February 29, 2024

Protecting Against AlphV BlackCat Ransomware

Learn how Darktrace AI is combating AlphV BlackCat ransomware, including the details of this ransomware and how to protect yourself from it.

Written by

Sam Lister

Specialist Security Researcher

Inside the SOC

Darktrace cyber analysts are world-class experts in threat intelligence, threat hunting and incident response, and provide 24/7 SOC support to thousands of Darktrace customers around the globe. Inside the SOC is exclusively authored by these experts, providing analysis of cyber incidents and threat trends, based on real-world experience in the field.

Written by

Sam Lister

Specialist Security Researcher

Feb 2024

As-a-Service malware trending

Throughout the course of 2023, “as-a-Service” strains of malware remained the most consistently observed threat type to affect Darktrace customers, mirroring their overall prominence across the cyber threat landscape. With this trend expected to continue throughout 2024, organizations and their security teams should be prepared to defend their network against increasingly versatile and tailorable malware-as-a-service (MaaS) and ransomware-as-a-service (RaaS) strains [1].

What is ALPHV ransomware?

The ALPHV ransomware, also known as ‘BlackCat’ or ‘Noberus’, is one example of a RaaS strain that has been prominent across the threat landscape over the last few years.

ALPHV is a ransomware strain coded in the Rust programming language. The ransomware is sold as part of the RaaS economy [2], with samples of the ransomware being provided and sold by a criminal group (the RaaS ‘operator’) to other cybercriminals (the RaaS ‘affiliates’) who then gain entry to organizations' networks with the intention of detonating the ransomware and demanding ransom payments.

ALPHV was likely first used in the wild back in November 2021 [3]. Since then, it has become one of the most prolific ransomware strains, with the Federal Bureau of Investigation (FBI) reporting nearly USD 300 million in ALPHV ransom payments as of September 2023 [4].

In December 2023, the FBI and the US Department of Justice announced a successful disruption campaign against the ALPHV group, which included a takedown of the their data leak site, and the release of a decryption tool for the ransomware strain [5], and in February 2024, the US Department of State announced a reward of up to USD 10 million for information leading to the identification or location of anyone occupying a key leadership position in the group operating the ALPHV ransomware strain [6].

The disruption campaign against the ransomware group appeared to have been successful, as evidenced by the recent, significant decline in ALPHV attacks, however, it would not be surprising for the group to simply return with new branding, in a similar vein to its apparent predecessors, DarkSide and BlackMatter [7].

How does ALPHV ransomware work?

ALPHV affiliates have been known to employ a variety of methods to progress towards their objective of detonating ALPHV ransomware [4]. In the latter half of 2023, ALPHV affiliates were observed using malicious advertising (i.e, malvertising) to deliver a Python-based backdoor-dropper known as 'Nitrogen' to users' devices [8][12]. These malvertising operations consisted in affiliates setting up malicious search engine adverts for tools such as WinSCP and AnyDesk.

Users' interactions with these adverts led them to sites resembling legitimate software distribution sites. Users' attempts to download software from these spoofed sites resulted in the delivery of a backdoor-dropping malware sample dubbed 'Nitrogen' to their devices. Nitrogen has been observed dropping a variety of command-and-control (C2) implants onto users' devices, including Cobalt Strike Beacon and Sliver C2. ALPHV affiliates often used the backdoor access afforded to them by these C2 implants to conduct reconnaissance and move laterally, in preparation for detonating ALPHV ransomware payloads.

Darktrace Detection of ALPHV Ransomware

During October 2023, Darktrace observed several cases of ALPHV affiliates attempting to infiltrate organizations' networks via the use of malvertising to socially engineer users into downloading and installing Nitrogen from impersonation websites such as 'wireshhark[.]com' and wìnscp[.]net (i.e, xn--wnscp-tsa[.]net).

While the attackers managed to bypass traditional security measures and evade detection by using a device from the customer’s IT team to perform its malicious activity, Darktrace DETECT™ swiftly identified the subtle indicators of compromise (IoCs) in the first instance. This swift detection of ALPHV, along with Cyber AI Analyst™ autonomously investigating the wide array of post-compromise activity, provided the customer with full visibility over the attack enabling them to promptly initiate their remediation and recovery efforts.

Unfortunately, in this incident, Darktrace RESPOND™ was not fully deployed within their environment, hindering its ability to autonomously counter emerging threats. Had RESPOND been fully operational here, it would have effectively contained the attack in its early stages, avoiding the eventual detonation of the ALPHV ransomware.

Figure 1: Timeline of the ALPHV ransomware attack.

In mid-October, a member of the IT team at a US-based Darktrace customer attempted to install the network traffic analysis software, Wireshark, onto their desktop. Due to the customer’s configuration, Darktrace's visibility over this device was limited to its internal traffic, despite this it was still able to identify and alert for a string of suspicious activity conducted by the device.

Initially, Darktrace observed the device making type A DNS requests for 'wiki.wireshark[.]org' immediately before making type A DNS requests for the domain names 'www.googleadservices[.]com', 'allpcsoftware[.]com', and 'wireshhark[.]com' (note the two 'h's). This pattern of activity indicates that the device’s user was redirected to the website, wireshhark[.]com, as a result of the user's interaction with a sponsored Google Search result pointing to allpcsoftware[.]com.

At the time of analysis, navigating to wireshhark[.]com directly from the browser search bar led to a YouTube video of Rick Astley's song "Never Gonna Give You Up". This suggests that the website, wireshhark[.]com, had been configured to redirect users to this video unless they had arrived at the website via the relevant sponsored Google Search result [8].

