ブログ
/
Network
/
November 6, 2023

How PlugX Malware Has Evolved & Adapted

Discover how Darktrace effectively detected and thwarted the PlugX remote access trojan in 2023 despite its highly evasive and adaptive nature.
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
Nahisha Nobregas
SOC Analyst
Default blog image
06
Nov 2023

What is PlugX Remote Access Trojan?

Understanding remote access trojans (RATs)

As malicious actors across the threat landscape continue to pursue more efficient and effective ways of compromising target networks, all while remaining undetected by security measures, it is unsurprising to see an increase in the use of remote access trojans (RATs) in recent years. RATs typically operate stealthily, evading security tools while offering threat actors remote control over infected devices, allowing attackers to execute a wide range of malicious activities like data theft or installing additional malware.

Definition and general functionality of RATs: A Remote Access Trojan (RAT) is a type of malware that enables unauthorized remote control of an infected computer. Once installed, RATs allow attackers to monitor user activities, steal sensitive information, manipulate files, and execute commands. RATs are typically distributed via phishing emails, malicious attachments, drive-by downloads, or exploiting software vulnerabilities. Due to their ability to provide comprehensive control over a compromised system, RATs pose a significant security threat to individuals and organizations.

Historical overview of PlugX

PlugX is one such example of a RAT that has attributed to Chinese threat actors such as Mustang Panda, since it first appeared in the wild back in 2008. It is known for its use in espionage, a modular and plug-in style approach to malware development. It has the ability to evolve with the latest tactics, techniques, and procedures (TTPs) that allow it to avoid the detection of traditional security tools as it implants itself target devices.

How does PlugX work?

The ultimate goal of any RAT is to remotely control affected devices with a wide range of capabilities, which in PlugX’s case has typically included rebooting systems, keylogging, managing critical system processes, and file upload/downloads. One technique PlugX heavily relies on is dynamic-link library (DLL) sideloading to infiltrate devices. This technique involves executing a malicious payload that is embedded within a benign executable found in a data link library (DLL) [1]. The embedded payload within the DLL is often encrypted or obfuscated to prevent detection.

What’s more, a new variant of PlugX was observed in the wild across Papua New Guinea, Ghana, Mongolia, Zimbabwe, and Nigeria in August 2022, that added several new capabilities to its toolbox.

Key capabilities of PlugX

The new variation is reported to continuously monitor affected environments for new USB devices to infect, allowing it to spread further through compromised networks [2]. It is then able to hide malicious files within a USB device by using a novel technique that prevents them from being viewed on Windows operating systems (OS). These hidden files can only be viewed on a Unix-like (.nix) OS, or by analyzing an affected USB devices with a forensic tool [2]. The new PlugX variant also has the ability to create a hidden directory, “RECYCLER.BIN”, containing a collection of stolen documents, likely in preparation for exfiltration via its command and control (C2) channels. [3]

Since December 2022, PlugX has been observed targeting networks in Europe through malware delivery via HTML smuggling campaigns, a technique that has been dubbed SmugX [4].

This evasive tactic allows threat actors to prepare and deploy malware via phishing campaigns by exploiting legitimate HTML5 and JavaScript features [5].

Darktrace Coverage of PlugX

Between January and March 2023, Darktrace observed activity relating to the PlugX RAT on multiple customers across the fleet. While PlugX’s TTPs may have bypassed traditional security tools, the anomaly-based detection capabilities of Darktrace allowed it to identify and alert the subtle deviations in the behavior of affected devices, while Darktrace was able to take immediate mitigative action against such anomalous activity and stop attackers in their tracks.  

C2 Communication

Between January and March 2023, Darktrace detected multiple suspicious connections related to the PlugX RAT within customer environments. When a device has been infected, it will typically communicate through C2 infrastructure established for the PlugX RAT. In most cases observed by Darktrace, affected devices exhibited suspicious C2 connections to rare endpoints that were assessed with moderate to high confidence to be linked to PlugX.

On the network of one Darktrace customer the observed communication was a mix of successful and unsuccessful connections at a high volume to rare endpoints on ports such as 110, 443, 5938, and 80. These ports are commonly associated with POP3, HTTPS, TeamViewer RDP / DynGate, and HTTP, respectively.  Figure 1 below showcases this pattern of activity.

Figure 1: Model Breach Event Log demonstrating various successful and unsuccessful connections to the PlugX C2 endpoint 103.56.53[.]46 via various destination ports.

