be on someone else’s watch. The bottom line One thing is for certain, though: there will be no slowing down the development of the sophisticated techniques being used by cybercriminals. A funding reduction in cybersecurity hands cybercriminals a significant opportunity, ensuring their activities will reap long-term rewards and maintain stability in their revenue stream.
welivesecurity.comJul 3, 2025source
Gamaredon in 2024: Cranking out spearphishing campaigns against Ukraine with an evolved toolset
Since Russia’s full-scale invasion of Ukraine in February 2022, cyberespionage has played a crucial role in the broader threatscape. Russia-aligned advanced persistent threat (APT) groups have relentlessly targeted Ukrainian entities, employing cyberattacks alongside disinformation campaigns. ESET Research has closely monitored these activities, regularly documenting cyber-operations carried out by various threat actors, including the highly active Gamaredon group. Key points of this blogpost: Gamaredon refocused exclusively on targeting Ukrainian governmental institutions in 2024, abandoning prior attempts against NATO countries. The group significantly increased the scale and frequency of spearphishing campaigns, employing new delivery methods such as malicious hyperlinks and LNK files executing PowerShell from Cloudflare-hosted domains. Gamaredon introduced six new malware tools, leveraging PowerShell and VBScript, designed primarily for stealth, persistence, and lateral movement. Existing tools received major upgrades, including enhanced obfuscation, improved stealth tactics, and sophisticated methods for lateral movement and data exfiltration. Gamaredon operators managed to hide almost their entire C&C infrastructure behind Cloudflare tunnels. Gamaredon increasingly relied on third-party services (Telegram, Telegraph, Cloudflare, Dropbox) and DNS-over-HTTPS (DoH) for protecting its C&C infrastructure. In our previous blogpost, Cyberespionage the Gamaredon way: Analysis of toolset used to spy on Ukraine in 2022 and 2023, we described Gamaredon’s aggressive cyberespionage activities against Ukrainian governmental institutions. As part of our continued investigation, we’ve conducted a thorough technical analysis of Gamaredon’s operations throughout 2024. The detailed results and technical insights are available in our latest white paper. Our research shows that the group remains highly active, consistently targeting Ukraine, but has notably adapted its tactics and tools. Targeting Ukraine exclusively Gamaredon, attributed by the Security Service of Ukraine (SSU) to the 18th Center of Information Security of Russia’s Federal Security Service (FSB), has targeted Ukrainian governmental institutions since at least 2013. While previous years saw occasional attempts against targets in other NATO countries, during 2024 Gamaredon operators returned their focus exclusively to Ukrainian institutions. This strongly aligns with the group’s historical objective as a cyberespionage actor aligned with Russian geopolitical interests. Given the ongoing war and geopolitical tensions, we expect Gamaredon’s targeting of Ukraine to continue unchanged in the foreseeable future. Spearphishing campaigns grow larger and more frequent Gamaredon’s spearphishing activities significantly intensified during the second half of 2024. Campaigns typically lasted one to five consecutive days, with emails containing malicious archives (RAR, ZIP, 7z) or XHTML files employing HTML smuggling techniques. These files delivered malicious HTA or LNK files that executed embedded VBScript downloaders such as PteroSand. Figure 1 depicts the number of unique samples of those HTA and LNK files delivered per month in Gamaredon spearphishing campaigns in 2024. Surprisingly, in October 2024, we observed a rare case where spearphishing emails included malicious hyperlinks rather than attachments – a deviation from Gamaredon’s usual tactics. Furthermore, Gamaredon introduced another novel technique: using malicious LNK files to execute PowerShell commands directly from Cloudflare-generated domains, bypassing some traditional detection mechanisms. Toolset evolution: New tools and significant improvements Gamaredon’s toolset underwent notable updates. While fewer new tools were introduced (six compared to eight in 2022 and nine in 2023), substantial resources went into updating and improving existing tools: New tools introduced in 2024 include: PteroDespair: A short-lived PowerShell reconnaissance tool discovered in January 2024, developed to collect diagnostic data on previously deployed malware. PteroTickle: A PowerShell weaponizer discovered in March 2024, targeting Python applications converted into executables on fixed and removable drives, facilitating lateral movement. It weaponizes Tcl scripts typically found in Python GUI apps using Tkinter and built with PyInstaller. PteroGraphin: Discovered in August 2024, this PowerShell tool initially used an uncommon persistence method involving Microsoft Excel add-ins. It creates an encrypted communication channel for payload delivery, through the Telegraph API. Later versions simplified persistence by using scheduled tasks instead. PteroStew: A new general-purpose VBScript downloader discovered in October 2024, similar to previously known downloaders (e.g., PteroSand, PteroRisk), but that notably stores its code in alternate data streams associated with benign files on the victim’s system. PteroQuark: Another VBScript downloader discovered in October 2024, introduced as a new component within the VBScript version of the PteroLNK weaponizer. PteroBox: A PowerShell file stealer discovered in November 2024, closely resembling PteroPSDoor but exfiltrating stolen files to Dropbox. It leverages WMI event subscriptions to detect newly inserted USB drives and exfiltrates selected files using the Dropbox API. The stolen files are meticulously tracked to avoid redundant uploads, highlighting Gamaredon’s increasing attention to stealth and efficiency. Major updates to existing tools in 2024 In addition to new tools, Gamaredon operators significantly upgraded existing tools in their arsenal: PteroPSDoor: A major upgrade introduced advanced stealth techniques, such as monitoring file changes via the IO.FileSystemWatcher object rather than continuously scanning