Although it was not possible to confirm this with certainty, it is highly likely that users who visited the website via the appropriate sponsored Google Search result were led to a fake website (wireshhark[.]com) posing as the legitimate website, wireshark[.]com. It seems that the actors who set up this fake version of wireshark[.]com were inspired by the well-known bait-and-switch technique known as 'rickrolling', where users are presented with a desirable lure (typically a hyperlink of some kind) which unexpectedly leads them to a music video of Rick Astley's "Never Gonna Give You Up".

After being redirected to wireshhark[.]com, the user unintentionally installed a malware sample which dropped what appears to be Cobalt Strike onto their device. The presence of Cobalt Strike on the user's desktop was evidenced by the subsequent type A DNS requests which the device made for the domain name 'pse[.]ac'. These DNS requests were responded to with the likely Cobalt Strike C2 server address, 194.169.175[.]132. Given that Darktrace only had visibility over the device’s internal traffic, it did not observe any C2 connections to this Cobalt Strike endpoint. However, the desktop's subsequent behavior suggests that a malicious actor had gained 'hands-on-keyboard' control of the device via an established C2 channel.

Figure 2: Advanced Search data showing an customer device being tricked into visiting the fake website, wireshhark[.]com.

Since the malicious actor had gained control of an IT member's device, they were able to abuse the privileged account credentials to spread Python payloads across the network via SMB and the Windows Management Instrumentation (WMI) service. The actor was also seen distributing the Windows Sys-Internals tool, PsExec, likely in an attempt to facilitate their lateral movement efforts. It was normal for this IT member's desktop to distribute files across the network via SMB, which meant that this malicious SMB activity was not, at first glance, out of place.

Figure 3: Advanced Search data showing that it was normal for the IT member's device to distribute files over SMB.

However, Darktrace DETECT recognized that the significant spike in file writes being performed here was suspicious, even though, on the surface, it seemed ‘normal’ for the device. Furthermore, Darktrace identified that the executable files being distributed were attempting to masquerade as a different file type, potentially in an attempt to evade the detection of traditional security tools.

Figure 4: Event Log data showing several Model Breaches being created in response to the IT member's DEVICE's SMB writes of Python-based executables.

An addition to DETECT’s identification of this unusual activity, Darktrace’s Cyber AI Analyst launched an autonomous investigation into the ongoing compromise and was able to link the SMB writes and the sharing of the executable Python payloads, viewing the connections as one lateral movement incident rather than a string of isolated events. After completing its investigation, Cyber AI Analyst was able to provide a detailed summary of events on one pane of glass, ensuring the customer could identify the affected device and begin their remediation.

Figure 5: Cyber AI Analyst investigation summary highlighting the IT member's desktop’s lateral movement activities.

C2 Activity

The Python payloads distributed by the IT member’s device were likely related to the Nitrogen malware, as evidenced by the payloads’ names and by the network behaviours which they engendered.

Figure 6: Advanced Search data showing the affected device reaching out to the C2 endpoint, pse[.]ac, and then distributing Python-based executable files to an internal domain controller.

The internal devices to which these Nitrogen payloads were distributed immediately went on to contact C2 infrastructure associated with Cobalt Strike. These C2 connections were made over SSL on ports 443 and 8443. Darktrace identified the attacker moving laterally to an internal SQL server and an internal domain controller.

Figure 7: Advanced Search data showing an internal SQL server contacting the Cobalt Strike C2 endpoint, 194.180.48[.]169, after receiving Python payloads from the IT member’s device.

Figure 8: Event Log data showing several DETECT model breaches triggering in response to an internal SQL server’s C2 connections to 194.180.48[.]169.

Once more, Cyber AI Analyst launched its own investigation into this activity and was able to successfully identify a series of separate SSL connections, linking them together into one wider C2 incident.

Figure 9: Cyber AI Analyst investigation summary highlighting C2 connections from the SQL server.

Darktrace observed the attacker using their 'hands-on-keyboard' access to these systems to elevate their privileges, conduct network reconnaissance (primarily port scanning), spread Python payloads further across the network, exfiltrate data from the domain controller and transfer a payload from GitHub to the domain controller.

Figure 10: Cyber AI Analyst investigation summary an IP address scan carried out by an internal domain controller.

Figure 12: Event Log data showing an internal domain controller contacting GitHub around the time that it was in communication with the C2 endpoint, 194.180.48[.]169.

Figure 13: Event Log data showing a DETECT model breach being created in response to an internal domain controller's large data upload to the C2 endpoint, 194.180.48[.]169.

After conducting extensive reconnaissance and lateral movement activities, the attacker was observed detonating ransomware with the organization's VMware environment, resulting in the successful encryption of the customer’s VMware vCenter server and VMware virtual machines. In this case, the attacker took around 24 hours to progress from initial access to ransomware detonation.

If the targeted organization had been signed up for Darktrace's Proactive Threat Notification (PTN) service, they would have been promptly notified of these suspicious activities by the Darktrace Security Operations Center (SOC) in the first instance, allowing them to quickly identify affected devices and quarantine them before the compromise could escalate.

Additionally, given the quantity of high-severe alerts that triggered in response to this attack, Darktrace RESPOND would, under normal circumstances, have inhibited the attacker's activities as soon as they were identified by DETECT. However, due to RESPOND not being configured to act on server devices within the customer’s network, the attacker was able to seamlessly move laterally through the organization's server environment and eventually detonate the ALPHV ransomware.

Nevertheless, Darktrace was able to successfully weave together multiple Cyber AI Analyst incidents which it generated into a thread representing the chain of behavior that made up this attack. The thread of Incident Events created by Cyber AI Analyst provided a substantial account of the attack and the steps involved in it, which significantly facilitated the customer’s post-incident investigation efforts.