On another customer’s network, Darktrace observed C2 communication involving multiple failed connection attempts to another rare external endpoint associated with PlugX. The device in this case was detected attempting connections to the endpoint, 45.142.166[.]112 on ports 110, 80, and 443 which caused the DETECT model ‘Anomalous Connection / Multiple Failed Connections to Rare Endpoint’ to breach. This model examines devices attempting connections to a rare external endpoint over a short period of time, and it breached in response to almost all PlugX C2 related activity detected by Darktrace. This highlights Darktrace DETECT’s unique ability to identify anomalous activity which appears benign or uncertain, rather than relying on traditional signature-based detections.

Figure 2: Device Event Log demonstrating various successful and unsuccessful connections to the PlugX C2 endpoint 45.142.166[.]112 via various destination on January 27, 2023.

New User Agent

Darktrace's Self-Learning AI approach to threat detection also allowed it to recognize connections to PlugX associated endpoints that utilized a new user agent. In almost all connections to PlugX endpoints detected by Darktrace, the same user agent, Mozilla/5.0 (Windows NT 10.0;Win64;x64)AppleWebKit/537.36, was observed, illustrating a clear pattern in PlugX-related activity

In one example from February 2023, an affected device successfully connected to an endpoint associated with PlugX, 45.142.166[.]112, while using the aforementioned new user agent, as depicted in Figure 3.

Figure 3: The Device Event log above showcases a successful connection to the PlugX associated IP address, 45.142.166[.]112 using the new user agent ‘Mozilla/5.0 (Windows NT 10.0;Win64;x64)AppleWebKit/537.36’.

On March 21, 2023, Darktrace observed similar activity on a separate customer’s network affected by connections to PlugX. This activity included connections to the same endpoint, 45.142.166[.]112. The connection was an HTTP POST request made via proxy with the same new user agent, ‘Mozilla/5.0 (Windows NT 10.0;Win64;x64)AppleWebKit/537.36’. When investigated further this user agent actually reveals very little about itself and appears to be missing a couple of common features that are typically contained in a user agent string, such as a web browser and its version or the mention of Safari before its build ID (‘537.36’).

Additionally, for this connection the URI observed consisted of a random string of 8 hexadecimal characters, namely ‘d819f07a’. This is a technique often used by malware to communicate with its C2 servers, while evading the detection of signature-based detection tools. Darktrace, however, recognized that this external connection to an endpoint with no hostname constituted anomalous behavior, and could have been indicative of a threat actor communicating with malicious infrastructure, thus the ‘Anomalous Connection / Possible Callback URI’ model was breached.

Figure 4: An affected device was detected using the new user agent, ‘Mozilla/5.0 (Windows NT 10.0;Win64;x64)AppleWebKit/537.36’ while connecting to the rare external endpoint 45.142.166[.]112 via proxy.

Numeric File Download

Darktrace’s detection of PlugX activity on another customer’s network, in February 2023, helped to demonstrate related patterns of activity within the C2 communication and tooling attack phases. Observed PlugX activity on this network followed the subsequent pattern; a connection to a PlugX endpoints is made, followed by a HTTP POST request to a numeric URI with a random string of 8 hexadecimal characters, as previously highlighted. Darktrace identified that this activity represented unusual ‘New Activity’ for this device, and thus treated it with suspicion.

Figure 5: New activity was identified by Darktrace in the Device Event Log shown above for connections to the endpoint 45.142.166[.]112 followed by HTTP POSTs to URIs “/8891431c” and “/ba12b866” on February 15, 2023.

The device in question continued to connect to the endpoint and make HTTP POST connections to various URIs relating to PlugX. Additionally, the user agent `Mozilla/5.0 (Windows NT 10.0;Win64;x64)AppleWebKit/537.36` was again detected for these connections. Figure 6 details the activity captured by Darktrace’s Cyber AI Analyst.

Figure 6: The image above showcases activity captured by Darktrace’s AI Analyst for PlugX connections made on February 15, 2023.

Darktrace detected that during these connections, the device in question attempted to download a suspicious file named only with numbers. The use of numeric file names is a technique often used by threat actors to obfuscate the download of malicious files or programs and bypass traditional security tools. Darktrace understood that the download of a numeric file, coupled with the use of an anomalous new user agent, mean the incident should be treated with suspicion. Fortunately, Darktrace RESPOND was enabled in autonomous response mode during this attack, meaning it was able to automatically block the device from downloading the file, or any other files, from the suspicious external location for a two-hour period, potentially preventing the download of PlugX’s malicious tooling.