directories, significantly reducing visibility. It also implemented WMI event subscriptions to detect new USB insertions, making file exfiltration more targeted and stealthier. Additionally, the latest versions store code exclusively in registry keys instead of in files, further complicating detection. PteroLNK (VBScript version): This tool was enhanced in early 2024 to weaponize not only USB drives but also mapped network drives, expanding its lateral movement capabilities. Throughout the second half of 2024, it received multiple incremental updates, including improved obfuscation, more complex methods for LNK file creation, and registry-based techniques to hide files and file extensions from victims. This weaponizer has become one of Gamaredon’s most frequently updated and actively maintained tools. PteroVDoor: This VBScript file stealer continued to be used in two variants (obfuscated and unobfuscated). Throughout 2024, Gamaredon operators repeatedly updated the tool, introducing new external platforms such as Codeberg repositories to dynamically distribute command and control (C&C) server information, complicating defensive measures. PteroPSLoad: Gamaredon notably transitioned back to using Cloudflare tunnels instead of ngrok for its C&C infrastructure. This marked the beginning of Gamaredon hiding almost its entire C&C infrastructure behind Cloudflare-generated domains, significantly enhancing its operational security. Unusual payloads: Russian propaganda via malware? A particularly intriguing finding was the discovery in July 2024 of a unique ad hoc VBScript payload, delivered by Gamaredon downloaders. This payload had no espionage functionality; rather, its sole purpose was to automatically open a Telegram propaganda channel named Guardians of Odessa, which spreads pro-Russian messaging targeting the Odessa region. While unusual for Gamaredon’s typical operations, we attribute this payload to Gamaredon with high confidence. Network infrastructure and evasion techniques Throughout 2024, Gamaredon showed persistent dedication to evading network-based defenses: The group continued, albeit at a reduced scale, to leverage fast-flux DNS techniques, frequently rotating IP addresses behind its domains. However, the number of domains that it registered declined notably from over 500 in 2023 to about 200 in 2024. Gamaredon increasingly relied on third-party services such as Telegram, Telegraph, Codeberg, and Cloudflare tunnels to obfuscate and dynamically distribute its C&C infrastructure. Cloudflare-generated subdomains became the group’s primary communication endpoints, with traditional domains relegated mostly to fallback use. Multiple DoH services (Google and Cloudflare) and third-party resolver websites (such as nslookup.io, who.is, dnswatch.info, and check-host.net) were regularly leveraged to bypass domain-based blocking. Gamaredon also introduced new techniques such as dropping embedded HTA and VBScript files into temporary directories and executing them separately to resolve C&C domains, further complicating automated detection efforts. Despite observable capacity limitations and abandoning older tools, Gamaredon remains a significant threat actor due to its continuous innovation, aggressive spearphishing campaigns, and persistent efforts to evade detections. As long as the Russia's war against Ukraine continues, we anticipate Gamaredon will persistently evolve its tactics and intensify its cyberespionage operations against Ukrainian institutions. For a detailed technical breakdown of Gamaredon’s 2024 activities, updates, and malware analyses, read our full white paper. A comprehensive list of indicators of compromise (IoCs) can be found in our GitHub repository and the Gamaredon white paper.
welivesecurity.comJul 2, 2025source
ESET APT Activity Report Q4 2024–Q1 2025: Malware sharing, wipers and exploits
In the latest episode of the ESET Research Podcast, ESET Distinguished Researcher Aryeh Goretsky is joined by ESET Security Awareness Specialist Rene Holt to dissect the key findings from ESET’s APT Activity Report. The first actor that steps into the limelight is UnsolicitedBooker, a China-aligned APT group that has demonstrated a level of persistence that truly puts the "P" in APT. This group targeted the same organization three times over several years, attempting to deploy its signature backdoor, MarsSnake. This example highlights the relentless focus of certain groups that will stop at nothing to achieve their objectives. The conversation then shifts to the challenges of attribution, particularly with the increasing trend of tool-sharing, primarily among China-aligned actors such as Worok. Their tactic is to muddy the waters by using overlapping toolsets sourced from digital quartermasters, intertwining their activities with those of other groups such as LuckyMouse and TA428. Turning to Russia-aligned actors, the discussion focuses on groups like Sednit, Gamaredon, and Sandworm. Sednit’s latest activity revolves around Operation RoundPress, which originally targeted the popular webmail service Roundcube but has recently expanded to other platforms such as Horde, MDaemon, and Zimbra. Sednit has been using targeted emails, exploiting flaws in these services, and employing cross-site scripting to attack defense companies located in Bulgaria and Ukraine. Gamaredon remains one of the most active APTs in Ukraine, constantly tweaking its obfuscation techniques to stay ahead of detection. Meanwhile, Sandworm has intensified its use of data-wiping malware, deploying a new wiper called ZEROLOT multiple times in the past six months. This wiper operates with surgical precision, erasing specific files and directories without immediately taking down the entire system—an approach that ensures the malware can complete its destructive mission. Aryeh and Rene also delve into the activities of North Korea-aligned and Iran-aligned groups. If you’re interested in more details, be sure to listen to this episode of the ESET Research Podcast or download the latest ESET APT Activity Report. Discussed topics: UnsolicitedBooker (MarsSnake) 1:45 Worok (and its digital quartermasters) 4:50 Sednit (Operation RoundCube) 9:55 Gamaredon 13:55 Sandworm (ZEROLOT wiper) 16:15 DeceptiveDevelopment (WeaselStore, ClickFix) 24:10 MuddyWater vs Lyceum 29:40
welivesecurity.comJul 1, 2025source