Figure 14: Darktrace's AI Analyst weaved together 33 of the Incident Events it created together into a thread representing the attacker’s chain of behavior.

Conclusion

It is expected for malicious cyber actors to revise and upgrade their methods to evade organizations’ improving security measures. The continued improvement of email security tools, for example, has likely created a need for attackers to develop new means of Initial Access, such as the use of Microsoft Teams-based malware delivery.

This fast-paced ALPHV ransomware attack serves as a further illustration of this trend, with the actor behind the attack using malvertising to convince an unsuspecting user to download the Python-based malware, Nitrogen, from a fake Wireshark site. Unbeknownst to the user, this stealthy malware dropped a C2 implant onto the user’s device, giving the malicious actor the ‘hands-on-keyboard’ access they needed to move laterally, conduct network reconnaissance, and ultimately detonate ALPHV ransomware.

Despite the non-traditional initial access methods used by this ransomware actor, Darktrace DETECT was still able to identify the unusual patterns of network traffic caused by the attacker’s post-compromise activities. The large volume of alerts created by Darktrace DETECT were autonomously investigated by Darktrace’s Cyber AI Analyst, which was able to weave together related activities of different devices into a comprehensive timeline of the attacker’s operation. Given the volume of DETECT alerts created in response to this ALPHV attack, it is expected that Darktrace RESPOND would have autonomously inhibited the attacker’s operation had the capability been appropriately configured.

As the first post-compromise activities Darktrace observed in this ALPHV attack were seemingly performed by a member of the customer’s IT team, it may have looked normal to a human or traditional signature and rules-based security tools. To Darktrace’s Self-Learning AI, however, the observed activities represented subtle deviations from the device’s normal pattern of life. This attack, and Darktrace’s detection of it, is therefore a prime illustration of the value that Self-Learning AI can bring to the task of detecting anomalies within organizations’ digital estates.

Credit to Sam Lister, Senior Cyber Analyst, Emma Foulger, Principal Cyber Analyst

Appendices

Darktrace DETECT Model Breaches

- Compliance / SMB Drive Write

- Compliance / High Priority Compliance Model Breach

- Anomalous File / Internal / Masqueraded Executable SMB Write

- Device / New or Uncommon WMI Activity

- Anomalous Connection / New or Uncommon Service Control

- Anomalous Connection / High Volume of New or Uncommon Service Control

- Device / New or Uncommon SMB Named Pipe

- Device / Multiple Lateral Movement Model Breaches

- Device / Large Number of Model Breaches

- SMB Writes of Suspicious Files (Cyber AI Analyst)

- Suspicious Remote WMI Activity (Cyber AI Analyst)

- Suspicious DCE-RPC Activity (Cyber AI Analyst)

- Compromise / Connection to Suspicious SSL Server

- Compromise / High Volume of Connections with Beacon Score

- Anomalous Connection / Suspicious Self-Signed SSL

- Anomalous Connection / Anomalous SSL without SNI to New External

- Compromise / Suspicious TLS Beaconing To Rare External

- Compromise / Beacon to Young Endpoint

- Compromise / SSL or HTTP Beacon

- Compromise / Agent Beacon to New Endpoint

- Device / Long Agent Connection to New Endpoint

- Compromise / SSL Beaconing to Rare Destination

- Compromise / Large Number of Suspicious Successful Connections

- Compromise / Slow Beaconing Activity To External Rare

- Anomalous Server Activity / Outgoing from Server

- Device / Multiple C2 Model Breaches

- Possible SSL Command and Control (Cyber AI Analyst)

- Unusual Repeated Connections (Cyber AI Analyst)

- Device / ICMP Address Scan

- Device / RDP Scan

- Device / Network Scan

- Device / Suspicious Network Scan Activity

- Scanning of Multiple Devices (Cyber AI Analyst)

- ICMP Address Scan (Cyber AI Analyst)

- Device / Anomalous Github Download

- Unusual Activity / Unusual External Data Transfer

- Device / Initial Breach Chain Compromise

MITRE ATT&CK Mapping

Resource Development techniques:

- Acquire Infrastructure: Malvertising (T1583.008)

Initial Access techniques:

- Drive-by Compromise (T1189)

Execution techniques:

- User Execution: Malicious File (T1204.002)

- System Services: Service Execution (T1569.002)

- Windows Management Instrumentation (T1047)

Defence Evasion techniques:

- Masquerading: Match Legitimate Name or Location (T1036.005)

Discovery techniques:

- Remote System Discovery (T1018)

- Network Service Discovery (T1046)

Lateral Movement techniques:

- Remote Services: SMB/Windows Admin Shares

- Lateral Tool Transfer (T1570)

Command and Control techniques:

- Application Layer Protocol: Web Protocols (T1071.001)

- Encrypted Channel: Asymmetric Cryptography (T1573.002)

- Non-Standard Port (T1571)

- Ingress Tool Channel (T1105)

Exfiltration techniques:

- Exfiltration Over C2 Channel (T1041)

Impact techniques:

- Data Encrypted for Impact (T1486)

List of Indicators of Compromise

- allpcsoftware[.]com

- wireshhark[.]com

- pse[.]ac • 194.169.175[.]132

- 194.180.48[.]169

- 193.42.33[.]14

- 141.98.6[.]195

References

[1] https://darktrace.com/threat-report-2023

[2] https://www.microsoft.com/en-us/security/blog/2022/05/09/ransomware-as-a-service-understanding-the-cybercrime-gig-economy-and-how-to-protect-yourself/

[3] https://www.bleepingcomputer.com/news/security/alphv-blackcat-this-years-most-sophisticated-ransomware/