Conclusion

Amid the continued evolution of PlugX from an espionage tool to a more widely available malware, it is essential that threat detection does not rely on a set of characteristics or indicators, but rather is focused on anomalies. Throughout these cases, Darktrace demonstrated the efficacy of its detection and alerting on emerging activity pertaining to a particularly stealthy and versatile RAT. Over the years, PlugX has continually looked to evolve and survive in the ever-changing threat landscape by adapting new capabilities and TTPs through which it can infect a system and spread to new devices without being noticed by security teams and their tools.

However, Darktrace’s Self-Learning AI allows it to gain a strong understanding of customer networks, learning what constitutes expected network behavior which in turn allows it to recognize the subtle deviations indicative of an ongoing compromise.

Darktrace’s ability to identify emerging threats through anomaly-based detection, rather than relying on established threat intelligence, uniquely positions it to detect and respond to highly adaptable and dynamic threats, like the PlugX malware, regardless of how it may evolve in the future.

Credit to: Nahisha Nobregas, SOC Analyst & Dylan Hinz, Cyber Analyst

Appendices

MITRE ATT&CK Framework

Execution

  • T1059.003 Command and Scripting Interpreter: Windows Command Shell

Persistence and Privilege Escalation

  • T1547.001 Boot or Logon AutoStart Execution: Registry Run Keys / Startup Folder
  • T1574.001 Hijack Execution Flow: DLL Search Order Hijacking
  • T1574.002 Hijack Execution Flow: DLL Side-Loading
  • T1543.003 Create or Modify System Process: Windows Service
  • T1140 Deobfuscate / Decode Files or Information
  • T1083 File and Directory Discovery

Defense Evasion

  • T1564.001 Hide Artifacts: Hidden Files and Directories
  • T1036.004 Masquerading: Task or Service
  • T1036.005 Masquerading: Match Legitimate Name or Location
  • T1027.006 Obfuscated Files or Information: HTML Smuggling

Credential Access

  • T1056.001 Input Capture: Keylogging

Collection

  • T1105 Ingress Tool Transfer

Command and Control

  • T1573.001 Encrypted Channel: Symmetric Cryptography
  • T1070.003 Mail Protocols
  • T1071.001 Web Protocol

DETECT Model Breaches

  • Anomalous Connection / Multiple Failed Connections to Rare Endpoint
  • Anomalous Connection / New User Agent to IP Without Hostname
  • Anomalous File / New User Agent Followed By Numeric File Download
  • Anomalous Connection / Possible Callback URL

Indicators of Compromise (IoCs)

IoC - Type - Description + Confidence

45.142.166[.]112 - IP - PlugX C2 Endpoint / moderate - high

103.56.53[.]46 - IP - PlugX C2 Endpoint / moderate - high

Mozilla/5.0 (Windows NT 10.0;Win64;x64)AppleWebKit/537.36 - User Agent - PlugX User Agent / moderate – high

/8891431c - URI - PlugX URI / moderate-high

/ba12b866 - URI - PlugX URI / moderate -high

References

1. https://www.crowdstrike.com/blog/dll-side-loading-how-to-combat-threat-actor-evasion-techniques/

2. https://unit42.paloaltonetworks.com/plugx-variants-in-usbs/

3. https://news.sophos.com/en-us/2023/03/09/border-hopping-plugx-usb-worm/

4. https://thehackernews.com/2023/07/chinese-hackers-use-html-smuggling-to.html

5. https://www.cyfirma.com/outofband/html-smuggling-a-stealthier-approach-to-deliver-malware/

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
Nahisha Nobregas
SOC Analyst

More in this series

No items found.

Blog

/

AI

/

July 13, 2026

Security After Signatures: Operating in a World of Pre‑CVE Disclosure Exploitation, Collapsed Trust Boundaries, and Autonomous Systems

Default blog imageDefault blog image

Three shifts have reshaped what it means to defend an enterprise securely.  

First, exploitation often begins before defenders have a Common Vulnerabilities and Exposures (CVE) identifier, a security advisory, or an entry in the Cybersecurity and Infrastructure Security Agency's (CISA) Known Exploited Vulnerabilities (KEV) catalog.

Secondly, the trust boundary has moved beyond the network edge into identities, tokens, APIs, and Software-as-a-Service (SaaS) workflows.  

Third, an increasing share of business activity is executed through automation, integrations, and AI agent-like systems that can act faster than teams can verify intent.  

If your security model still relies on detecting known bad artefacts, triaging isolated alerts, and waiting for confirmation before acting, you are already behind the threat.  

This is not a failure of security teams; it’s a failure of the operating model to keep pace with how the environment has changed.

A SOC built around alerts and signatures assumes that malicious activity will eventually surface as an event. In real incidents, however, the decisive evidence is rarely a single event. Instead, it is a chain of individually explainable actions that only appears malicious once you connect the dots across identity, non-human identity, cloud, email, SaaS, operational technology (OT), and network telemetry.

The defenders succeeding today observe behaviors, link them into sequences, understand what those sequences mean, and contain impact before the full story unfolds. That is the operating model the current threat environment demands.  

Exploitation before disclosure

The first shift is the straightforward: the time to exploit has dropped to nearly zero.  

In one example, Darktrace observed a sequence of subtle but strategically significant anomalies within a customer environment that later aligned with exploitation of CVE‑2025‑0994 in Trimble Cityworks by likely Chinese-nexus threat actors. Behavioral indicators were visible at least 18 days before public disclosure, with related anomalies emerging 40 to 50 days earlier during the intrusion window.  

This case illustrates a familiar pattern: clusters of weak‑signal anomalies combing to form an actionable picture of intrusion long before a CVE is published. Such activity reflects long‑horizon, option‑preserving operator models often associated with mature state‑linked activity.  

Figure 1: Darktrace’s detection of malicious exploitation of CVE 2025-0994, later tied to Chinese-nexus threat actors targeting critical national infrastructure (CNI) in the US, weeks before public disclosure.

Throughout 2025 and 2026, Darktrace has continued to observe the value of anomaly-based detections across a range of incidents.

CVE CVE public disclosure date Darktrace detection date Days between detection of exploitation and CVE public disclosure
CVE-2025-0994 Trimble Cityworks 2025-02-06 2025-01-19 18 days
CVE-2025-24183 Apache 2025-03-10 2025-02-18 20 days
CVE-2025-10035 Fortra GoAnywhere 2025-09-18 2025-09-11 7 days
CVE-2026-0257 PAN-OS 2026-05-13

Identity is the real control plane

The second shift is that identity has replaced perimeter as the primary control plane. As Darktrace’s Annual Threat Report 2026 illustrated, identity remains the main challenge in defending against modern intrusions. A clear example is the Adversary-in-the-Middle (AiTM) case published by Darktrace in December 2025. A phishing email led to the compromise of an Office 365 account. Session hijacking bypassed multi-factor authentication (MFA), and the compromised account was used for follow-on phishing and persistence activities including the creation of malicious email rules.  

Every step in that sequence mattered. A successful login alone does not prove legitimacy. An inbox rule, on its own, may not appear catastrophic. Mail activity, viewed in isolation, may seem operationally normal. But the behavioral chain tells a different story: credential theft, token abuse, persistence, and onward compromise through a trusted identity.  

This is why the question is no longer “Did the user authenticate successfully”. The more important question is, “Does this identity action make sense right now, in this context, given what came before it?” The AiTM case shows how identity can be compromised. In practice, however, attacks rarely remained confined to identity alone.  

In another Darktrace case, a compromised SaaS account triggered activity across the email, SaaS, and network layers, including inbox rule changes, phishing propagation, and connections to suspicious infrastructure. Viewed in isolation, none of these events were decisive. Together, however,  they formed a behavioral sequence that revealed the intrusion, with the full attack story automatically correlated and surfaced to defenders by Darktrace’s Cyber AI Analyst.  

Figure 2: Cyber AI Analyst correlated and appended additional events to the incident, including other users who connected to the suspicious redirect link after outbound phishing emails were sent.

AI accelerates the threat  

The third shift is the one many teams still underestimate: trusted tooling, integrations, and AI agent-like systems can create actions that appear legitimate but are strategically dangerous.  

The shift becomes clearer when examining how governments are now framing AI risk. In 2026, guidance published by CISA, UK’s National Cyber Security Centre (NCSC) and Five Eyes partners warned that agentic systems expand attack surfaces, accumulate privilege, and can behave in ways that are difficult to predict or explain [1]. The advice is simple: assume unexpected behavior and design controls around it.  

The real risk is not AI usage. It is unknown autonomy: systems with credentials, data access, and action paths that can execute workflow steps without sufficient behavioral validation, traceability, or human oversight. Darktrace’s Model Context Protocol (MCP) risk analysis provides a useful framework for understanding this challenge. Over-privileged agents, content injection, and tool abuse become high-consequence risks when connected systems can dynamically retrieve data, execute actions, and communicate externally.  

Whether security teams like it or not, AI is already in the enterprise. It will help drive innovation, but it will also be abused, whether accidentally or maliciously. In each of the cases below, AI either scaled the attacker, built the tooling, or existed within the environment as something to exploit or misuse.