[4] https://www.cisa.gov/news-events/cybersecurity-advisories/aa23-353a

[5] https://www.justice.gov/opa/pr/justice-department-disrupts-prolific-alphvblackcat-ransomware-variant

[6] https://www.state.gov/u-s-department-of-state-announces-reward-offers-for-criminal-associates-of-the-alphv-blackcat-ransomware-variant/

[7] https://www.bleepingcomputer.com/news/security/blackcat-alphv-ransomware-linked-to-blackmatter-darkside-gangs/

[8] https://www.trendmicro.com/en_us/research/23/f/malvertising-used-as-entry-vector-for-blackcat-actors-also-lever.html

[9] https://news.sophos.com/en-us/2023/07/26/into-the-tank-with-nitrogen/

[10] https://www.esentire.com/blog/persistent-connection-established-nitrogen-campaign-leverages-dll-side-loading-technique-for-c2-communication

[11] https://www.esentire.com/blog/nitrogen-campaign-2-0-reloads-with-enhanced-capabilities-leading-to-alphv-blackcat-ransomware

[12] https://www.esentire.com/blog/the-notorious-alphv-blackcat-ransomware-gang-is-attacking-corporations-and-public-entities-using-google-ads-laced-with-malware-warns-esentire

執筆者

Sam Lister

Specialist Security Researcher

Inside the SOC

Written by

Sam Lister

Specialist Security Researcher

Senior Director of Product

Watch the NIS2 Webinar

Blog

AI

August 19, 2026

When AI Becomes the Lure: A Fake Gemini Installer Delivers Vidar

Key takeaways

‍

Darktrace observed a customer download a fake Google Gemini installer hosted on Google Colab, resulting in the execution of the Vidar information stealer.
Darktrace identified the compromise through behavioral indicators, including suspicious process activity, anomalous network communications, and indicators of credential theft, before autonomously containing the threat.
The incident highlights how threat actors are increasingly exploiting trusted platforms and a growing interest in AI tools to distribute malware through seemingly legitimate software acquisition workflows.

‍

The Growing Abuse of Generative AI

As organizations are increasingly adopting generative AI tools into their daily workflows, attackers are adapting their distribution methods accordingly too. As part of their day-to-day work, users are now searching for AI assistants, programming tools, browser extensions, desktop applications, and productivity integrations.

Recent reports have highlighted campaigns that use fake AI software and AI-related installers to distribute malware and steal credentials [1]. Researchers have documented campaigns that exploit fake AI-themed websites and services to distribute information stealers and backdoors [2]. Security researchers have also observed attackers disguising malware as legitimate installers for AI software to increase the likelihood of victim interaction and execution [3].

In July 2026, Darktrace observed one such case within a customer environment in the Europe, Middle East and Africa (EMEA) region, where attackers used a fake generative AI installer to deliver the prolific information stealer Vidar. This incident highlights how threat actors are exploiting interest in AI services to distribute established malware using increasingly convincing social engineering techniques.

How a Fake Gemini Installer Delivered Vidar

Initial Access: From Search Result to Malware Download

Unlike many malware campaigns that begin with a phishing email, this activity appears to have originated from a user searching for and downloading software.

Darktrace first observed unusual activity on the customer network after a suspicious executable file was launched from a user’s Download folder. Further investigation revealed that the file purported to be a Google Gemini installer and was named “Download_Google_Gemini_For_Windows.exe”.

During the initial analysis, it was noted that the top search result for the suspicious filename associated pointed to a file hosted on Google Colab, a cloud-based Jupyter notebook platform, commonly used by developers, researchers, and data scientists to run code and machine learning workloads through a web browser. By leveraging another trusted Google platform, the attacker increased the likelihood that users would perceive the download as legitimate, making the lure more convincing to those searching for Gemini-related software.

‍

Figure 1: The Google Colab page containing a download prompt for the fake Google Gemini installer.

Further investigation of the Google Colab page revealed that the download prompt redirected users to a secondary site, hxxps://micronsoftwares[.]com, which posed as a "Windows Software Hub" download page and offered the fake Gemini installer for download.

‍

Figure 2: The secondary website posing as a "Windows Software Hub" download page, which likely hosted the fake Gemini installer.

While the investigation did not uncover any HTTP or file-download telemetry data that conclusively identified the download source, SSL communication sessions with Google Colab were detected immediately before the suspicious file was executed. The timing of these connections suggests that the user interacted with the Colab resource before being redirected to the secondary site from which the executable was downloaded.

‍

The user was not simply tricked into opening an email attachment; instead, the attacker embedded malicious content into a process many users would consider entirely legitimate: searching for and downloading software associated with a trusted platform.

‍

Weaponizing Trusted Platforms

At the time of review (July 15, 2026), Darktrace's Threat Research team confirmed that the Google Colab page was still active and prompting users to download a ZIP archive containing the binary file.

‍

The archive also appeared to contain a README file instructing users to run the binary file with administrator privileges and add it to their antivirus software’s exception lists. These instructions suggest that the campaign relied heavily on social engineering, convincing users to take actions that would facilitate malware execution and potentially bypass security checks.

The use of a legitimate platform also complicates the user’s decision-making. Downloads associated with a trusted service are often perceived as less suspicious than those hosted on unfamiliar domains. When combined with the branding of a widely used AI tool, the lure becomes even more convincing.

Malware Analysis

Darktrace’s Threat Research team identified the executable file as the information-stealing malware Vidar. Analysis revealed that the binary file was a newer Go-compiled variant that communicated with Telegram-based infrastructure. Darktrace’s researchers also identified dtm[.]kijangturbo88[.]top as a command-and-control (C2) endpoint associated with the activity. While the malware itself was not novel, the lure and delivery mechanism was.

For a deeper look at the information stealer, see Darktrace’s 2023 analysis of Vidar.