1. AI as an Attack Multiplier

In one campaign targeting Mexican government entities, a single operator used commercial AI platforms to generate exploits, automate reconnaissance, and process large volumes of data, compressing work that would traditionally have required an entire team into a single workflow [2].  

Darktrace is also observing this trend further down the stack. In one case, Darktrace identified AI-generated malware exploiting React2Shell, where an attacker used a Large Language Model (LLM) to produce working exploit code and deploy it at scale.  

[darktrace.com], [darktrace.com]

2. AI as an Attack Surface

Attempted AI exploitation is now appearing within customer environments. In one case involving an automation technology manufacturer, a compromised LLM proxy was seemingly used as a stepping stone to access additional AI services. When that attempt failed, the attacker pivoted to cryptomining.

What is clear is that the AI layer has already become an asset worth probing, exploiting, and pivoting through. It is also clear that defenders benefit from rapidly understanding how these activities connect. In this case, Cyber AI Analyst automatically pieced together the intrusion, while Darktrace’s Managed Threat Detection service alerted to the customer, enabling the activity to be contained before it could progress further.

Figure 3: Cyber AI Analyst's investigation into a compromised LLM proxy that was abused for cryptomining activity.

AI as a trusted but dangerous actor

This does not require a cinematic vision of “rogue AI.” The Salesloft incident provides a more grounded example, where AI and automation operate with legitimate access but served malicious intent. In that case, attackers abused compromised OAuth tokens associated with the Drift AI chat agent to export significant volumes of data from Salesforce environments.  

The activity resembled legitimate API usage and relied on trusted SaaS integrations rather than malware or other obvious signs of intrusion. That is precisely the challenge. Traditional security controls are good at detecting forced entry, but far less effective when a trusted application integration behaves in a way that is technically permitted yet operationally harmful.  

In these scenarios, the security challenge shifts from validating access to validating behavior.

This is what that looks like in practice: AI-linked identities executing legitimate actions that require behavioral validation rather than access validation.

Figure 4: Darktrace / SECURE AI highlights anomalous activity across AI identities, surfacing critical behavior that requires validation and containment.

Early observations from Darktrace / SECURE AI deployments reinforce this reality. Across Darktrace's observed fleet, AI service connections per deployment increased 13% during the first half of 2026, reaching over 16 million connections overall. The typical organisation now interacts with seven different AI providers, evidence that AI is no longer operating at the edges of the enterprise. It is increasingly woven into day-to-day business activity.

The most common risks are not compromised models or advanced AI attacks. Instead, they stem from employees and business functions exposing sensitive information through entirely legitimate-looking interactions. Darktrace has observed repeated submission of personally identifiable information (PII), tax information, identification documents, and medical data into LLM prompts, alongside widespread use of unsanctioned (shadow) AI services and growing AI activity from mobile devices.  

For defenders, the challenge is increasingly one of context: understanding when legitimate business use crosses into material risk, while preserving privacy and user trust.

Conclusion

Across all three shifts, the pattern is the same: behavior precedes understanding. Security teams are not losing because adversaries have become invisible. An increasingly outdated security model assumes that malicious activity will reveal itself cleanly and early. It no longer does.  

In 2026 and beyond, defenders win by understanding behavioral sequences, continuously validating trust, and acting before certainty becomes hindsight. That is security after signatures. That is security in the AI era.

Credit to: Daniel Levy, Threat Hunting Data Scientist

Edited by: Ryan Trail, Content Manager

References

[1] https://www.cyber.gov.au/business-government/secure-design/artificial-intelligence/careful-adoption-of-agentic-ai-services  

[2]https://www.latimes.com/business/story/2026-02-26/hacker-used-anthropics-claude-ai-to-steal-mexican-government-data

Continue reading
About the author
Nathaniel Jones
VP, Security & AI Strategy, Field CISO

Blog

/

AI

/

July 10, 2026

AIインフラがアタックサーフェスの一部に

Default blog imageDefault blog image

AIインフラとアタックサーフェスの進化

多くの組織が生成AIを実運用環境に導入するなかで、企業のクラウド環境内に新たなインフラのレイヤーが出現しています。それはAIゲートウェイです。AIゲートウェイはユーザー、アプリケーション、基盤モデルの間に位置し、多くの場合クラウドの特権アクセスを保持し、さまざまなAIサービスへのアクセスを大規模に管理しています。

AIゲートウェイとは?