‍

**Figure 3:** Darktrace’s detection of the unusual outbound connection associated with the fake Gemini installer.

Shortly after execution, the process established communications with the external IP address 91.98.98[.]86 via port 443, directly linking the executable to suspicious network activity observed on the device. Subsequent open-source intelligence (OSINT) analysis of the revealed multiple malicious associations [5].

Additional Darktrace detections included unusual SSL activity from the affected device. Analysis of related SSL telemetry identified 91.98.111[.]49 as additional infrastructure associated with the activity [6].

Subsequent alerts from the customer's Microsoft Defender for Endpoint integration later confirmed activity consistent with the theft of browser credentials and other sensitive data from the affected endpoint.

Taken together, these detections provided a clear picture of the attack, from the execution of a suspicious file and unusual network connections to indicators of C2 activity and credential theft.

‍

**Figure 4:** Darktrace’s detection of anomalous activity following the execution of the fake Gemini installer, seen in the Model Alert Event Log.

Darktrace's Autonomous Response

Following the detection, Darktrace’s Autonomous Response took immediate containment action, including blocking communication with suspicious external infrastructure, including 91.98.98[.]86, and quarantining the compromised device.

Despite the apparent legitimacy of the activity, with the installer hosted on a trusted platform and resembling a routine software download, Darktrace was able to detect and contain the attack because the device's behavior deviated from its normal pattern.

‍

**Figure 5:** Automated containment actions implemented by Darktrace's Autonomous Response following the detection of activity associated with the fake Gemini installer.

Conclusion

This investigation highlights how threat actors continue to adapt established malware delivery techniques to emerging technology trends. While the malware itself was not new, the distribution method was. By disguising Vidar as a Google Gemini installer and hosting the malicious content on a trusted platform, the attack aligned its lure with a growing behavioral trend: users actively searching for AI tools and services as part of their day-to-day work.

Although fake installers are not a new phenomenon, the rapid rise of generative AI has created new opportunities for threat actors. Rather than relying solely on traditional delivery methods, attackers can now target users who are actively searching for AI applications. As AI adoption continues to accelerate across enterprise environments, organizations should remain alert to campaigns that exploit this interest through fake applications, malicious websites, manipulated search results, the misuse of trusted platforms, and AI-themed social engineering.

Credit to Rushanth Ramanathan (Cyber Analyst) Joanna Ng (Detection Engineer)

Edited by Ryan Traill (Content Manager)

Appendices

Darktrace Model Detections

Security Integration / C2 Activity and Integration Detection
Endpoint / New Suspicious Executable Launched
Endpoint / Process Connection / Unusual Connection from New Process
Anomalous Connection / Rare External SSL Self-Signed
Security Integration / High Severity Integration Detection
Antigena / Network / Significant Anomaly / Antigena Significant Security Integration and Network Activity Block

•Antigena / Network / Significant Anomaly / Antigena Significant Anomaly from Client Block

‍

List of Indicators of Compromise (IoCs)

‍

IoC	Type	Description
`Download_Google_Gemini_For_Windows.exe`	File	Fake Gemini-themed installer observed during the investigation.
`GoogleAppInstaller.exe`	File	Related executable identified through endpoint telemetry.
`91.98.98[.]86`	IP Address	External destination contacted by the malicious executable.
`91.98.111[.]49`	IP Address	Related infrastructure identified through SSL certificate pivoting.
`dtm[.]kijangturbo88[.]top`	Domain	Command-and-control endpoint identified during malware analysis.
`1e13c2c9eac72daf63fd00a9946878949e159ae6ec51b54ec64f942d79d61913`	SHA256	Malware sample associated with the fake Gemini installer.

MITRE ATT@CK Mapping

‍

MITRE ATT&CK Mapping	Tactic	Technique
Initial Access	`T1204`	User Execution
Execution	`T1204.002`	User Execution: Malicious File
Defence Evasion	`T1036`	Masquerading
Credential Access	`T1555`	Credentials from Password Stores
Credential Access	`T1555.003`	Credentials from Web Browsers
Command and Control	`T1071`	Application Layer Protocol
Exfiltration	`T1041`	Exfiltration Over C2 Channel

About the author

Rajendra Rushanth

Cyber Analyst

Blog

AI

August 13, 2026

信頼が新たなアタックサーフェスである理由：ダークトレースの2026年度中間脅威アップデート

2026年初頭、ダークトレースのアナリストによって専用に設計されたReact2Shellハニーポットは、その展開から2時間未満で侵害を受けました。この1つのデータをとって見ても2026年前半の脅威ランドスケープのペースを表していると言うことができますが、スピードは1つの側面に過ぎません。

過去6か月間の変化は、従来のマルウェアや脆弱性を中心とした攻撃から、信頼されるアイデンティティ、プラットフォーム、およびインフラの悪用への移行を示しています。多くの組織が広範囲にAIを導入するのに伴い、アイデンティティ、SaaSプラットフォーム、クラウドに対する権限、自動化フレームワーク、および非人間アイデンティティが、攻撃者に好まれる攻撃経路となっています。

攻撃者は、外部から侵入するよりも、防御者やユーザーが信頼するよう条件づけられている関係、サービス、認証済みチャネル内で活動する傾向が高まっています。

2025年と比較して何が変わったか？

2025年には、アイデンティティが新たな境界となり、攻撃者はますます従来のエクスプロイト手法を使うよりも信頼されるアカウント、SaaSプラットフォーム、そしてAIによって可能になった新手の手法を使うようになりました。2026年前半は、この進化の次の段階にあると言えます。アイデンティの悪用は依然として中心的ですが、信頼の問題は今やアカウントをはるかに超えてEメール認証、クラウドに対する権限、ソフトウェアサプライチェーン、AIゲートウェイ、リモート管理ツール、そして非人間アイデンティティにも及んでいます。

Theme	2025 (Mid-Year / Annual)	H1 2026
Identity	Credentials remained the weak link; identity emerged as the new perimeter.	Identity remains the entry point, but trust has become the new attack surface.
Cloud & SaaS	SaaS-targeted ransomware continued to rise.	Cloud and SaaS became the attacker's preferred operating environment.
AI	Large Language Models (LLMs) were suspected of influencing phishing shifts.	LLM-generated malware, compromised AI proxies, and the abuse of AI identities emerged.
Attack Surface Scale & Speed	Exponential growth of Common Vulnerabilities and Exposures (CVEs), with public proof-of-concepts appearing faster.	Cloud and AI adoption expanded the attack surface, while AI accelerated exploitation. One honeypot was compromised in under two hours.
Supply Chain	Legitimate services were increasingly abused.	Trusted maintainers and CI/CD workflows were weaponized.

アイデンティティとEメール：危険にさらされる信頼のシグナル

Eメールは信頼されるアイデンティティへの最も確実なルートであり、攻撃者はノイズを作り出すことよりも品質の向上に投資していることをデータが示しています。2026年前半において、フィッシングEメールの67%がDMARC認証を通過していました。ほとんどのフィッシングの試みに対して、阻止するには認証だけではもはや不十分なのです。フィッシングEメールの25.8%でVIPユーザーが標的となっており、22025年と同様に25%を上回る値ですが、期間全体を通じて上昇傾向にあります。重要な点は、フィッシングの巧妙さが高まり続けていることです：フィッシングEメールの37％が大量のテキストを含んでおり、2025年前半の32％から増加しています。また、39％は新手のソーシャルエンジニアリングテクニックを特徴としており、攻撃者が特定の標的にあわせてさらなるカスタマイズを行っていることを示唆しています。

ダークトレースの顧客に最も幅広く影響を与えている脅威もアイデンティティに焦点を当てたものであり、専用のStealCおよびAMOSインフォスティーラーを使った攻撃がこの半年間に多く見られています。これらのマルウェアの蔓延も、根本的にはアイデンティティの問題です。インフォスティーラーによって収集された認証情報がしばしば最初のアクセスベクトルとして使用され、攻撃チェーンの後の段階で格段に影響の大きい侵害につながります。また、マルウェアの投下には技術的弱点の悪用をほとんど必要としない点が重要です。ユーザー自身の手で悪意あるコードを実行させるClickFixソーシャルエンジニアリング手法は、依然としてよく見られます。最近観測されたある攻撃キャンペーンは17か国のダークトレース顧客で確認され、最も影響を受けたのは米国でした。この侵害はソフトウェアの欠陥から始まったものではなく、信頼されるユーザーが信頼されるアクションを実行したことが端緒となりました。

サプライチェーン：信頼が大規模に武器化される

3月と4月には共通の教訓が確認されました。それは信頼がサプライチェーン脆弱性となったということです。Axiosの侵害では幅広く使用されているメンテナーへの信頼が悪用され、Trivyキャンペーンは信頼されるCI/CDインフラ、リリースアーティファクト、およびコンテナイメージを利用して、正当な開発ワークフローに悪意あるコードを押し込みました。

これを最もはっきりと示した事例は、2月から3月にかけて観測されたキャンペーンです。多数のデバイスがHola VPNを使用している間に悪意のあるペイロードをダウンロードし、その後、Holaの配信パイプラインの侵害に関連していることが判明しました。ダークトレースの脅威調査チームは、公開アドバイザリがリリースされる前に、複数の顧客において繰り返し発生している異常な動作から、この侵害に関連する活動を特定しました。

最近では、攻撃者が正当なブロックチェーンインフラを悪用して、AMOSやPhexia等のインフォスティーラーを投下しているケースが見られます。セキュリティリソースが限られているユーザーによく使用されている一般的なVPN等のツールは、正当なC2インフラと組み合わせることで、攻撃者がはるかに広範な被害者層に到達することを可能にする一方、防御側は関連するエンドポイントを単純にブロックすることはできないため、やっかいな問題になります。

防御者にとっての課題は、もはや悪意のあるインフラを特定することではなく、信頼されるインフラが悪意のある動作を始めたときにそれを認識できるかどうかということです。

クラウドとSaaS：標的から作戦領域へ

5月から6月には、デバイス登録、クラウドデータ窃取、SaaS悪用、RDPの拡大、リモート管理ツール関連の活動が見られ、攻撃者がクラウドやSaaSを単なる標的としてではなく、むしろ好ましい作戦環境としてますます認識していることが示唆されました。

ダークトレース顧客でのある事例では、1つの侵害されたSaaSアカウントがEメール、SaaS、ネットワークレイヤーにわたる活動を引き起こし、これには受信トレイルールの変更、フィッシングの拡散、疑わしいインフラへの接続が含まれていました。これらの兆候のいずれも単独では決定的とは言えませんでしたが、総合すると明らかに侵入を示していました。これらの環境において、攻撃者はますます信頼管理をを回避する必要がなくなっています。信頼は、侵害されたアイデンティティ、委任されたアクセス権、および正当な管理ツールを通じて引き継ぐことができるからです。これは、2025年に見られたSaaSを標的としたランサムウェアの傾向の自然な進化であると言えます。ますます多くのケースにおいて、ビジネスが運営されるのと同じプラットフォームが、敵対者が作戦を展開するプラットフォームとなっています。