AIゲートウェイはユーザー、アプリケーション、基盤モデルの間に位置し、多くの場合クラウドの特権アクセスを保持し、さまざまなAIサービスへのアクセスを大規模に管理しています。

こうした役割から、AIゲートウェイは企業のアタックサーフェスのますます重要な一部になりつつあります。AIゲートウェイが侵害されれば、攻撃者に対して計算リソースへのアクセスだけでなく、クラウドアイデンティティ、モデルサービス、機密性の高いプロンプト、そして他の接続されたシステムへのアクセスも提供してしまいます。

このブログでは、Amazon Bedrock サービスに接続されたAIゲートウェイが侵害され、その後暗号通貨マイニングインフラとの通信が観測された事例をダークトレースがどのように調査したかを解説します。問題のインスタンスは、その構成、ならびに関連するIAM(Identity and Access Management)ロールから、Amazon BedrockでホスティングされるAIサービスへのゲートウェイとして機能していることがわかりました。疑わしい侵害アクティビティが発生した後、このホストは既知の暗号通貨マイニングインフラに繰り返し通信を行い、その後シャットダウンされた様子が観測されました。Darktrace はこのアクティビティを検知し、Enhanced MonitoringおよびManaged Threat Detectionサービスを通じてエスカレーションを行いました。

この事例では最終的影響は不正な暗号通貨マイニングでしたが、このインシデントが注目に値するのはその発生場所です。侵害されたアセットは、クラウドインフラ、アイデンティティ、各種AIサービスの交差する場所に位置していました。最近の調査では、LiteLLM等のAIゲートウェイが、認証情報、モデルへのアクセス、クラウド権限を中央管理するその能力から、攻撃者にとって魅力的な標的となる可能性が明らかになっています。このアクティビティと公開されているLiteLLM脆弱性を直接結びつける証拠は見つかっていませんが、このインシデントは、AIインフラを個別のアプリケーション層として見るのではなく、重要なアタックサーフェスの一部として扱う必要性があることを表しています[1]。

暗号通貨マイニングがクラウド侵害後のアクティビティとしてよく見られる背景

暗号通貨マイニングはクラウド環境において、侵害後のアクティビティとして収益性の高いものとなり得ます。クラウド資産にアクセスできるようになった後、攻撃者はマイニングソフトウェアを展開して被害者の計算リソースを悪用し金銭的利益を得ることができます。この種のアクティビティは多くの場合機会主義的なものであり、露出したサービス、弱い認証情報、漏洩したアクセスキー、脆弱なアプリケーション、あるいはクラウドワークロードの設定ミスなどを標的として実行されます。

典型的なクラウド上での暗号通貨マイニング侵入には次のようなアクティビティが含まれます:

  • 露出したあるいは脆弱なクラウドインフラの特定
  • 露出したサービス、認証情報、またはアプリケーションの脆弱性を通じたアクセスの獲得
  • マイニングソフトウェアのダウンロードおよび実行
  • マイニングプールインフラへのアウトバウンド接続を繰り返し確立
  • アクティビティが検知され停止されるまで継続して計算リソースを消費

この事例において注目すべき要素は暗号通貨マイニングだけではありません。それが発生した場所が、AI関連アクティビティをサポートするクラウドインフラ上だったことです。この事例は、AIサービスを実現するためのアセットも、よくあるクラウド侵害リスクにさらされる可能性があることを示しています。

Amazon Bedrockに接続されたAIゲートウェイの侵害を調査

2026年6月12日、DarktraceはLiteLLM-Proxyという名前のAmazon Web Service (AWS) EC2インスタンスから暗号通貨マイニング発生中とみられるアクティビティを観測しました。このインスタンスはLiteLLMアクティビティをサポートしており、Amazon Bedrockリソースへのアクセス権を有するインスタンスプロファイルと関連付けられていました。  

AIゲートウェイは大規模言語モデルへのアクセスを中央管理するよう設計されており、多くの場合AIアプリケーションに対する認証、ルーティング、ログ、ポリシー適用を扱っています。セキュリティの視点から見ると、クラウド権限、モデルアクセス、アプリケーションワークフローを単一の制御ポイントに集約する役割も果たしています。その結果、AIゲートウェイの侵害は、侵害されたホストだけにとどまらない影響を及ぼす可能性があります。

確定的な初期アクセスベクトルは確認できませんでしたが、このアクティビティはインターネットに接続されているシステムの侵害でよく見られる次のような順序に従っていました。ブルートフォースアクセス、ペイロードの投下、そしてマイニングプールインフラに対する繰り返しのアウトバウンド接続です。

ステージ1: インターネットに露出したSSHからの初期アクセス

暗号通貨マイニングアクティビティが観測される前、LiteLLM-Proxy EC2インスタンスはSSH(ポート22)が0.0.0.0/0に対して開かれ、外部に公開されていました。