AI：攻撃の加速装置でありアタックサーフェス、そして信頼される、しかしリスクの高いアクター

信頼がアタックサーフェスなら、AIはそれが最も急激に拡大している領域であると言えます。ダークトレースの顧客基盤全体において、2026年前半にAIサービスへの接続は平均13％増加し、接続数は1600万回以上、標準的な組織は7つの異なるAIプロバイダーとやり取りするようになっています。そしてAIはもはや企業のごく一部ではありません。日々の業務に組み込まれています。この変化が3つの問題を生み出しましたが、これらはすべて2026年前半にダークトレースによって観測されています。

1. 攻撃倍増装置としてのAI

ダークトレースは、AIによって生成されたReact2Shellを悪用するマルウェアを特定しました。これは攻撃者がLLMを使用して有効なエクスプロイトコードを生成し大規模に展開したものです。脅威ランドスケープ全体で見ても同様の活動が増えており、効果的な攻撃作戦への参入障壁が崩れつつあることを示唆しています。最近のJadePuffer事例にも見られるように、エージェント型脅威アクターがインターネットに接続されたサーバーの脆弱性をエクスプロイトし、その後完全に自動化されたランサムウェア攻撃が開始されており、AIは脆弱性の公開から実際のエクスプロイトまでの過程を加速させています[1]。

2. アタックサーフェスとしてのAI

今やAIレイヤー自体も、調べてみる価値があります。あるオートメーション技術メーカーにおいて、侵害されたLLMプロキシが他のAIサービスへの踏み台として利用され、それが失敗すると攻撃者は暗号通貨マイニングに切り替えたという事例がありました。DarktraceのCyber AI Analystはこの侵入インシデントを明らかにし、Managed Threat Detectionサービスにより顧客への通知が行われたことにより、事態がそれ以上進行する前に封じ込めることができました。実務者にとっての教訓は明確です：AIゲートウェイ、プロキシ、モデルエンドポイントは、本番環境のクラウドワークロードと同様に扱うべきです。なぜなら、攻撃者は既にそうしているからです。

3. 信頼される、ただしリスクもあるアクターとしてのAI

Darktrace/ SECURE AIの観測結果からわかることは、最も一般的な現実世界のリスクはさらに判別が難しいということです。従業員が、個人識別情報（PII）、税務記録、身分証明書、会社の財務データ、人事記録、個人の医療データをLLMのプロンプトに入力することや、シャドーAIの蔓延、そしてモバイルデバイスからのAI使用の増加が挙げられます。28日間で約28,000人のユーザーから送信された約280,000件のプロンプトのうち、Darktraceはこれらのプロンプトの約1%（2,945件）に機密性の高いデータが含まれていることを特定しました。*

‍_{*プロンプトデータはユーザーのプライバシーを保護するために、集積および匿名化された形で分析されました。}

防御者にとって、課題はコンテキストです。つまり、正当なビジネス利用が重大なリスクに変わるタイミングを、プライバシーやユーザーの信頼を損なうことなく見極めることです。組織がAIシステムをますます信頼し、AIがマシンスピードで機密情報にアクセス、処理、共有するようになるなかで、アイデンティティ、アプリケーション、クラウドインフラと同様に、AIも保護および監視されなければなりません。

スピードと地政学：迅速な作戦、長期的な目標

2026年前半に行われたいくつかの調査では、攻撃者が新たに公開された脆弱性をいかに迅速に実用化し、パッチ適用サイクルが完了する前にOAST（Out-of-Band Application Security Testing）インフラおよび信頼されたクラウドサービスを通じてエクスプロイト検証を行っているかが明らかになりました。React2Shellは2時間で侵害され、BeyondTrustのエクスプロイトも1日未満で発生しました。このような状況の中で、国家を背後に持つ脅威アクターは、引き続き正当なサービス、クラウドインフラ、および信頼関係を通じた、長期的なアクセス、情報収集、および事前配置に重点を置いています。中国、ロシア、イラン、北朝鮮（DPRK）を背後に持つ作戦は共通の特徴を持っています。それは即座に混乱を起こすことよりも、永続化と戦略的な配置を優先するということです。

中国：ダークトレースは、中国系アクターが信頼されるサービス、DLLサイドローディング、およびモジュール型の侵入チェーンを通じた長期的なアクセスの獲得を優先している状況を確認しています。これはCrimson Echoレポートでも解説され、Twill Typhoonの手法に関連するアクティビティと一致しています。

イラン：ダークトレースによるZionSiphonの調査は、イランに関連するアクターのOT環境への関心を浮き彫りにし、諜報目的とインフラ破壊能力を融合させている実態を明らかにしています。

ロシア：ダークトレースによる調査、および業界全体のさまざまな報告は、ロシアがウクライナ関連の支援に関する長期的な情報収集のために、信頼関係とサプライチェーンを標的としていることを明らかにしています[2]。

北朝鮮：ダークトレースは、脆弱性の迅速な武器化と永続的アクセス技術を組み合わせた北朝鮮関連の活動を観測ししています。これにはAxiosのサプライチェーン侵害、React2Shellエクスプロイト、ステルス性のmacOS侵入が含まれています。

目的は脅威アクターによって異なっていましたが、手法は非常に一貫していました。信頼されるサービス、正当なインフラ、そして永続的アクセスは、即時の混乱よりも価値が高いことが示されています。

防御者のシフト

アイデンティティの侵害、サプライチェーン攻撃、SaaSの悪用、AIインフラの標的化、国家が支援する作戦、いずれのケースにおいても、攻撃者は防御のコントロールを突破するのではなく、信頼されたシステムを通じて行動することで成功を収めることが増えています。信頼されるユーザー、信頼されるソフトウェア、信頼されるインフラチャ、そしてますます信頼されるAIシステムは、すべて有効な攻撃経路となりました。