図1:EC2インスタンスがSSHポート22に対してすべてのインバウンドトラフィックを許可している設定ミスをDarktraceが警告

暗号通貨マイニングアクティビティに先立って、Darktraceはこのインスタンスに対する大量のインバウンド接続の試みが外部IPアドレス(主に145.241.123[.]102)からポート22に対して行われていることを観測しました。これはブルートフォースアクティビティを示唆するものです [2]。これらの接続の多くは短命であり、数秒しか続いておらず、スキャニングまたはログインの失敗を示していました。

図2:Darktraceがデバイスのポート22に対する不審なインバウンド接続試行を検知

入手できたテレメトリーではこれらのインバウンドSSH接続のいずれかが認証の成功につながったかどうかの確認に至らず、このアクティビティが初期アクセスベクトルであると断定することはできませんでした。しかしながら、SSHの露出、外部IPアドレスからのインバウンド接続、それに続くマイニングアクティビティは、SSHがアクセス経路の可能性が高いことを示唆しています。

ステージ2: AIゲートウェイへのXMRigマルウェアのダウンロード

最初に観測されたマイニングプールへの接続の後、このEC2インスタンスは3.42 MBのデータをポート80上のHTTP接続を介して外部エンドポイント185.62.1[.]8にダウンロードしました。このエンドポイントは暗号通貨マイニングマルウェアXMRigを含むZIPファイルをホスティングしていました[3][4]。ホストレベルのログは入手できなかったため、ダークトレースはマイニングツールがどのように実行されたか、あるいは前のSSHアクティビティがペイロード投下を直接的に可能にしたかどうかを確認できませんでした。しかしながら、ダウンロードのタイミングとその後ほどなくマイニングプールへの接続が繰り返されたことは、このインスタンスが侵害されて不正な計算アクティビティに使われたという評価を裏付けています。

ステージ3 – 侵害されたAIゲートウェイが暗号通貨マイニングインフラと通信

わずか数分後、DarktraceはLiteLLM-ProxyEC2インスタンスがHTTPs(ポート443)でホスト名pool.hasvault[.]proに対して接続していることを確認しました。最初の接続の後、同じホスト名に対して繰り返しアウトバウンド接続が観測されました。これは、侵害されたホストがマイニングインフラと通信しワークを受け取り、結果を送信するという、暗号通貨マイニングプールとの通信のパターンと一致しています。

このアクティビティがDarktraceのEnhanced Monitoringモデル“Compromise / HighPriority Crypto Currency Mining”をトリガーし、ダークトレースのSOCにより顧客に対してエスカレーションされました。また、このアクティビティはCyber AI Analystによって分析され、関連するイベントが1つの調査ナラティブにまとめられました。これにより、影響を受けたクラウドアセットからマニングプールへの繰り返しの接続を特定することができました。

図3:CyberAI Analystによる暗号通貨マイニングアクティビティの調査  

ポート443上のHTTPSの使用にも注目すべきです。なぜならば、単独で見れば、このトラフィックそのものは疑わしく見えないかもしれないからです。しかしこのケースでは、接続先、接続の量、そして類似のアクティビティが他にないことなどが、この通信を疑わしいものとして特定するのに必要な、動作のコンテキストを提供することになりました。

ステージ4: Managed Threat Detectionサービスによるリソース乱用の特定

暗号通貨マイニングアクティビティがダークトレースのManaged Threat Detectionサービスにより検知され、ダークトレースのSOCによりレビューされました。レビューの結果、このアクティビティは顧客向けにエスカレーションされました。このエスカレーションにより、顧客はAWS環境で現在発生中のリソースの乱用について、タイムリーな通知を受けることができました。

ステージ5: クラウド認証情報の不正使用とみられる疑わしいIAMアクティビティ

これとは別に、6月13日、Darktraceは別のIAMユーザーから発生した疑わしいアクティビティを検知しました。

図4: DarktraceのAdvanced Search機能が別のIAMユーザーが実行した疑わしいアクティビティをハイライト

まず、このユーザーは “GetSendQuota”イベントを試行している様子が見られました。このアクションは少なくとも過去3か月間にこのアカウントによって実行されたことのないアクションです。また、このコマンドのソースIPアドレスは14.176.1[.]47でした。地理位置情報はベトナムであり、このユーザーのアクティビティがAmazon IPアドレスから最も多く見られた場所です。さらに、このアクティビティに対してAWS CLIが使用されており、これもこのユーザーにとって通常とは異なる振る舞いでした。このことは、Darktraceの“IaaS / Unusual Activity / UnusualAWS CLI Activity”モデルによって検知されました。

図5: Darktraceによる “GetSendQuota” イベントの検知

このIAMユーザーからは、長期アクセスキーを使った疑わしいアクティビティがさらに観測されました。中でも、“InvokeModel” および “ListFoundationModels”コマンドの失敗が検知されており、モデル列挙や起動などAmazon Bedrockサービスとのやり取りを試行したことがわかります。これは前日観測されたLiteLLM侵害への関連を思わせますが、2つのイベントを確定的に結びつける証拠は不十分でした。

“CreateUser”コマンドの試行も注目に値します。なぜなら要求されたユーザー名は意味が薄いものであり、新しいアカウントを作成することにより永続性を確立する試みと見られるからです。このアクティビティはDarktraceのモデル“IaaS / Admin / New AWS UserAccount Creation”をトリガーしました。

図6:Darktraceによる“CreateUser” イベントの検知

2つのインシデント間に結びつきは確認できなかったものの、このIAMアクティビティには重要な意味があります。これは、クラウド侵害の調査においてワークロードのテレメトリーとコントロールプレーンのテレメトリーの両方を取り入れることの重要性を表しています。EC2暗号通貨マイニングアクティビティが計算リソースの乱用を示す一方、IAMアクティビティは認証情報の侵害や長期アクセスキーの不正使用、そしてクラウトサービスの不正使用の可能性を示唆しているからです。

AIインフラ保護のための重要な教訓

このインシデントの重大性は暗号通貨マイニングアクティビティそのものではなく、それが発生した場所にあります。侵害されたシステムはAmazon Bedrockサービスへのアクセス権を持つAIゲートウェイとして機能し、クラウドインフラ、アイデンティティ、そしてさまざまなAIオペレーションの交差する場所に位置していました。組織がAI機能を実運用環境に導入していくなかで、これらのプラットフォームは、露出したサービス、認証情報窃取、クラウドの設定ミスなどを通じて攻撃者がすでに狙っているアタックサーフェスの一部となりつつあるのです。

このケースでは詳細な侵入経路は特定されておらず、ワークロードの侵害と調査中に検知された疑わしいIAMアクティビティの間に決定的なつながりは確認されませんでしたが、これらのイベントは全体的な現状を裏付けています。つまり、AIインフラは個別のテクノロジースタックとして扱うのではなく、クラウド環境全体の一部として保護しなければならないとうことです。

このケースでは、最も目立った侵害の兆候は暗号通貨マイニングインフラとの通信でした。しかしここで得られたより重要な教訓は、このインシデントの全貌が理解される前にDarktraceのビヘイビア分析により明らかになった、高い権限を持つAI関連アセットを取り巻くリスクです。AIゲートウェイによりクラウド権限、モデルアクセス、アプリケーションワークフローがますます集約されるなかで、防御者は個別のアラートに集中するよりも、ワークロード、アイデンティティ、サービスの間でどのように動作がつながっているかを理解することに重点を置く必要があるでしょう。

協力:Angel Arribas Lopez (Associate Principal Cyber Analyst)、Nathaniel Jones (Field CISO/VP Threat Research)、Emma Foulger (Global Threat Ops)、Mark Turner(Security Researcher)

編集:Ryan Traill (Content Manager)

付録

Darktraceによるモデル検知結果

·       Compromise / High Priority Crypto Currency Mining

·       Compromise / Monero Mining

·       Device / Internet Facing Device with High Priority Alert

·       IaaS / Unusual Activity / Unusual AWS CLI Activity

·       IaaS / Admin / New AWS User Account Creation

MITRE ATT&CK マッピング

初期アクセス – 外部リモートサービス – T1133

初期アクセス – 有効なアカウント – T1078

実行 – コマンドおよびスクリプトインタプリタ – T1059

永続化 – アカウント作成 – T1136

探索 – クラウドサービス探索 – T1526

影響 – リソースハイジャッキング– T1496

参考資料

[1] https://docs.litellm.ai/blog/security-update-march-2026

[2] https://www.abuseipdb.com/check/145.241.123.102

[3] https://urlscan.io/search/#185.62.1.8

[4] https://www.virustotal.com/gui/file/85de36ff66fae9f4b059cbedf6d36e017ebc26c828f99f911a96e78636f21200/community

Continue reading
About the author
Angel Arribas Lopez
Associate Principal Cyber Analyst
あなたのデータ × DarktraceのAI
唯一無二のDarktrace AIで、ネットワークセキュリティを次の次元へ