防御者にとっての課題はもはや、単にあるアクションが許可されるかどうかを判断することではなく、そのアクションがより広いコンテキストの中で正当かどうかを見極めることです。認証、評判、出自は依然として重要ですが、それだけではもはや十分とは言えません。攻撃者がますます信頼されるシステム内で活動するようになる中で、最も強力なシグナルは多くの場合動作の逸脱です。つまり、信頼されているアクティビティが期待される振る舞いと一致しなくなった時にそれを識別することです。

本稿の執筆にはNathaniel Jones（SVP, Global Threat Intelligence）、Emma Foulger（Global Threat Research Operations Lead）、Justin Torres （Senior Cyber Analyst） Daniel Levy（Threat Hunting Data Scientist）が協力しました。

‍
編集：Ryan Traill（Content Manager）

‍

付録1：脅威調査手法

ダークトレースの脅威調査チームは、顧客の運用環境に対する詳細な調査を行ってアクティブな脅威を識別し、主要な侵害インジケーター（IoC）を特定し、関連する脅威インテリジェンスを提供しています。この調査はダークトレースの異常ベース検知に基づいたもので、脅威調査チームによる徹底した分析およびコンテキスト化が行われています。検知された脅威は関連する顧客のセキュリティチームに直ちに報告されます。顧客がダークトレースの自律遮断テクノロジーを使用している場合、これらの脅威は速やかに緩和され、エスカレーションが阻止されます。

2026年1月1日から6月30日までの期間、ダークトレースは顧客ベースにおいて多種多様なサイバー脅威を調査しました。その多くは同様のTTPおよびIoCが短い期間内に一定の数の顧客に影響を及ぼした、複数の顧客を標的とした攻撃作戦的活動であったことが判明しています。

Eメールに関連する統計は、2026年1月1日から6月30日までの間に、すべてのクラウドホスト型顧客環境におけるDarktrace / EMAILの集約されたデータから導出されています。特異な観測値を除外するために、集約を行う前に標準的なデータ品質フィルタリングが適用されています。地域別の統計は、このデータセットの関連サブセットに基づいています。

付録2：キャンペーン - 地域およびセクター別の傾向

過去6か月間の脅威ランドスケープは上記の大まかなテーマにより定義されていますが、ダークトレースの顧客基盤全体でのキャンペーンクラスタリングにより、それらがセクター、地域、産業ごとにどのように異なっているかが明らかになりました。

ダークトレースの脅威調査チームは、顧客ベースに影響を与えるさまざまな脅威を調査しています。この調査を通じて、共通のTTPやインフラが観察され、短期間で多くの顧客に影響を与える、攻撃キャンペーンのような活動のクラスターが特定されました。

セクターおよび産業は、一貫した分類を確保するために標準産業分類（SIC）システムを使用して分類されています。本レポートのセクターおよび地域別の考察は、より広範な世界的な傾向を反映している一方で、ダークトレースの顧客基盤の分布にも影響を受けています。例えば、金融、製造、教育分野はダークトレースの顧客の中で多く、これらのセクターで観測される事例数が多くなる可能性があります。これは必ずしも特定のセクターにおけるリスクが高いことを示すものではなく、顧客の分布を反映しています。同様に、地域の傾向はダークトレースの顧客の地理的分布によって影響を受ける可能性があります。

2026年前半にダークトレースの脅威調査チームによって特定されたキャンペーンクラスターの分析では、明確な地域別の傾向が明らかになりました：

ヨーロッパ、中東、アフリカ（EMEA）がすべてのキャンペーンクラスター事例の60%を占めており支配的でした。
アメリカ大陸（AMS）は、その次に影響が大きかった地域でありキャンペーンクラスター事例の30%を占めていました。
アジア太平洋および日本（APJ）地域はキャンペーンクラスターの影響をあまり受けていませんでした。これは脅威アクターがこの地域の優先度を低くし、代わりに他の地域に注力した可能性を示しています。

産業セクターの標的化も地域によって大きく異なりました：

EMEA地域では、情報通信分野が大きく影響を受け、全体の25%を占めました。
対照的に、AMSでは標的はより均等に分布しており、教育、行政機関および防衛、金融および保険の各セクターがいずれもAMS地域の事例の20%以上を占めていました。
APJ地域ではキャンペーン活動はより均等に分散しており、特定のセクターが支配的な標的として浮上することはありませんでした。

いくつかの国はそれぞれの地域内でも際立った特徴がありました：

米国はAMS内のすべてのキャンペーンクラスターの60%を占めていました。
日本はAPJ全体のキャンペーン事例の40%を占めました。
EMEA地域では、英国とジンバブエがそれぞれ特定された事例の23%を占めており、両国ともさまざまなキャンペーンのタイプに影響を受けています。

‍

Inside the SOCおよび2026年度脅威調査の月次傾向：アクセスからインパクトまで

Month	Dominant Themes
January	Voice phishing, VPS infrastructure, WebSocket C2, RMM abuse, ransomware, infostealers (StealC), and trojanized installers (7-Zip).
February	Voice phishing, VPN intrusion, edge infrastructure compromise (BeyondTrust), and RMM abuse.
March	Sustained supply chain compromise (Hola VPN, Axios, Trivy), malicious browser extensions, phishing, and discovery tools.
April	Account creation abuse, payload delivery, VPN credential abuse, Fortinet exploitation, and botnet activity.
May	PowerShell, EtherHiding, data exfiltration, VPN access, business email compromise (BEC), ClickFix, and infostealers (AMOS).
June	RDP abuse, device registration, RMM usage, voice phishing, cloud data theft, botnet activity, blockchain abuse, ClickFix, and infostealers (AMOS).